{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 5.050383286332764,
  "global_step": 25177,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.0,
      "learning_rate": 9.92063492063492e-07,
      "loss": 10.8952,
      "theoretical_loss": 20.81281780154715,
      "tokens_seen": 65536
    },
    {
      "epoch": 0.0,
      "learning_rate": 1.984126984126984e-06,
      "loss": 10.8925,
      "theoretical_loss": 17.566201104328645,
      "tokens_seen": 131072
    },
    {
      "epoch": 0.0,
      "learning_rate": 2.9761904761904763e-06,
      "loss": 10.8113,
      "theoretical_loss": 15.939477092836569,
      "tokens_seen": 196608
    },
    {
      "epoch": 0.0,
      "learning_rate": 3.968253968253968e-06,
      "loss": 10.7096,
      "theoretical_loss": 14.89231675598857,
      "tokens_seen": 262144
    },
    {
      "epoch": 0.0,
      "learning_rate": 4.96031746031746e-06,
      "loss": 10.516,
      "theoretical_loss": 14.136216937762974,
      "tokens_seen": 327680
    },
    {
      "epoch": 0.0,
      "learning_rate": 5.9523809523809525e-06,
      "loss": 10.5507,
      "theoretical_loss": 13.552561472550224,
      "tokens_seen": 393216
    },
    {
      "epoch": 0.0,
      "learning_rate": 6.944444444444444e-06,
      "loss": 10.1736,
      "theoretical_loss": 13.08180900140119,
      "tokens_seen": 458752
    },
    {
      "epoch": 0.0,
      "learning_rate": 7.936507936507936e-06,
      "loss": 10.0954,
      "theoretical_loss": 12.690129625483323,
      "tokens_seen": 524288
    },
    {
      "epoch": 0.0,
      "learning_rate": 8.928571428571428e-06,
      "loss": 9.9354,
      "theoretical_loss": 12.356592463873625,
      "tokens_seen": 589824
    },
    {
      "epoch": 0.0,
      "learning_rate": 9.92063492063492e-06,
      "loss": 9.9176,
      "theoretical_loss": 12.067412607035077,
      "tokens_seen": 655360
    },
    {
      "epoch": 0.0,
      "learning_rate": 1.0912698412698412e-05,
      "loss": 9.8329,
      "theoretical_loss": 11.813066231101676,
      "tokens_seen": 720896
    },
    {
      "epoch": 0.0,
      "learning_rate": 1.1904761904761905e-05,
      "loss": 9.5392,
      "theoretical_loss": 11.586719208706729,
      "tokens_seen": 786432
    },
    {
      "epoch": 0.0,
      "learning_rate": 1.2896825396825396e-05,
      "loss": 9.6937,
      "theoretical_loss": 11.383314140186787,
      "tokens_seen": 851968
    },
    {
      "epoch": 0.0,
      "learning_rate": 1.3888888888888888e-05,
      "loss": 9.4581,
      "theoretical_loss": 11.199011702111871,
      "tokens_seen": 917504
    },
    {
      "epoch": 0.0,
      "learning_rate": 1.4880952380952381e-05,
      "loss": 9.4229,
      "theoretical_loss": 11.030833917977912,
      "tokens_seen": 983040
    },
    {
      "epoch": 0.0,
      "learning_rate": 1.5873015873015872e-05,
      "loss": 9.4341,
      "theoretical_loss": 10.87642808645695,
      "tokens_seen": 1048576
    },
    {
      "epoch": 0.0,
      "learning_rate": 1.6865079365079364e-05,
      "loss": 9.3235,
      "theoretical_loss": 10.733905740062724,
      "tokens_seen": 1114112
    },
    {
      "epoch": 0.0,
      "learning_rate": 1.7857142857142855e-05,
      "loss": 9.2909,
      "theoretical_loss": 10.60172987623028,
      "tokens_seen": 1179648
    },
    {
      "epoch": 0.0,
      "learning_rate": 1.884920634920635e-05,
      "loss": 9.2573,
      "theoretical_loss": 10.478634172356642,
      "tokens_seen": 1245184
    },
    {
      "epoch": 0.0,
      "learning_rate": 1.984126984126984e-05,
      "loss": 9.159,
      "theoretical_loss": 10.36356394376333,
      "tokens_seen": 1310720
    },
    {
      "epoch": 0.0,
      "learning_rate": 2.0833333333333333e-05,
      "loss": 9.0708,
      "theoretical_loss": 10.255632220896747,
      "tokens_seen": 1376256
    },
    {
      "epoch": 0.0,
      "learning_rate": 2.1825396825396824e-05,
      "loss": 9.2882,
      "theoretical_loss": 10.15408655327002,
      "tokens_seen": 1441792
    },
    {
      "epoch": 0.0,
      "learning_rate": 2.2817460317460315e-05,
      "loss": 8.981,
      "theoretical_loss": 10.058283561732598,
      "tokens_seen": 1507328
    },
    {
      "epoch": 0.0,
      "learning_rate": 2.380952380952381e-05,
      "loss": 9.1814,
      "theoretical_loss": 9.967669178840278,
      "tokens_seen": 1572864
    },
    {
      "epoch": 0.0,
      "objective/train/docs_used": 67480,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 9.154304504394531,
      "objective/train/theoretical_loss": 9.881763126393109,
      "objective/train/tokens_used": 22098400,
      "theoretical_loss": 9.881763126393109,
      "tokens_seen": 1638400
    },
    {
      "epoch": 0.0,
      "learning_rate": 2.48015873015873e-05,
      "loss": 9.2744,
      "theoretical_loss": 9.881763126393109,
      "tokens_seen": 1638400
    },
    {
      "epoch": 0.0,
      "learning_rate": 2.5793650793650793e-05,
      "loss": 9.1924,
      "theoretical_loss": 9.80014659154056,
      "tokens_seen": 1703936
    },
    {
      "epoch": 0.0,
      "learning_rate": 2.6785714285714284e-05,
      "loss": 9.1481,
      "theoretical_loss": 9.722452346907446,
      "tokens_seen": 1769472
    },
    {
      "epoch": 0.0,
      "learning_rate": 2.7777777777777776e-05,
      "loss": 8.8718,
      "theoretical_loss": 9.648356759081546,
      "tokens_seen": 1835008
    },
    {
      "epoch": 0.0,
      "learning_rate": 2.876984126984127e-05,
      "loss": 9.0305,
      "theoretical_loss": 9.577573271145639,
      "tokens_seen": 1900544
    },
    {
      "epoch": 0.0,
      "learning_rate": 2.9761904761904762e-05,
      "loss": 8.9527,
      "theoretical_loss": 9.509847046764852,
      "tokens_seen": 1966080
    },
    {
      "epoch": 0.0,
      "learning_rate": 3.075396825396825e-05,
      "loss": 8.8801,
      "theoretical_loss": 9.444950537631936,
      "tokens_seen": 2031616
    },
    {
      "epoch": 0.0,
      "learning_rate": 3.1746031746031745e-05,
      "loss": 8.7157,
      "theoretical_loss": 9.382679790910457,
      "tokens_seen": 2097152
    },
    {
      "epoch": 0.0,
      "learning_rate": 3.273809523809524e-05,
      "loss": 8.6865,
      "theoretical_loss": 9.32285135423398,
      "tokens_seen": 2162688
    },
    {
      "epoch": 0.0,
      "learning_rate": 3.373015873015873e-05,
      "loss": 8.7653,
      "theoretical_loss": 9.265299666660276,
      "tokens_seen": 2228224
    },
    {
      "epoch": 0.0,
      "learning_rate": 3.472222222222222e-05,
      "loss": 8.6092,
      "theoretical_loss": 9.209874847444755,
      "tokens_seen": 2293760
    },
    {
      "epoch": 0.0,
      "learning_rate": 3.571428571428571e-05,
      "loss": 8.7534,
      "theoretical_loss": 9.156440812508292,
      "tokens_seen": 2359296
    },
    {
      "epoch": 0.0,
      "learning_rate": 3.670634920634921e-05,
      "loss": 8.5526,
      "theoretical_loss": 9.10487366241335,
      "tokens_seen": 2424832
    },
    {
      "epoch": 0.0,
      "learning_rate": 3.76984126984127e-05,
      "loss": 8.4992,
      "theoretical_loss": 9.055060296533734,
      "tokens_seen": 2490368
    },
    {
      "epoch": 0.0,
      "learning_rate": 3.8690476190476195e-05,
      "loss": 8.5932,
      "theoretical_loss": 9.006897216643829,
      "tokens_seen": 2555904
    },
    {
      "epoch": 0.0,
      "learning_rate": 3.968253968253968e-05,
      "loss": 8.5997,
      "theoretical_loss": 8.960289489909357,
      "tokens_seen": 2621440
    },
    {
      "epoch": 0.0,
      "learning_rate": 4.067460317460318e-05,
      "loss": 8.3689,
      "theoretical_loss": 8.915149846640611,
      "tokens_seen": 2686976
    },
    {
      "epoch": 0.0,
      "learning_rate": 4.1666666666666665e-05,
      "loss": 8.3045,
      "theoretical_loss": 8.871397892478225,
      "tokens_seen": 2752512
    },
    {
      "epoch": 0.0,
      "learning_rate": 4.265873015873016e-05,
      "loss": 8.3696,
      "theoretical_loss": 8.828959418153499,
      "tokens_seen": 2818048
    },
    {
      "epoch": 0.0,
      "learning_rate": 4.365079365079365e-05,
      "loss": 8.3294,
      "theoretical_loss": 8.787765792778412,
      "tokens_seen": 2883584
    },
    {
      "epoch": 0.0,
      "learning_rate": 4.464285714285714e-05,
      "loss": 7.9479,
      "theoretical_loss": 8.747753428911455,
      "tokens_seen": 2949120
    },
    {
      "epoch": 0.0,
      "learning_rate": 4.563492063492063e-05,
      "loss": 7.9269,
      "theoretical_loss": 8.708863309520833,
      "tokens_seen": 3014656
    },
    {
      "epoch": 0.0,
      "learning_rate": 4.6626984126984126e-05,
      "loss": 7.9812,
      "theoretical_loss": 8.671040568508847,
      "tokens_seen": 3080192
    },
    {
      "epoch": 0.0,
      "learning_rate": 4.761904761904762e-05,
      "loss": 8.0261,
      "theoretical_loss": 8.634234117735474,
      "tokens_seen": 3145728
    },
    {
      "epoch": 0.0,
      "learning_rate": 4.8611111111111115e-05,
      "loss": 8.2961,
      "theoretical_loss": 8.598396314536323,
      "tokens_seen": 3211264
    },
    {
      "epoch": 0.0,
      "objective/train/docs_used": 72733,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 8.41878604888916,
      "objective/train/theoretical_loss": 8.563482664611069,
      "objective/train/tokens_used": 23736800,
      "theoretical_loss": 8.563482664611069,
      "tokens_seen": 3276800
    },
    {
      "epoch": 0.0,
      "learning_rate": 4.96031746031746e-05,
      "loss": 8.2388,
      "theoretical_loss": 8.563482664611069,
      "tokens_seen": 3276800
    },
    {
      "epoch": 0.0,
      "learning_rate": 5.05952380952381e-05,
      "loss": 7.8592,
      "theoretical_loss": 8.529451555895115,
      "tokens_seen": 3342336
    },
    {
      "epoch": 0.0,
      "learning_rate": 5.1587301587301586e-05,
      "loss": 7.7882,
      "theoretical_loss": 8.496264019646002,
      "tokens_seen": 3407872
    },
    {
      "epoch": 0.0,
      "learning_rate": 5.257936507936508e-05,
      "loss": 8.0092,
      "theoretical_loss": 8.463883515497187,
      "tokens_seen": 3473408
    },
    {
      "epoch": 0.0,
      "learning_rate": 5.357142857142857e-05,
      "loss": 7.8125,
      "theoretical_loss": 8.432275737672779,
      "tokens_seen": 3538944
    },
    {
      "epoch": 0.0,
      "learning_rate": 5.4563492063492063e-05,
      "loss": 8.1055,
      "theoretical_loss": 8.401408439930716,
      "tokens_seen": 3604480
    },
    {
      "epoch": 0.0,
      "learning_rate": 5.555555555555555e-05,
      "loss": 7.6534,
      "theoretical_loss": 8.371251277120209,
      "tokens_seen": 3670016
    },
    {
      "epoch": 0.0,
      "learning_rate": 5.6547619047619046e-05,
      "loss": 7.8082,
      "theoretical_loss": 8.341775661511075,
      "tokens_seen": 3735552
    },
    {
      "epoch": 0.0,
      "learning_rate": 5.753968253968254e-05,
      "loss": 7.7956,
      "theoretical_loss": 8.31295463228533,
      "tokens_seen": 3801088
    },
    {
      "epoch": 0.0,
      "learning_rate": 5.8531746031746036e-05,
      "loss": 7.7599,
      "theoretical_loss": 8.284762736781182,
      "tokens_seen": 3866624
    },
    {
      "epoch": 0.0,
      "learning_rate": 5.9523809523809524e-05,
      "loss": 7.683,
      "theoretical_loss": 8.257175922251864,
      "tokens_seen": 3932160
    },
    {
      "epoch": 0.0,
      "learning_rate": 6.051587301587302e-05,
      "loss": 7.5917,
      "theoretical_loss": 8.230171437050114,
      "tokens_seen": 3997696
    },
    {
      "epoch": 0.0,
      "learning_rate": 6.15079365079365e-05,
      "loss": 7.6448,
      "theoretical_loss": 8.20372774027797,
      "tokens_seen": 4063232
    },
    {
      "epoch": 0.0,
      "learning_rate": 6.25e-05,
      "loss": 7.5636,
      "theoretical_loss": 8.177824419053046,
      "tokens_seen": 4128768
    },
    {
      "epoch": 0.0,
      "learning_rate": 6.349206349206349e-05,
      "loss": 7.3527,
      "theoretical_loss": 8.152442112639616,
      "tokens_seen": 4194304
    },
    {
      "epoch": 0.0,
      "learning_rate": 6.448412698412699e-05,
      "loss": 7.6386,
      "theoretical_loss": 8.1275624427775,
      "tokens_seen": 4259840
    },
    {
      "epoch": 0.0,
      "learning_rate": 6.547619047619048e-05,
      "loss": 7.4188,
      "theoretical_loss": 8.10316794961571,
      "tokens_seen": 4325376
    },
    {
      "epoch": 0.0,
      "learning_rate": 6.646825396825397e-05,
      "loss": 7.5008,
      "theoretical_loss": 8.07924203272264,
      "tokens_seen": 4390912
    },
    {
      "epoch": 0.0,
      "learning_rate": 6.746031746031745e-05,
      "loss": 7.5319,
      "theoretical_loss": 8.055768896701416,
      "tokens_seen": 4456448
    },
    {
      "epoch": 0.0,
      "learning_rate": 6.845238095238096e-05,
      "loss": 7.048,
      "theoretical_loss": 8.032733500989007,
      "tokens_seen": 4521984
    },
    {
      "epoch": 0.0,
      "learning_rate": 6.944444444444444e-05,
      "loss": 7.2613,
      "theoretical_loss": 8.010121513461836,
      "tokens_seen": 4587520
    },
    {
      "epoch": 0.0,
      "learning_rate": 7.043650793650793e-05,
      "loss": 7.2432,
      "theoretical_loss": 7.987919267509379,
      "tokens_seen": 4653056
    },
    {
      "epoch": 0.0,
      "learning_rate": 7.142857142857142e-05,
      "loss": 7.2,
      "theoretical_loss": 7.966113722271801,
      "tokens_seen": 4718592
    },
    {
      "epoch": 0.0,
      "learning_rate": 7.242063492063492e-05,
      "loss": 7.2711,
      "theoretical_loss": 7.944692425767988,
      "tokens_seen": 4784128
    },
    {
      "epoch": 0.0,
      "learning_rate": 7.341269841269842e-05,
      "loss": 7.2298,
      "theoretical_loss": 7.9236434806675184,
      "tokens_seen": 4849664
    },
    {
      "epoch": 0.0,
      "objective/train/docs_used": 77606,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 7.1540398597717285,
      "objective/train/theoretical_loss": 7.902955512484067,
      "objective/train/tokens_used": 25375200,
      "theoretical_loss": 7.902955512484067,
      "tokens_seen": 4915200
    },
    {
      "epoch": 0.0,
      "learning_rate": 7.440476190476191e-05,
      "loss": 7.0498,
      "theoretical_loss": 7.902955512484067,
      "tokens_seen": 4915200
    },
    {
      "epoch": 0.0,
      "learning_rate": 7.53968253968254e-05,
      "loss": 6.8919,
      "theoretical_loss": 7.882617639989203,
      "tokens_seen": 4980736
    },
    {
      "epoch": 0.0,
      "learning_rate": 7.63888888888889e-05,
      "loss": 7.1864,
      "theoretical_loss": 7.862619447664628,
      "tokens_seen": 5046272
    },
    {
      "epoch": 0.0,
      "learning_rate": 7.738095238095239e-05,
      "loss": 7.153,
      "theoretical_loss": 7.842950960027937,
      "tokens_seen": 5111808
    },
    {
      "epoch": 0.0,
      "learning_rate": 7.837301587301588e-05,
      "loss": 7.0043,
      "theoretical_loss": 7.823602617682313,
      "tokens_seen": 5177344
    },
    {
      "epoch": 0.0,
      "learning_rate": 7.936507936507937e-05,
      "loss": 7.067,
      "theoretical_loss": 7.804565254954165,
      "tokens_seen": 5242880
    },
    {
      "epoch": 0.0,
      "learning_rate": 8.035714285714287e-05,
      "loss": 7.0251,
      "theoretical_loss": 7.7858300789950725,
      "tokens_seen": 5308416
    },
    {
      "epoch": 0.0,
      "learning_rate": 8.134920634920635e-05,
      "loss": 6.7699,
      "theoretical_loss": 7.767388650235364,
      "tokens_seen": 5373952
    },
    {
      "epoch": 0.0,
      "learning_rate": 8.234126984126984e-05,
      "loss": 6.8388,
      "theoretical_loss": 7.749232864086619,
      "tokens_seen": 5439488
    },
    {
      "epoch": 0.0,
      "learning_rate": 8.333333333333333e-05,
      "loss": 6.9893,
      "theoretical_loss": 7.731354933799318,
      "tokens_seen": 5505024
    },
    {
      "epoch": 0.0,
      "learning_rate": 8.432539682539683e-05,
      "loss": 6.819,
      "theoretical_loss": 7.71374737438992,
      "tokens_seen": 5570560
    },
    {
      "epoch": 0.0,
      "learning_rate": 8.531746031746032e-05,
      "loss": 6.6608,
      "theoretical_loss": 7.696402987558934,
      "tokens_seen": 5636096
    },
    {
      "epoch": 0.0,
      "learning_rate": 8.630952380952381e-05,
      "loss": 6.5322,
      "theoretical_loss": 7.679314847528181,
      "tokens_seen": 5701632
    },
    {
      "epoch": 0.0,
      "learning_rate": 8.73015873015873e-05,
      "loss": 6.7129,
      "theoretical_loss": 7.662476287731328,
      "tokens_seen": 5767168
    },
    {
      "epoch": 0.0,
      "learning_rate": 8.82936507936508e-05,
      "loss": 6.7745,
      "theoretical_loss": 7.645880888297279,
      "tokens_seen": 5832704
    },
    {
      "epoch": 0.0,
      "learning_rate": 8.928571428571429e-05,
      "loss": 6.8321,
      "theoretical_loss": 7.629522464270861,
      "tokens_seen": 5898240
    },
    {
      "epoch": 0.0,
      "learning_rate": 9.027777777777777e-05,
      "loss": 6.6706,
      "theoretical_loss": 7.613395054519696,
      "tokens_seen": 5963776
    },
    {
      "epoch": 0.0,
      "learning_rate": 9.126984126984126e-05,
      "loss": 6.6371,
      "theoretical_loss": 7.59749291128028,
      "tokens_seen": 6029312
    },
    {
      "epoch": 0.0,
      "learning_rate": 9.226190476190476e-05,
      "loss": 6.6738,
      "theoretical_loss": 7.581810490299888,
      "tokens_seen": 6094848
    },
    {
      "epoch": 0.0,
      "learning_rate": 9.325396825396825e-05,
      "loss": 6.6876,
      "theoretical_loss": 7.5663424415343705,
      "tokens_seen": 6160384
    },
    {
      "epoch": 0.0,
      "learning_rate": 9.424603174603175e-05,
      "loss": 6.5013,
      "theoretical_loss": 7.551083600364949,
      "tokens_seen": 6225920
    },
    {
      "epoch": 0.0,
      "learning_rate": 9.523809523809524e-05,
      "loss": 6.5264,
      "theoretical_loss": 7.536028979299919,
      "tokens_seen": 6291456
    },
    {
      "epoch": 0.0,
      "learning_rate": 9.623015873015874e-05,
      "loss": 6.4506,
      "theoretical_loss": 7.521173760129762,
      "tokens_seen": 6356992
    },
    {
      "epoch": 0.0,
      "learning_rate": 9.722222222222223e-05,
      "loss": 6.4093,
      "theoretical_loss": 7.506513286506497,
      "tokens_seen": 6422528
    },
    {
      "epoch": 0.0,
      "learning_rate": 9.821428571428572e-05,
      "loss": 6.3955,
      "theoretical_loss": 7.492043056920249,
      "tokens_seen": 6488064
    },
    {
      "epoch": 0.0,
      "objective/train/docs_used": 82513,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 6.486067771911621,
      "objective/train/theoretical_loss": 7.4777587180480305,
      "objective/train/tokens_used": 27013600,
      "theoretical_loss": 7.4777587180480305,
      "tokens_seen": 6553600
    },
    {
      "epoch": 0.0,
      "learning_rate": 9.92063492063492e-05,
      "loss": 6.6458,
      "theoretical_loss": 7.4777587180480305,
      "tokens_seen": 6553600
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00010019841269841271,
      "loss": 6.6237,
      "theoretical_loss": 7.463656058451462,
      "tokens_seen": 6619136
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0001011904761904762,
      "loss": 6.5474,
      "theoretical_loss": 7.449731002601916,
      "tokens_seen": 6684672
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00010218253968253968,
      "loss": 6.3712,
      "theoretical_loss": 7.435979605213019,
      "tokens_seen": 6750208
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00010317460317460317,
      "loss": 6.7222,
      "theoretical_loss": 7.422398045861905,
      "tokens_seen": 6815744
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00010416666666666667,
      "loss": 6.7042,
      "theoretical_loss": 7.408982623881875,
      "tokens_seen": 6881280
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00010515873015873016,
      "loss": 6.5424,
      "theoretical_loss": 7.395729753510345,
      "tokens_seen": 6946816
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00010615079365079365,
      "loss": 6.4443,
      "theoretical_loss": 7.3826359592770325,
      "tokens_seen": 7012352
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00010714285714285714,
      "loss": 6.4673,
      "theoretical_loss": 7.369697871618373,
      "tokens_seen": 7077888
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00010813492063492064,
      "loss": 6.7002,
      "theoretical_loss": 7.3569122227050885,
      "tokens_seen": 7143424
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00010912698412698413,
      "loss": 6.2387,
      "theoretical_loss": 7.3442758424706875,
      "tokens_seen": 7208960
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00011011904761904761,
      "loss": 6.3967,
      "theoretical_loss": 7.331785654829519,
      "tokens_seen": 7274496
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0001111111111111111,
      "loss": 6.3247,
      "theoretical_loss": 7.319438674073677,
      "tokens_seen": 7340032
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0001121031746031746,
      "loss": 6.4478,
      "theoretical_loss": 7.307232001438824,
      "tokens_seen": 7405568
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00011309523809523809,
      "loss": 6.3822,
      "theoretical_loss": 7.295162821829564,
      "tokens_seen": 7471104
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00011408730158730158,
      "loss": 6.57,
      "theoretical_loss": 7.283228400695652,
      "tokens_seen": 7536640
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00011507936507936508,
      "loss": 6.351,
      "theoretical_loss": 7.271426081050832,
      "tokens_seen": 7602176
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00011607142857142858,
      "loss": 6.1891,
      "theoretical_loss": 7.259753280626623,
      "tokens_seen": 7667712
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00011706349206349207,
      "loss": 6.2692,
      "theoretical_loss": 7.24820748915387,
      "tokens_seen": 7733248
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00011805555555555556,
      "loss": 6.3516,
      "theoretical_loss": 7.236786265765262,
      "tokens_seen": 7798784
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00011904761904761905,
      "loss": 6.1213,
      "theoretical_loss": 7.225487236512497,
      "tokens_seen": 7864320
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00012003968253968255,
      "loss": 6.3916,
      "theoretical_loss": 7.21430809199212,
      "tokens_seen": 7929856
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00012103174603174604,
      "loss": 6.4588,
      "theoretical_loss": 7.2032465850744005,
      "tokens_seen": 7995392
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00012202380952380953,
      "loss": 6.1372,
      "theoretical_loss": 7.192300528730015,
      "tokens_seen": 8060928
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.000123015873015873,
      "loss": 6.2326,
      "theoretical_loss": 7.1814677939495155,
      "tokens_seen": 8126464
    },
    {
      "epoch": 0.0,
      "objective/train/docs_used": 87567,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 6.155417442321777,
      "objective/train/theoretical_loss": 7.1707463077509646,
      "objective/train/tokens_used": 28652000,
      "theoretical_loss": 7.1707463077509646,
      "tokens_seen": 8192000
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0001240079365079365,
      "loss": 6.3427,
      "theoretical_loss": 7.1707463077509646,
      "tokens_seen": 8192000
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.000125,
      "loss": 6.1035,
      "theoretical_loss": 7.160134051271272,
      "tokens_seen": 8257536
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0001259920634920635,
      "loss": 6.2199,
      "theoretical_loss": 7.149629057937138,
      "tokens_seen": 8323072
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00012698412698412698,
      "loss": 6.1004,
      "theoretical_loss": 7.139229411711638,
      "tokens_seen": 8388608
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00012797619047619048,
      "loss": 6.084,
      "theoretical_loss": 7.128933245412794,
      "tokens_seen": 8454144
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00012896825396825398,
      "loss": 5.9073,
      "theoretical_loss": 7.118738739100616,
      "tokens_seen": 8519680
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00012996031746031748,
      "loss": 6.1843,
      "theoretical_loss": 7.1086441185293445,
      "tokens_seen": 8585216
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00013095238095238096,
      "loss": 5.9909,
      "theoretical_loss": 7.09864765366177,
      "tokens_seen": 8650752
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00013194444444444446,
      "loss": 5.8824,
      "theoretical_loss": 7.088747657242693,
      "tokens_seen": 8716288
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00013293650793650793,
      "loss": 6.1145,
      "theoretical_loss": 7.078942483428749,
      "tokens_seen": 8781824
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00013392857142857144,
      "loss": 6.1675,
      "theoretical_loss": 7.069230526471966,
      "tokens_seen": 8847360
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0001349206349206349,
      "loss": 6.1415,
      "theoretical_loss": 7.059610219454568,
      "tokens_seen": 8912896
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0001359126984126984,
      "loss": 6.1579,
      "theoretical_loss": 7.0500800330726685,
      "tokens_seen": 8978432
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0001369047619047619,
      "loss": 6.2736,
      "theoretical_loss": 7.040638474466625,
      "tokens_seen": 9043968
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00013789682539682541,
      "loss": 6.0392,
      "theoretical_loss": 7.031284086095933,
      "tokens_seen": 9109504
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0001388888888888889,
      "loss": 6.2167,
      "theoretical_loss": 7.022015444656678,
      "tokens_seen": 9175040
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0001398809523809524,
      "loss": 5.9301,
      "theoretical_loss": 7.012831160039609,
      "tokens_seen": 9240576
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00014087301587301586,
      "loss": 5.9091,
      "theoretical_loss": 7.003729874327071,
      "tokens_seen": 9306112
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00014186507936507937,
      "loss": 6.1878,
      "theoretical_loss": 6.994710260827057,
      "tokens_seen": 9371648
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00014285714285714284,
      "loss": 5.9379,
      "theoretical_loss": 6.98577102314278,
      "tokens_seen": 9437184
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00014384920634920634,
      "loss": 5.8658,
      "theoretical_loss": 6.976910894276189,
      "tokens_seen": 9502720
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00014484126984126984,
      "loss": 5.9478,
      "theoretical_loss": 6.968128635764015,
      "tokens_seen": 9568256
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00014583333333333335,
      "loss": 6.1221,
      "theoretical_loss": 6.959423036844894,
      "tokens_seen": 9633792
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00014682539682539685,
      "loss": 5.9375,
      "theoretical_loss": 6.950792913656309,
      "tokens_seen": 9699328
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00014781746031746032,
      "loss": 5.9392,
      "theoretical_loss": 6.942237108460029,
      "tokens_seen": 9764864
    },
    {
      "epoch": 0.0,
      "objective/train/docs_used": 92738,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 5.9081549644470215,
      "objective/train/theoretical_loss": 6.9337544888949,
      "objective/train/tokens_used": 30290400,
      "theoretical_loss": 6.9337544888949,
      "tokens_seen": 9830400
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00014880952380952382,
      "loss": 5.9356,
      "theoretical_loss": 6.9337544888949,
      "tokens_seen": 9830400
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0001498015873015873,
      "loss": 6.1536,
      "theoretical_loss": 6.925343947255817,
      "tokens_seen": 9895936
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0001507936507936508,
      "loss": 5.9417,
      "theoretical_loss": 6.917004399797798,
      "tokens_seen": 9961472
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00015178571428571427,
      "loss": 6.0151,
      "theoretical_loss": 6.908734786064147,
      "tokens_seen": 10027008
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0001527777777777778,
      "loss": 6.0507,
      "theoretical_loss": 6.900534068237688,
      "tokens_seen": 10092544
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00015376984126984128,
      "loss": 5.9881,
      "theoretical_loss": 6.89240123051416,
      "tokens_seen": 10158080
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00015476190476190478,
      "loss": 6.0754,
      "theoretical_loss": 6.884335278496871,
      "tokens_seen": 10223616
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00015575396825396825,
      "loss": 5.7266,
      "theoretical_loss": 6.87633523861175,
      "tokens_seen": 10289152
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00015674603174603175,
      "loss": 5.9508,
      "theoretical_loss": 6.868400157541997,
      "tokens_seen": 10354688
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00015773809523809523,
      "loss": 5.7484,
      "theoretical_loss": 6.860529101681551,
      "tokens_seen": 10420224
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00015873015873015873,
      "loss": 5.9385,
      "theoretical_loss": 6.85272115660663,
      "tokens_seen": 10485760
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0001597222222222222,
      "loss": 5.9719,
      "theoretical_loss": 6.844975426564642,
      "tokens_seen": 10551296
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00016071428571428573,
      "loss": 5.9549,
      "theoretical_loss": 6.8372910339797945,
      "tokens_seen": 10616832
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0001617063492063492,
      "loss": 6.0978,
      "theoretical_loss": 6.829667118974749,
      "tokens_seen": 10682368
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0001626984126984127,
      "loss": 5.7867,
      "theoretical_loss": 6.8221028389077185,
      "tokens_seen": 10747904
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00016369047619047618,
      "loss": 5.8499,
      "theoretical_loss": 6.814597367924395,
      "tokens_seen": 10813440
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00016468253968253969,
      "loss": 5.9166,
      "theoretical_loss": 6.807149896524181,
      "tokens_seen": 10878976
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00016567460317460316,
      "loss": 5.7915,
      "theoretical_loss": 6.799759631140145,
      "tokens_seen": 10944512
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00016666666666666666,
      "loss": 5.7431,
      "theoretical_loss": 6.7924257937322245,
      "tokens_seen": 11010048
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00016765873015873016,
      "loss": 5.8272,
      "theoretical_loss": 6.785147621393148,
      "tokens_seen": 11075584
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00016865079365079366,
      "loss": 5.6441,
      "theoretical_loss": 6.777924365966638,
      "tokens_seen": 11141120
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00016964285714285717,
      "loss": 5.733,
      "theoretical_loss": 6.770755293677423,
      "tokens_seen": 11206656
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00017063492063492064,
      "loss": 5.8183,
      "theoretical_loss": 6.763639684772625,
      "tokens_seen": 11272192
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00017162698412698414,
      "loss": 5.7902,
      "theoretical_loss": 6.756576833174123,
      "tokens_seen": 11337728
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00017261904761904762,
      "loss": 5.7469,
      "theoretical_loss": 6.749566046141486,
      "tokens_seen": 11403264
    },
    {
      "epoch": 0.0,
      "objective/train/docs_used": 97786,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 6.325038433074951,
      "objective/train/theoretical_loss": 6.7426066439450905,
      "objective/train/tokens_used": 31928800,
      "theoretical_loss": 6.7426066439450905,
      "tokens_seen": 11468800
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00017361111111111112,
      "loss": 5.8377,
      "theoretical_loss": 6.7426066439450905,
      "tokens_seen": 11468800
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0001746031746031746,
      "loss": 6.0515,
      "theoretical_loss": 6.735697959549075,
      "tokens_seen": 11534336
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0001755952380952381,
      "loss": 5.8069,
      "theoretical_loss": 6.728839338303761,
      "tokens_seen": 11599872
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0001765873015873016,
      "loss": 5.6908,
      "theoretical_loss": 6.722030137647226,
      "tokens_seen": 11665408
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0001775793650793651,
      "loss": 5.5015,
      "theoretical_loss": 6.715269726815689,
      "tokens_seen": 11730944
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00017857142857142857,
      "loss": 5.6095,
      "theoretical_loss": 6.7085574865624125,
      "tokens_seen": 11796480
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00017956349206349207,
      "loss": 5.588,
      "theoretical_loss": 6.701892808884824,
      "tokens_seen": 11862016
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00018055555555555555,
      "loss": 5.7956,
      "theoretical_loss": 6.695275096759559,
      "tokens_seen": 11927552
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00018154761904761905,
      "loss": 5.759,
      "theoretical_loss": 6.68870376388518,
      "tokens_seen": 11993088
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00018253968253968252,
      "loss": 5.7904,
      "theoretical_loss": 6.682178234432274,
      "tokens_seen": 12058624
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00018353174603174602,
      "loss": 5.8681,
      "theoretical_loss": 6.675697942800715,
      "tokens_seen": 12124160
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00018452380952380953,
      "loss": 5.6942,
      "theoretical_loss": 6.669262333383815,
      "tokens_seen": 12189696
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00018551587301587303,
      "loss": 5.612,
      "theoretical_loss": 6.662870860339158,
      "tokens_seen": 12255232
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0001865079365079365,
      "loss": 5.8235,
      "theoretical_loss": 6.656522987365879,
      "tokens_seen": 12320768
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0001875,
      "loss": 5.7374,
      "theoretical_loss": 6.6502181874881705,
      "tokens_seen": 12386304
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0001884920634920635,
      "loss": 5.624,
      "theoretical_loss": 6.643955942844831,
      "tokens_seen": 12451840
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00018948412698412698,
      "loss": 5.7172,
      "theoretical_loss": 6.637735744484626,
      "tokens_seen": 12517376
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00019047619047619048,
      "loss": 5.6884,
      "theoretical_loss": 6.631557092167304,
      "tokens_seen": 12582912
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00019146825396825398,
      "loss": 5.5524,
      "theoretical_loss": 6.625419494170049,
      "tokens_seen": 12648448
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00019246031746031748,
      "loss": 5.9368,
      "theoretical_loss": 6.619322467099223,
      "tokens_seen": 12713984
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00019345238095238096,
      "loss": 5.6623,
      "theoretical_loss": 6.613265535707211,
      "tokens_seen": 12779520
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00019444444444444446,
      "loss": 5.4595,
      "theoretical_loss": 6.607248232714213,
      "tokens_seen": 12845056
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00019543650793650793,
      "loss": 5.2926,
      "theoretical_loss": 6.60127009863481,
      "tokens_seen": 12910592
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00019642857142857144,
      "loss": 5.5284,
      "theoretical_loss": 6.59533068160918,
      "tokens_seen": 12976128
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0001974206349206349,
      "loss": 5.7083,
      "theoretical_loss": 6.589429537238785,
      "tokens_seen": 13041664
    },
    {
      "epoch": 0.0,
      "objective/train/docs_used": 102516,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 5.887484073638916,
      "objective/train/theoretical_loss": 6.583566228426414,
      "objective/train/tokens_used": 33567200,
      "theoretical_loss": 6.583566228426414,
      "tokens_seen": 13107200
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0001984126984126984,
      "loss": 5.7401,
      "theoretical_loss": 6.583566228426414,
      "tokens_seen": 13107200
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00019940476190476191,
      "loss": 5.6309,
      "theoretical_loss": 6.5777403252204305,
      "tokens_seen": 13172736
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00020039682539682542,
      "loss": 5.5628,
      "theoretical_loss": 6.571951404663098,
      "tokens_seen": 13238272
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0002013888888888889,
      "loss": 5.7721,
      "theoretical_loss": 6.566199050642863,
      "tokens_seen": 13303808
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0002023809523809524,
      "loss": 5.3532,
      "theoretical_loss": 6.560482853750463,
      "tokens_seen": 13369344
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00020337301587301587,
      "loss": 5.6038,
      "theoretical_loss": 6.554802411138745,
      "tokens_seen": 13434880
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00020436507936507937,
      "loss": 5.8779,
      "theoretical_loss": 6.549157326386091,
      "tokens_seen": 13500416
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00020535714285714284,
      "loss": 5.8416,
      "theoretical_loss": 6.54354720936333,
      "tokens_seen": 13565952
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00020634920634920634,
      "loss": 5.3702,
      "theoretical_loss": 6.537971676104026,
      "tokens_seen": 13631488
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00020734126984126985,
      "loss": 5.8494,
      "theoretical_loss": 6.532430348678068,
      "tokens_seen": 13697024
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00020833333333333335,
      "loss": 5.6039,
      "theoretical_loss": 6.5269228550684195,
      "tokens_seen": 13762560
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00020932539682539685,
      "loss": 5.5831,
      "theoretical_loss": 6.521448829050978,
      "tokens_seen": 13828096
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00021031746031746032,
      "loss": 5.5982,
      "theoretical_loss": 6.516007910077416,
      "tokens_seen": 13893632
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00021130952380952382,
      "loss": 5.4289,
      "theoretical_loss": 6.51059974316095,
      "tokens_seen": 13959168
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0002123015873015873,
      "loss": 5.5763,
      "theoretical_loss": 6.50522397876491,
      "tokens_seen": 14024704
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0002132936507936508,
      "loss": 5.5622,
      "theoretical_loss": 6.499880272694068,
      "tokens_seen": 14090240
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00021428571428571427,
      "loss": 5.6438,
      "theoretical_loss": 6.494568285988618,
      "tokens_seen": 14155776
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0002152777777777778,
      "loss": 5.662,
      "theoretical_loss": 6.489287684820745,
      "tokens_seen": 14221312
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00021626984126984128,
      "loss": 5.3087,
      "theoretical_loss": 6.484038140393699,
      "tokens_seen": 14286848
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00021726190476190478,
      "loss": 5.3336,
      "theoretical_loss": 6.4788193288433105,
      "tokens_seen": 14352384
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00021825396825396825,
      "loss": 5.4532,
      "theoretical_loss": 6.473630931141869,
      "tokens_seen": 14417920
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00021924603174603176,
      "loss": 5.6727,
      "theoretical_loss": 6.468472633004308,
      "tokens_seen": 14483456
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00022023809523809523,
      "loss": 5.3546,
      "theoretical_loss": 6.463344124796616,
      "tokens_seen": 14548992
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00022123015873015873,
      "loss": 5.469,
      "theoretical_loss": 6.45824510144643,
      "tokens_seen": 14614528
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0002222222222222222,
      "loss": 5.7059,
      "theoretical_loss": 6.45317526235573,
      "tokens_seen": 14680064
    },
    {
      "epoch": 0.0,
      "objective/train/docs_used": 107633,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 5.52135705947876,
      "objective/train/theoretical_loss": 6.448134311315593,
      "objective/train/tokens_used": 35205600,
      "theoretical_loss": 6.448134311315593,
      "tokens_seen": 14745600
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00022321428571428573,
      "loss": 5.493,
      "theoretical_loss": 6.448134311315593,
      "tokens_seen": 14745600
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0002242063492063492,
      "loss": 5.5841,
      "theoretical_loss": 6.443121956422939,
      "tokens_seen": 14811136
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0002251984126984127,
      "loss": 5.329,
      "theoretical_loss": 6.438137909999214,
      "tokens_seen": 14876672
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00022619047619047618,
      "loss": 5.5932,
      "theoretical_loss": 6.433181888510964,
      "tokens_seen": 14942208
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00022718253968253969,
      "loss": 5.3925,
      "theoretical_loss": 6.428253612492239,
      "tokens_seen": 15007744
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00022817460317460316,
      "loss": 5.3742,
      "theoretical_loss": 6.4233528064687855,
      "tokens_seen": 15073280
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00022916666666666666,
      "loss": 5.6301,
      "theoretical_loss": 6.418479198883969,
      "tokens_seen": 15138816
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00023015873015873016,
      "loss": 5.0119,
      "theoretical_loss": 6.413632522026391,
      "tokens_seen": 15204352
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00023115079365079367,
      "loss": 5.4315,
      "theoretical_loss": 6.40881251195914,
      "tokens_seen": 15269888
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00023214285714285717,
      "loss": 5.401,
      "theoretical_loss": 6.404018908450656,
      "tokens_seen": 15335424
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00023313492063492064,
      "loss": 5.2261,
      "theoretical_loss": 6.399251454907132,
      "tokens_seen": 15400960
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00023412698412698414,
      "loss": 5.44,
      "theoretical_loss": 6.394509898306452,
      "tokens_seen": 15466496
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00023511904761904762,
      "loss": 5.4848,
      "theoretical_loss": 6.389793989133574,
      "tokens_seen": 15532032
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00023611111111111112,
      "loss": 5.5739,
      "theoretical_loss": 6.385103481317387,
      "tokens_seen": 15597568
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0002371031746031746,
      "loss": 5.3712,
      "theoretical_loss": 6.380438132168923,
      "tokens_seen": 15663104
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0002380952380952381,
      "loss": 5.3527,
      "theoretical_loss": 6.375797702320966,
      "tokens_seen": 15728640
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0002390873015873016,
      "loss": 5.4,
      "theoretical_loss": 6.371181955668966,
      "tokens_seen": 15794176
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0002400793650793651,
      "loss": 5.3654,
      "theoretical_loss": 6.366590659313248,
      "tokens_seen": 15859712
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00024107142857142857,
      "loss": 5.4837,
      "theoretical_loss": 6.36202358350248,
      "tokens_seen": 15925248
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00024206349206349207,
      "loss": 5.2794,
      "theoretical_loss": 6.357480501578371,
      "tokens_seen": 15990784
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00024305555555555555,
      "loss": 5.2403,
      "theoretical_loss": 6.352961189921553,
      "tokens_seen": 16056320
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00024404761904761905,
      "loss": 5.3106,
      "theoretical_loss": 6.348465427898629,
      "tokens_seen": 16121856
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00024503968253968255,
      "loss": 5.3799,
      "theoretical_loss": 6.343992997810366,
      "tokens_seen": 16187392
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.000246031746031746,
      "loss": 5.3256,
      "theoretical_loss": 6.33954368484097,
      "tokens_seen": 16252928
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00024702380952380955,
      "loss": 5.5652,
      "theoretical_loss": 6.33511727700846,
      "tokens_seen": 16318464
    },
    {
      "epoch": 0.0,
      "objective/train/docs_used": 112492,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 5.672967910766602,
      "objective/train/theoretical_loss": 6.330713565116083,
      "objective/train/tokens_used": 36844000,
      "theoretical_loss": 6.330713565116083,
      "tokens_seen": 16384000
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.000248015873015873,
      "loss": 5.5336,
      "theoretical_loss": 6.330713565116083,
      "tokens_seen": 16384000
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0002490079365079365,
      "loss": 5.1247,
      "theoretical_loss": 6.326332342704751,
      "tokens_seen": 16449536
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00025,
      "loss": 4.9896,
      "theoretical_loss": 6.32197340600647,
      "tokens_seen": 16515072
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002509920634920635,
      "loss": 5.323,
      "theoretical_loss": 6.3176365538987636,
      "tokens_seen": 16580608
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000251984126984127,
      "loss": 5.0875,
      "theoretical_loss": 6.313321587860021,
      "tokens_seen": 16646144
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00025297619047619046,
      "loss": 5.4551,
      "theoretical_loss": 6.309028311925785,
      "tokens_seen": 16711680
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00025396825396825396,
      "loss": 5.3549,
      "theoretical_loss": 6.304756532645939,
      "tokens_seen": 16777216
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00025496031746031746,
      "loss": 5.1,
      "theoretical_loss": 6.300506059042775,
      "tokens_seen": 16842752
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00025595238095238096,
      "loss": 5.4827,
      "theoretical_loss": 6.296276702569918,
      "tokens_seen": 16908288
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002569444444444444,
      "loss": 5.4362,
      "theoretical_loss": 6.292068277072099,
      "tokens_seen": 16973824
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00025793650793650796,
      "loss": 5.3388,
      "theoretical_loss": 6.28788059874573,
      "tokens_seen": 17039360
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00025892857142857146,
      "loss": 5.2424,
      "theoretical_loss": 6.283713486100297,
      "tokens_seen": 17104896
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00025992063492063497,
      "loss": 5.3369,
      "theoretical_loss": 6.279566759920507,
      "tokens_seen": 17170432
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002609126984126984,
      "loss": 5.3566,
      "theoretical_loss": 6.275440243229228,
      "tokens_seen": 17235968
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002619047619047619,
      "loss": 5.1989,
      "theoretical_loss": 6.271333761251142,
      "tokens_seen": 17301504
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002628968253968254,
      "loss": 5.2989,
      "theoretical_loss": 6.267247141377137,
      "tokens_seen": 17367040
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002638888888888889,
      "loss": 5.1155,
      "theoretical_loss": 6.2631802131294085,
      "tokens_seen": 17432576
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00026488095238095237,
      "loss": 5.0634,
      "theoretical_loss": 6.259132808127246,
      "tokens_seen": 17498112
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00026587301587301587,
      "loss": 5.2512,
      "theoretical_loss": 6.255104760053497,
      "tokens_seen": 17563648
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00026686507936507937,
      "loss": 5.2924,
      "theoretical_loss": 6.251095904621689,
      "tokens_seen": 17629184
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00026785714285714287,
      "loss": 5.1212,
      "theoretical_loss": 6.247106079543801,
      "tokens_seen": 17694720
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002688492063492063,
      "loss": 5.2689,
      "theoretical_loss": 6.243135124498652,
      "tokens_seen": 17760256
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002698412698412698,
      "loss": 5.2765,
      "theoretical_loss": 6.239182881100916,
      "tokens_seen": 17825792
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002708333333333333,
      "loss": 5.2352,
      "theoretical_loss": 6.235249192870732,
      "tokens_seen": 17891328
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002718253968253968,
      "loss": 5.5256,
      "theoretical_loss": 6.231333905203899,
      "tokens_seen": 17956864
    },
    {
      "epoch": 0.01,
      "objective/train/docs_used": 117713,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 5.518486022949219,
      "objective/train/theoretical_loss": 6.227436865342643,
      "objective/train/tokens_used": 38482400,
      "theoretical_loss": 6.227436865342643,
      "tokens_seen": 18022400
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002728174603174603,
      "loss": 5.1989,
      "theoretical_loss": 6.227436865342643,
      "tokens_seen": 18022400
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002738095238095238,
      "loss": 5.243,
      "theoretical_loss": 6.223557922346955,
      "tokens_seen": 18087936
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002748015873015873,
      "loss": 5.205,
      "theoretical_loss": 6.219696927066456,
      "tokens_seen": 18153472
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00027579365079365083,
      "loss": 5.068,
      "theoretical_loss": 6.215853732112821,
      "tokens_seen": 18219008
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00027678571428571433,
      "loss": 5.1643,
      "theoretical_loss": 6.212028191832702,
      "tokens_seen": 18284544
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002777777777777778,
      "loss": 5.1496,
      "theoretical_loss": 6.208220162281178,
      "tokens_seen": 18350080
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002787698412698413,
      "loss": 5.0765,
      "theoretical_loss": 6.204429501195701,
      "tokens_seen": 18415616
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002797619047619048,
      "loss": 4.9861,
      "theoretical_loss": 6.20065606797053,
      "tokens_seen": 18481152
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002807539682539683,
      "loss": 5.2004,
      "theoretical_loss": 6.19689972363164,
      "tokens_seen": 18546688
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00028174603174603173,
      "loss": 5.3659,
      "theoretical_loss": 6.1931603308120975,
      "tokens_seen": 18612224
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00028273809523809523,
      "loss": 5.1814,
      "theoretical_loss": 6.189437753727901,
      "tokens_seen": 18677760
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00028373015873015873,
      "loss": 5.1199,
      "theoretical_loss": 6.185731858154261,
      "tokens_seen": 18743296
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00028472222222222223,
      "loss": 5.2977,
      "theoretical_loss": 6.182042511402313,
      "tokens_seen": 18808832
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002857142857142857,
      "loss": 5.1583,
      "theoretical_loss": 6.17836958229627,
      "tokens_seen": 18874368
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002867063492063492,
      "loss": 5.0914,
      "theoretical_loss": 6.1747129411509825,
      "tokens_seen": 18939904
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002876984126984127,
      "loss": 4.8355,
      "theoretical_loss": 6.171072459749913,
      "tokens_seen": 19005440
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002886904761904762,
      "loss": 5.1587,
      "theoretical_loss": 6.1674480113235095,
      "tokens_seen": 19070976
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002896825396825397,
      "loss": 5.0374,
      "theoretical_loss": 6.163839470527964,
      "tokens_seen": 19136512
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002906746031746032,
      "loss": 5.3875,
      "theoretical_loss": 6.160246713424372,
      "tokens_seen": 19202048
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002916666666666667,
      "loss": 5.1875,
      "theoretical_loss": 6.156669617458243,
      "tokens_seen": 19267584
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002926587301587302,
      "loss": 5.3545,
      "theoretical_loss": 6.153108061439397,
      "tokens_seen": 19333120
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002936507936507937,
      "loss": 5.1569,
      "theoretical_loss": 6.149561925522211,
      "tokens_seen": 19398656
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00029464285714285714,
      "loss": 5.2313,
      "theoretical_loss": 6.146031091186222,
      "tokens_seen": 19464192
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00029563492063492064,
      "loss": 5.3309,
      "theoretical_loss": 6.142515441217064,
      "tokens_seen": 19529728
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00029662698412698414,
      "loss": 5.2886,
      "theoretical_loss": 6.1390148596877605,
      "tokens_seen": 19595264
    },
    {
      "epoch": 0.01,
      "objective/train/docs_used": 122717,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 5.544928073883057,
      "objective/train/theoretical_loss": 6.135529231940326,
      "objective/train/tokens_used": 40120800,
      "theoretical_loss": 6.135529231940326,
      "tokens_seen": 19660800
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00029761904761904765,
      "loss": 5.3891,
      "theoretical_loss": 6.135529231940326,
      "tokens_seen": 19660800
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002986111111111111,
      "loss": 5.1961,
      "theoretical_loss": 6.132058444567705,
      "tokens_seen": 19726336
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002996031746031746,
      "loss": 5.166,
      "theoretical_loss": 6.128602385396022,
      "tokens_seen": 19791872
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003005952380952381,
      "loss": 5.1897,
      "theoretical_loss": 6.125160943467138,
      "tokens_seen": 19857408
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003015873015873016,
      "loss": 5.1994,
      "theoretical_loss": 6.121734009021521,
      "tokens_seen": 19922944
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00030257936507936505,
      "loss": 5.0567,
      "theoretical_loss": 6.118321473481398,
      "tokens_seen": 19988480
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00030357142857142855,
      "loss": 5.1657,
      "theoretical_loss": 6.114923229434213,
      "tokens_seen": 20054016
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00030456349206349205,
      "loss": 5.1727,
      "theoretical_loss": 6.111539170616359,
      "tokens_seen": 20119552
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003055555555555556,
      "loss": 4.7328,
      "theoretical_loss": 6.108169191897195,
      "tokens_seen": 20185088
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00030654761904761905,
      "loss": 5.2345,
      "theoretical_loss": 6.104813189263336,
      "tokens_seen": 20250624
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00030753968253968255,
      "loss": 5.1383,
      "theoretical_loss": 6.101471059803204,
      "tokens_seen": 20316160
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00030853174603174605,
      "loss": 5.1325,
      "theoretical_loss": 6.098142701691856,
      "tokens_seen": 20381696
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00030952380952380956,
      "loss": 4.9321,
      "theoretical_loss": 6.094828014176053,
      "tokens_seen": 20447232
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000310515873015873,
      "loss": 5.3554,
      "theoretical_loss": 6.091526897559593,
      "tokens_seen": 20512768
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003115079365079365,
      "loss": 5.1831,
      "theoretical_loss": 6.088239253188885,
      "tokens_seen": 20578304
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003125,
      "loss": 5.2792,
      "theoretical_loss": 6.084964983438763,
      "tokens_seen": 20643840
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003134920634920635,
      "loss": 4.9628,
      "theoretical_loss": 6.0817039916985465,
      "tokens_seen": 20709376
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000314484126984127,
      "loss": 5.0786,
      "theoretical_loss": 6.078456182358325,
      "tokens_seen": 20774912
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00031547619047619046,
      "loss": 5.3443,
      "theoretical_loss": 6.075221460795472,
      "tokens_seen": 20840448
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00031646825396825396,
      "loss": 5.2776,
      "theoretical_loss": 6.071999733361386,
      "tokens_seen": 20905984
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00031746031746031746,
      "loss": 4.7741,
      "theoretical_loss": 6.068790907368448,
      "tokens_seen": 20971520
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00031845238095238096,
      "loss": 5.0769,
      "theoretical_loss": 6.0655948910771915,
      "tokens_seen": 21037056
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003194444444444444,
      "loss": 5.2176,
      "theoretical_loss": 6.062411593683687,
      "tokens_seen": 21102592
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00032043650793650796,
      "loss": 5.2326,
      "theoretical_loss": 6.059240925307134,
      "tokens_seen": 21168128
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00032142857142857147,
      "loss": 5.2254,
      "theoretical_loss": 6.056082796977648,
      "tokens_seen": 21233664
    },
    {
      "epoch": 0.01,
      "objective/train/docs_used": 127954,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 4.949924468994141,
      "objective/train/theoretical_loss": 6.052937120624258,
      "objective/train/tokens_used": 41759200,
      "theoretical_loss": 6.052937120624258,
      "tokens_seen": 21299200
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00032242063492063497,
      "loss": 5.1471,
      "theoretical_loss": 6.052937120624258,
      "tokens_seen": 21299200
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003234126984126984,
      "loss": 4.9762,
      "theoretical_loss": 6.049803809063083,
      "tokens_seen": 21364736
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003244047619047619,
      "loss": 4.7374,
      "theoretical_loss": 6.0466827759857145,
      "tokens_seen": 21430272
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003253968253968254,
      "loss": 4.8236,
      "theoretical_loss": 6.04357393594778,
      "tokens_seen": 21495808
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003263888888888889,
      "loss": 5.0896,
      "theoretical_loss": 6.040477204357686,
      "tokens_seen": 21561344
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00032738095238095237,
      "loss": 4.9351,
      "theoretical_loss": 6.037392497465552,
      "tokens_seen": 21626880
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00032837301587301587,
      "loss": 4.9554,
      "theoretical_loss": 6.034319732352309,
      "tokens_seen": 21692416
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00032936507936507937,
      "loss": 5.2635,
      "theoretical_loss": 6.031258826918979,
      "tokens_seen": 21757952
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00033035714285714287,
      "loss": 5.0873,
      "theoretical_loss": 6.0282096998761245,
      "tokens_seen": 21823488
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003313492063492063,
      "loss": 5.1317,
      "theoretical_loss": 6.025172270733464,
      "tokens_seen": 21889024
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003323412698412698,
      "loss": 4.7562,
      "theoretical_loss": 6.0221464597896475,
      "tokens_seen": 21954560
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003333333333333333,
      "loss": 5.0833,
      "theoretical_loss": 6.0191321881221995,
      "tokens_seen": 22020096
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003343253968253968,
      "loss": 4.9952,
      "theoretical_loss": 6.016129377577614,
      "tokens_seen": 22085632
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003353174603174603,
      "loss": 5.141,
      "theoretical_loss": 6.01313795076161,
      "tokens_seen": 22151168
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003363095238095238,
      "loss": 4.8368,
      "theoretical_loss": 6.010157831029533,
      "tokens_seen": 22216704
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00033730158730158733,
      "loss": 5.0532,
      "theoretical_loss": 6.007188942476907,
      "tokens_seen": 22282240
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00033829365079365083,
      "loss": 5.0954,
      "theoretical_loss": 6.0042312099301425,
      "tokens_seen": 22347776
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00033928571428571433,
      "loss": 5.0574,
      "theoretical_loss": 6.001284558937368,
      "tokens_seen": 22413312
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003402777777777778,
      "loss": 4.8983,
      "theoretical_loss": 5.998348915759426,
      "tokens_seen": 22478848
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003412698412698413,
      "loss": 4.9248,
      "theoretical_loss": 5.995424207360987,
      "tokens_seen": 22544384
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003422619047619048,
      "loss": 4.8788,
      "theoretical_loss": 5.992510361401818,
      "tokens_seen": 22609920
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003432539682539683,
      "loss": 5.0268,
      "theoretical_loss": 5.989607306228168,
      "tokens_seen": 22675456
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00034424603174603173,
      "loss": 4.9311,
      "theoretical_loss": 5.986714970864292,
      "tokens_seen": 22740992
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00034523809523809523,
      "loss": 4.9288,
      "theoretical_loss": 5.983833285004112,
      "tokens_seen": 22806528
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00034623015873015873,
      "loss": 5.1659,
      "theoretical_loss": 5.980962179002983,
      "tokens_seen": 22872064
    },
    {
      "epoch": 0.01,
      "objective/train/docs_used": 130738,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 5.168276786804199,
      "objective/train/theoretical_loss": 5.978101583869607,
      "objective/train/tokens_used": 43397600,
      "theoretical_loss": 5.978101583869607,
      "tokens_seen": 22937600
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00034722222222222224,
      "loss": 4.6862,
      "theoretical_loss": 5.978101583869607,
      "tokens_seen": 22937600
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003482142857142857,
      "loss": 4.9681,
      "theoretical_loss": 5.975251431258057,
      "tokens_seen": 23003136
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003492063492063492,
      "loss": 5.004,
      "theoretical_loss": 5.972411653459913,
      "tokens_seen": 23068672
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003501984126984127,
      "loss": 5.0517,
      "theoretical_loss": 5.9695821833965335,
      "tokens_seen": 23134208
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003511904761904762,
      "loss": 5.2897,
      "theoretical_loss": 5.966762954611432,
      "tokens_seen": 23199744
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003521825396825397,
      "loss": 5.0365,
      "theoretical_loss": 5.963953901262764,
      "tokens_seen": 23265280
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003531746031746032,
      "loss": 4.9465,
      "theoretical_loss": 5.961154958115937,
      "tokens_seen": 23330816
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003541666666666667,
      "loss": 5.1461,
      "theoretical_loss": 5.958366060536315,
      "tokens_seen": 23396352
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003551587301587302,
      "loss": 4.8917,
      "theoretical_loss": 5.955587144482044,
      "tokens_seen": 23461888
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003561507936507937,
      "loss": 4.9521,
      "theoretical_loss": 5.952818146496978,
      "tokens_seen": 23527424
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00035714285714285714,
      "loss": 4.8648,
      "theoretical_loss": 5.950059003703704,
      "tokens_seen": 23592960
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00035813492063492064,
      "loss": 5.081,
      "theoretical_loss": 5.94730965379668,
      "tokens_seen": 23658496
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00035912698412698415,
      "loss": 4.9531,
      "theoretical_loss": 5.944570035035458,
      "tokens_seen": 23724032
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00036011904761904765,
      "loss": 4.8974,
      "theoretical_loss": 5.941840086238027,
      "tokens_seen": 23789568
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003611111111111111,
      "loss": 4.9338,
      "theoretical_loss": 5.939119746774228,
      "tokens_seen": 23855104
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003621031746031746,
      "loss": 4.9457,
      "theoretical_loss": 5.936408956559284,
      "tokens_seen": 23920640
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003630952380952381,
      "loss": 4.8877,
      "theoretical_loss": 5.933707656047414,
      "tokens_seen": 23986176
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003640873015873016,
      "loss": 5.2136,
      "theoretical_loss": 5.93101578622554,
      "tokens_seen": 24051712
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00036507936507936505,
      "loss": 5.1442,
      "theoretical_loss": 5.928333288607086,
      "tokens_seen": 24117248
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00036607142857142855,
      "loss": 4.939,
      "theoretical_loss": 5.925660105225867,
      "tokens_seen": 24182784
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00036706349206349205,
      "loss": 4.9773,
      "theoretical_loss": 5.92299617863006,
      "tokens_seen": 24248320
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003680555555555556,
      "loss": 4.8856,
      "theoretical_loss": 5.920341451876267,
      "tokens_seen": 24313856
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00036904761904761905,
      "loss": 4.8102,
      "theoretical_loss": 5.9176958685236585,
      "tokens_seen": 24379392
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00037003968253968255,
      "loss": 4.9364,
      "theoretical_loss": 5.9150593726282015,
      "tokens_seen": 24444928
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00037103174603174606,
      "loss": 4.7276,
      "theoretical_loss": 5.912431908736972,
      "tokens_seen": 24510464
    },
    {
      "epoch": 0.01,
      "objective/train/docs_used": 131522,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 5.190422534942627,
      "objective/train/theoretical_loss": 5.909813421882534,
      "objective/train/tokens_used": 45036000,
      "theoretical_loss": 5.909813421882534,
      "tokens_seen": 24576000
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00037202380952380956,
      "loss": 4.8887,
      "theoretical_loss": 5.909813421882534,
      "tokens_seen": 24576000
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000373015873015873,
      "loss": 4.9041,
      "theoretical_loss": 5.907203857577422,
      "tokens_seen": 24641536
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003740079365079365,
      "loss": 4.8034,
      "theoretical_loss": 5.9046031618086765,
      "tokens_seen": 24707072
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000375,
      "loss": 5.0023,
      "theoretical_loss": 5.902011281032472,
      "tokens_seen": 24772608
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003759920634920635,
      "loss": 5.0488,
      "theoretical_loss": 5.899428162168808,
      "tokens_seen": 24838144
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000376984126984127,
      "loss": 5.0453,
      "theoretical_loss": 5.896853752596286,
      "tokens_seen": 24903680
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00037797619047619046,
      "loss": 4.813,
      "theoretical_loss": 5.894288000146949,
      "tokens_seen": 24969216
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00037896825396825396,
      "loss": 4.8831,
      "theoretical_loss": 5.891730853101199,
      "tokens_seen": 25034752
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00037996031746031746,
      "loss": 5.263,
      "theoretical_loss": 5.88918226018278,
      "tokens_seen": 25100288
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00038095238095238096,
      "loss": 5.1195,
      "theoretical_loss": 5.8866421705538325,
      "tokens_seen": 25165824
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003819444444444444,
      "loss": 4.9757,
      "theoretical_loss": 5.8841105338100155,
      "tokens_seen": 25231360
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00038293650793650797,
      "loss": 5.1901,
      "theoretical_loss": 5.881587299975694,
      "tokens_seen": 25296896
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00038392857142857147,
      "loss": 4.8656,
      "theoretical_loss": 5.8790724194991935,
      "tokens_seen": 25362432
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00038492063492063497,
      "loss": 5.0726,
      "theoretical_loss": 5.876565843248124,
      "tokens_seen": 25427968
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003859126984126984,
      "loss": 5.1417,
      "theoretical_loss": 5.8740675225047525,
      "tokens_seen": 25493504
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003869047619047619,
      "loss": 5.1355,
      "theoretical_loss": 5.871577408961457,
      "tokens_seen": 25559040
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003878968253968254,
      "loss": 5.0327,
      "theoretical_loss": 5.869095454716231,
      "tokens_seen": 25624576
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003888888888888889,
      "loss": 5.139,
      "theoretical_loss": 5.866621612268246,
      "tokens_seen": 25690112
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00038988095238095237,
      "loss": 4.9834,
      "theoretical_loss": 5.864155834513486,
      "tokens_seen": 25755648
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00039087301587301587,
      "loss": 5.187,
      "theoretical_loss": 5.8616980747404295,
      "tokens_seen": 25821184
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00039186507936507937,
      "loss": 5.0414,
      "theoretical_loss": 5.859248286625787,
      "tokens_seen": 25886720
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003928571428571429,
      "loss": 5.0615,
      "theoretical_loss": 5.856806424230314,
      "tokens_seen": 25952256
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003938492063492063,
      "loss": 4.9987,
      "theoretical_loss": 5.854372441994654,
      "tokens_seen": 26017792
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003948412698412698,
      "loss": 5.0551,
      "theoretical_loss": 5.851946294735258,
      "tokens_seen": 26083328
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003958333333333333,
      "loss": 4.8932,
      "theoretical_loss": 5.849527937640345,
      "tokens_seen": 26148864
    },
    {
      "epoch": 0.01,
      "objective/train/docs_used": 132646,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 4.532196998596191,
      "objective/train/theoretical_loss": 5.8471173262659235,
      "objective/train/tokens_used": 46674400,
      "theoretical_loss": 5.8471173262659235,
      "tokens_seen": 26214400
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003968253968253968,
      "loss": 4.9406,
      "theoretical_loss": 5.8471173262659235,
      "tokens_seen": 26214400
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003978174603174603,
      "loss": 5.0843,
      "theoretical_loss": 5.84471441653186,
      "tokens_seen": 26279936
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00039880952380952383,
      "loss": 5.0651,
      "theoretical_loss": 5.842319164718004,
      "tokens_seen": 26345472
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00039980158730158733,
      "loss": 5.2335,
      "theoretical_loss": 5.83993152746036,
      "tokens_seen": 26411008
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00040079365079365083,
      "loss": 4.9606,
      "theoretical_loss": 5.83755146174731,
      "tokens_seen": 26476544
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00040178571428571433,
      "loss": 4.8502,
      "theoretical_loss": 5.835178924915889,
      "tokens_seen": 26542080
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004027777777777778,
      "loss": 5.1473,
      "theoretical_loss": 5.832813874648102,
      "tokens_seen": 26607616
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004037698412698413,
      "loss": 4.9566,
      "theoretical_loss": 5.8304562689673,
      "tokens_seen": 26673152
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004047619047619048,
      "loss": 4.7435,
      "theoretical_loss": 5.828106066234588,
      "tokens_seen": 26738688
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004057539682539683,
      "loss": 5.1525,
      "theoretical_loss": 5.825763225145295,
      "tokens_seen": 26804224
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00040674603174603173,
      "loss": 5.0371,
      "theoretical_loss": 5.823427704725473,
      "tokens_seen": 26869760
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00040773809523809523,
      "loss": 5.0707,
      "theoretical_loss": 5.82109946432846,
      "tokens_seen": 26935296
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00040873015873015874,
      "loss": 4.9852,
      "theoretical_loss": 5.818778463631473,
      "tokens_seen": 27000832
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00040972222222222224,
      "loss": 5.0129,
      "theoretical_loss": 5.816464662632243,
      "tokens_seen": 27066368
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004107142857142857,
      "loss": 4.8632,
      "theoretical_loss": 5.8141580216457065,
      "tokens_seen": 27131904
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004117063492063492,
      "loss": 5.2237,
      "theoretical_loss": 5.811858501300729,
      "tokens_seen": 27197440
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004126984126984127,
      "loss": 4.7749,
      "theoretical_loss": 5.809566062536868,
      "tokens_seen": 27262976
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004136904761904762,
      "loss": 4.9421,
      "theoretical_loss": 5.807280666601191,
      "tokens_seen": 27328512
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004146825396825397,
      "loss": 5.3685,
      "theoretical_loss": 5.805002275045111,
      "tokens_seen": 27394048
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004156746031746032,
      "loss": 4.9844,
      "theoretical_loss": 5.8027308497212875,
      "tokens_seen": 27459584
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004166666666666667,
      "loss": 5.0175,
      "theoretical_loss": 5.800466352780546,
      "tokens_seen": 27525120
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004176587301587302,
      "loss": 5.2404,
      "theoretical_loss": 5.798208746668847,
      "tokens_seen": 27590656
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004186507936507937,
      "loss": 4.9851,
      "theoretical_loss": 5.795957994124291,
      "tokens_seen": 27656192
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00041964285714285714,
      "loss": 5.0305,
      "theoretical_loss": 5.7937140581741575,
      "tokens_seen": 27721728
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00042063492063492065,
      "loss": 4.947,
      "theoretical_loss": 5.791476902131985,
      "tokens_seen": 27787264
    },
    {
      "epoch": 0.01,
      "objective/train/docs_used": 133227,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 4.931251525878906,
      "objective/train/theoretical_loss": 5.789246489594688,
      "objective/train/tokens_used": 48312800,
      "theoretical_loss": 5.789246489594688,
      "tokens_seen": 27852800
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00042162698412698415,
      "loss": 5.0547,
      "theoretical_loss": 5.789246489594688,
      "tokens_seen": 27852800
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00042261904761904765,
      "loss": 5.0383,
      "theoretical_loss": 5.787022784439701,
      "tokens_seen": 27918336
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004236111111111111,
      "loss": 5.0555,
      "theoretical_loss": 5.784805750822171,
      "tokens_seen": 27983872
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004246031746031746,
      "loss": 4.825,
      "theoretical_loss": 5.782595353172176,
      "tokens_seen": 28049408
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004255952380952381,
      "loss": 5.0085,
      "theoretical_loss": 5.780391556191977,
      "tokens_seen": 28114944
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004265873015873016,
      "loss": 4.7223,
      "theoretical_loss": 5.778194324853311,
      "tokens_seen": 28180480
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00042757936507936505,
      "loss": 4.9563,
      "theoretical_loss": 5.776003624394711,
      "tokens_seen": 28246016
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00042857142857142855,
      "loss": 4.8565,
      "theoretical_loss": 5.773819420318858,
      "tokens_seen": 28311552
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00042956349206349205,
      "loss": 4.653,
      "theoretical_loss": 5.771641678389971,
      "tokens_seen": 28377088
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004305555555555556,
      "loss": 4.9409,
      "theoretical_loss": 5.769470364631225,
      "tokens_seen": 28442624
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00043154761904761905,
      "loss": 5.1167,
      "theoretical_loss": 5.767305445322201,
      "tokens_seen": 28508160
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00043253968253968256,
      "loss": 4.933,
      "theoretical_loss": 5.765146886996363,
      "tokens_seen": 28573696
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00043353174603174606,
      "loss": 4.8283,
      "theoretical_loss": 5.762994656438579,
      "tokens_seen": 28639232
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00043452380952380956,
      "loss": 5.0357,
      "theoretical_loss": 5.760848720682651,
      "tokens_seen": 28704768
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000435515873015873,
      "loss": 5.0796,
      "theoretical_loss": 5.758709047008894,
      "tokens_seen": 28770304
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004365079365079365,
      "loss": 4.8349,
      "theoretical_loss": 5.756575602941732,
      "tokens_seen": 28835840
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004375,
      "loss": 5.0019,
      "theoretical_loss": 5.75444835624733,
      "tokens_seen": 28901376
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004384920634920635,
      "loss": 4.8212,
      "theoretical_loss": 5.752327274931249,
      "tokens_seen": 28966912
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000439484126984127,
      "loss": 4.9358,
      "theoretical_loss": 5.750212327236129,
      "tokens_seen": 29032448
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00044047619047619046,
      "loss": 4.9691,
      "theoretical_loss": 5.7481034816394105,
      "tokens_seen": 29097984
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00044146825396825396,
      "loss": 4.9269,
      "theoretical_loss": 5.7460007068510635,
      "tokens_seen": 29163520
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00044246031746031746,
      "loss": 5.0363,
      "theoretical_loss": 5.74390397181136,
      "tokens_seen": 29229056
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00044345238095238096,
      "loss": 4.6776,
      "theoretical_loss": 5.741813245688668,
      "tokens_seen": 29294592
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004444444444444444,
      "loss": 4.8552,
      "theoretical_loss": 5.739728497877267,
      "tokens_seen": 29360128
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00044543650793650797,
      "loss": 5.2274,
      "theoretical_loss": 5.737649697995197,
      "tokens_seen": 29425664
    },
    {
      "epoch": 0.01,
      "objective/train/docs_used": 134422,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 5.431413650512695,
      "objective/train/theoretical_loss": 5.7355768158821245,
      "objective/train/tokens_used": 49951200,
      "theoretical_loss": 5.7355768158821245,
      "tokens_seen": 29491200
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00044642857142857147,
      "loss": 5.112,
      "theoretical_loss": 5.7355768158821245,
      "tokens_seen": 29491200
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00044742063492063497,
      "loss": 5.1022,
      "theoretical_loss": 5.73350982159724,
      "tokens_seen": 29556736
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004484126984126984,
      "loss": 5.0531,
      "theoretical_loss": 5.731448685417178,
      "tokens_seen": 29622272
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004494047619047619,
      "loss": 4.8186,
      "theoretical_loss": 5.729393377833956,
      "tokens_seen": 29687808
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004503968253968254,
      "loss": 4.9572,
      "theoretical_loss": 5.7273438695529535,
      "tokens_seen": 29753344
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004513888888888889,
      "loss": 4.9607,
      "theoretical_loss": 5.725300131490888,
      "tokens_seen": 29818880
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00045238095238095237,
      "loss": 4.8473,
      "theoretical_loss": 5.7232621347738455,
      "tokens_seen": 29884416
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00045337301587301587,
      "loss": 5.0085,
      "theoretical_loss": 5.721229850735305,
      "tokens_seen": 29949952
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00045436507936507937,
      "loss": 4.938,
      "theoretical_loss": 5.719203250914208,
      "tokens_seen": 30015488
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004553571428571429,
      "loss": 4.8577,
      "theoretical_loss": 5.717182307053037,
      "tokens_seen": 30081024
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004563492063492063,
      "loss": 5.0517,
      "theoretical_loss": 5.715166991095922,
      "tokens_seen": 30146560
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004573412698412698,
      "loss": 4.8027,
      "theoretical_loss": 5.713157275186761,
      "tokens_seen": 30212096
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004583333333333333,
      "loss": 5.0271,
      "theoretical_loss": 5.71115313166738,
      "tokens_seen": 30277632
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004593253968253968,
      "loss": 5.0041,
      "theoretical_loss": 5.709154533075688,
      "tokens_seen": 30343168
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00046031746031746033,
      "loss": 4.9309,
      "theoretical_loss": 5.707161452143879,
      "tokens_seen": 30408704
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00046130952380952383,
      "loss": 4.9085,
      "theoretical_loss": 5.7051738617966326,
      "tokens_seen": 30474240
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00046230158730158733,
      "loss": 4.6103,
      "theoretical_loss": 5.7031917351493515,
      "tokens_seen": 30539776
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00046329365079365083,
      "loss": 5.0814,
      "theoretical_loss": 5.701215045506411,
      "tokens_seen": 30605312
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00046428571428571433,
      "loss": 5.0333,
      "theoretical_loss": 5.699243766359421,
      "tokens_seen": 30670848
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004652777777777778,
      "loss": 5.0393,
      "theoretical_loss": 5.697277871385534,
      "tokens_seen": 30736384
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004662698412698413,
      "loss": 4.813,
      "theoretical_loss": 5.695317334445736,
      "tokens_seen": 30801920
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004672619047619048,
      "loss": 4.8031,
      "theoretical_loss": 5.693362129583184,
      "tokens_seen": 30867456
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004682539682539683,
      "loss": 4.6799,
      "theoretical_loss": 5.691412231021549,
      "tokens_seen": 30932992
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00046924603174603173,
      "loss": 4.9719,
      "theoretical_loss": 5.689467613163388,
      "tokens_seen": 30998528
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00047023809523809523,
      "loss": 4.7362,
      "theoretical_loss": 5.687528250588518,
      "tokens_seen": 31064064
    },
    {
      "epoch": 0.01,
      "objective/train/docs_used": 135224,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 4.8863983154296875,
      "objective/train/theoretical_loss": 5.6855941180524265,
      "objective/train/tokens_used": 51589600,
      "theoretical_loss": 5.6855941180524265,
      "tokens_seen": 31129600
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00047123015873015874,
      "loss": 4.893,
      "theoretical_loss": 5.6855941180524265,
      "tokens_seen": 31129600
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00047222222222222224,
      "loss": 4.9991,
      "theoretical_loss": 5.683665190484683,
      "tokens_seen": 31195136
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004732142857142857,
      "loss": 5.1718,
      "theoretical_loss": 5.681741442987381,
      "tokens_seen": 31260672
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004742063492063492,
      "loss": 4.9702,
      "theoretical_loss": 5.679822850833591,
      "tokens_seen": 31326208
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004751984126984127,
      "loss": 4.9426,
      "theoretical_loss": 5.677909389465831,
      "tokens_seen": 31391744
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004761904761904762,
      "loss": 4.5549,
      "theoretical_loss": 5.676001034494554,
      "tokens_seen": 31457280
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004771825396825397,
      "loss": 4.8956,
      "theoretical_loss": 5.674097761696653,
      "tokens_seen": 31522816
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004781746031746032,
      "loss": 4.7593,
      "theoretical_loss": 5.672199547013983,
      "tokens_seen": 31588352
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004791666666666667,
      "loss": 4.7632,
      "theoretical_loss": 5.670306366551898,
      "tokens_seen": 31653888
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004801587301587302,
      "loss": 4.8197,
      "theoretical_loss": 5.6684181965778,
      "tokens_seen": 31719424
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004811507936507937,
      "loss": 5.0023,
      "theoretical_loss": 5.666535013519715,
      "tokens_seen": 31784960
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00048214285714285715,
      "loss": 4.265,
      "theoretical_loss": 5.6646567939648715,
      "tokens_seen": 31850496
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00048313492063492065,
      "loss": 4.6841,
      "theoretical_loss": 5.6627835146583045,
      "tokens_seen": 31916032
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00048412698412698415,
      "loss": 4.9755,
      "theoretical_loss": 5.660915152501465,
      "tokens_seen": 31981568
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00048511904761904765,
      "loss": 4.838,
      "theoretical_loss": 5.659051684550857,
      "tokens_seen": 32047104
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004861111111111111,
      "loss": 4.8988,
      "theoretical_loss": 5.657193088016677,
      "tokens_seen": 32112640
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004871031746031746,
      "loss": 4.8451,
      "theoretical_loss": 5.655339340261474,
      "tokens_seen": 32178176
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004880952380952381,
      "loss": 4.9273,
      "theoretical_loss": 5.653490418798825,
      "tokens_seen": 32243712
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004890873015873016,
      "loss": 4.6617,
      "theoretical_loss": 5.651646301292022,
      "tokens_seen": 32309248
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004900793650793651,
      "loss": 4.6989,
      "theoretical_loss": 5.649806965552774,
      "tokens_seen": 32374784
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004910714285714286,
      "loss": 4.6449,
      "theoretical_loss": 5.6479723895399205,
      "tokens_seen": 32440320
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000492063492063492,
      "loss": 4.9494,
      "theoretical_loss": 5.6461425513581665,
      "tokens_seen": 32505856
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004930555555555556,
      "loss": 4.6846,
      "theoretical_loss": 5.6443174292568195,
      "tokens_seen": 32571392
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004940476190476191,
      "loss": 4.6929,
      "theoretical_loss": 5.6424970016285485,
      "tokens_seen": 32636928
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004950396825396826,
      "loss": 4.6288,
      "theoretical_loss": 5.640681247008156,
      "tokens_seen": 32702464
    },
    {
      "epoch": 0.01,
      "objective/train/docs_used": 136821,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 4.2028489112854,
      "objective/train/theoretical_loss": 5.638870144071353,
      "objective/train/tokens_used": 53228000,
      "theoretical_loss": 5.638870144071353,
      "tokens_seen": 32768000
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000496031746031746,
      "loss": 4.6949,
      "theoretical_loss": 5.638870144071353,
      "tokens_seen": 32768000
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004970238095238095,
      "loss": 4.5009,
      "theoretical_loss": 5.637063671633564,
      "tokens_seen": 32833536
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000498015873015873,
      "loss": 4.7564,
      "theoretical_loss": 5.635261808648728,
      "tokens_seen": 32899072
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004990079365079365,
      "loss": 5.0523,
      "theoretical_loss": 5.6334645342081195,
      "tokens_seen": 32964608
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0005,
      "loss": 4.6206,
      "theoretical_loss": 5.631671827539186,
      "tokens_seen": 33030144
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004999899699097292,
      "loss": 4.7802,
      "theoretical_loss": 5.629883668004389,
      "tokens_seen": 33095680
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004999799398194584,
      "loss": 4.7891,
      "theoretical_loss": 5.628100035100061,
      "tokens_seen": 33161216
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004999699097291876,
      "loss": 4.8953,
      "theoretical_loss": 5.626320908455279,
      "tokens_seen": 33226752
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004999598796389167,
      "loss": 4.5953,
      "theoretical_loss": 5.6245462678307385,
      "tokens_seen": 33292288
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000499949849548646,
      "loss": 4.67,
      "theoretical_loss": 5.622776093117652,
      "tokens_seen": 33357824
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004999398194583751,
      "loss": 4.8824,
      "theoretical_loss": 5.621010364336651,
      "tokens_seen": 33423360
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004999297893681044,
      "loss": 4.7663,
      "theoretical_loss": 5.619249061636698,
      "tokens_seen": 33488896
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004999197592778335,
      "loss": 4.7263,
      "theoretical_loss": 5.61749216529402,
      "tokens_seen": 33554432
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004999097291875627,
      "loss": 4.9792,
      "theoretical_loss": 5.615739655711037,
      "tokens_seen": 33619968
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004998996990972919,
      "loss": 4.9381,
      "theoretical_loss": 5.61399151341532,
      "tokens_seen": 33685504
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004998896690070211,
      "loss": 4.9476,
      "theoretical_loss": 5.6122477190585425,
      "tokens_seen": 33751040
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004998796389167503,
      "loss": 4.5086,
      "theoretical_loss": 5.610508253415453,
      "tokens_seen": 33816576
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004998696088264795,
      "loss": 4.5371,
      "theoretical_loss": 5.6087730973828585,
      "tokens_seen": 33882112
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004998595787362087,
      "loss": 4.5089,
      "theoretical_loss": 5.6070422319786095,
      "tokens_seen": 33947648
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004998495486459378,
      "loss": 4.8347,
      "theoretical_loss": 5.605315638340606,
      "tokens_seen": 34013184
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000499839518555667,
      "loss": 4.7779,
      "theoretical_loss": 5.603593297725807,
      "tokens_seen": 34078720
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004998294884653962,
      "loss": 4.7224,
      "theoretical_loss": 5.601875191509249,
      "tokens_seen": 34144256
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004998194583751254,
      "loss": 4.8306,
      "theoretical_loss": 5.600161301183084,
      "tokens_seen": 34209792
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004998094282848546,
      "loss": 5.0491,
      "theoretical_loss": 5.598451608355614,
      "tokens_seen": 34275328
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004997993981945837,
      "loss": 4.6312,
      "theoretical_loss": 5.596746094750342,
      "tokens_seen": 34340864
    },
    {
      "epoch": 0.01,
      "objective/train/docs_used": 137555,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 4.4468183517456055,
      "objective/train/theoretical_loss": 5.595044742205037,
      "objective/train/tokens_used": 54866400,
      "theoretical_loss": 5.595044742205037,
      "tokens_seen": 34406400
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004997893681043129,
      "loss": 4.5163,
      "theoretical_loss": 5.595044742205037,
      "tokens_seen": 34406400
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004997793380140421,
      "loss": 4.9292,
      "theoretical_loss": 5.5933475326707995,
      "tokens_seen": 34471936
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004997693079237714,
      "loss": 4.8149,
      "theoretical_loss": 5.591654448211143,
      "tokens_seen": 34537472
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004997592778335005,
      "loss": 4.954,
      "theoretical_loss": 5.589965471001077,
      "tokens_seen": 34603008
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004997492477432298,
      "loss": 4.9477,
      "theoretical_loss": 5.5882805833262115,
      "tokens_seen": 34668544
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004997392176529588,
      "loss": 4.7965,
      "theoretical_loss": 5.586599767581859,
      "tokens_seen": 34734080
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004997291875626881,
      "loss": 5.0125,
      "theoretical_loss": 5.584923006272151,
      "tokens_seen": 34799616
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004997191574724173,
      "loss": 4.7564,
      "theoretical_loss": 5.583250282009159,
      "tokens_seen": 34865152
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004997091273821465,
      "loss": 4.7508,
      "theoretical_loss": 5.581581577512031,
      "tokens_seen": 34930688
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004996990972918757,
      "loss": 4.7624,
      "theoretical_loss": 5.579916875606134,
      "tokens_seen": 34996224
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004996890672016048,
      "loss": 4.6123,
      "theoretical_loss": 5.578256159222196,
      "tokens_seen": 35061760
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000499679037111334,
      "loss": 4.8336,
      "theoretical_loss": 5.576599411395472,
      "tokens_seen": 35127296
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004996690070210632,
      "loss": 4.9514,
      "theoretical_loss": 5.574946615264906,
      "tokens_seen": 35192832
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004996589769307924,
      "loss": 4.7192,
      "theoretical_loss": 5.5732977540723105,
      "tokens_seen": 35258368
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004996489468405216,
      "loss": 4.7461,
      "theoretical_loss": 5.571652811161542,
      "tokens_seen": 35323904
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004996389167502507,
      "loss": 4.6751,
      "theoretical_loss": 5.570011769977693,
      "tokens_seen": 35389440
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004996288866599799,
      "loss": 4.6333,
      "theoretical_loss": 5.568374614066299,
      "tokens_seen": 35454976
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004996188565697091,
      "loss": 4.6241,
      "theoretical_loss": 5.566741327072535,
      "tokens_seen": 35520512
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004996088264794383,
      "loss": 4.8779,
      "theoretical_loss": 5.565111892740433,
      "tokens_seen": 35586048
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004995987963891675,
      "loss": 4.4276,
      "theoretical_loss": 5.563486294912105,
      "tokens_seen": 35651584
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004995887662988968,
      "loss": 4.6749,
      "theoretical_loss": 5.56186451752697,
      "tokens_seen": 35717120
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004995787362086258,
      "loss": 4.665,
      "theoretical_loss": 5.560246544620993,
      "tokens_seen": 35782656
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004995687061183551,
      "loss": 4.7076,
      "theoretical_loss": 5.558632360325929,
      "tokens_seen": 35848192
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004995586760280842,
      "loss": 4.7311,
      "theoretical_loss": 5.557021948868571,
      "tokens_seen": 35913728
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004995486459378135,
      "loss": 4.6598,
      "theoretical_loss": 5.555415294570011,
      "tokens_seen": 35979264
    },
    {
      "epoch": 0.01,
      "objective/train/docs_used": 138886,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 4.407133102416992,
      "objective/train/theoretical_loss": 5.553812381844907,
      "objective/train/tokens_used": 56504800,
      "theoretical_loss": 5.553812381844907,
      "tokens_seen": 36044800
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004995386158475427,
      "loss": 4.5711,
      "theoretical_loss": 5.553812381844907,
      "tokens_seen": 36044800
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004995285857572718,
      "loss": 4.7261,
      "theoretical_loss": 5.552213195200755,
      "tokens_seen": 36110336
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000499518555667001,
      "loss": 4.7503,
      "theoretical_loss": 5.550617719237167,
      "tokens_seen": 36175872
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004995085255767302,
      "loss": 4.7123,
      "theoretical_loss": 5.549025938645155,
      "tokens_seen": 36241408
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004994984954864594,
      "loss": 4.8397,
      "theoretical_loss": 5.547437838206435,
      "tokens_seen": 36306944
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004994884653961886,
      "loss": 4.8057,
      "theoretical_loss": 5.545853402792717,
      "tokens_seen": 36372480
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004994784353059178,
      "loss": 4.711,
      "theoretical_loss": 5.544272617365014,
      "tokens_seen": 36438016
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004994684052156469,
      "loss": 4.758,
      "theoretical_loss": 5.542695466972956,
      "tokens_seen": 36503552
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004994583751253761,
      "loss": 4.9674,
      "theoretical_loss": 5.541121936754111,
      "tokens_seen": 36569088
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004994483450351053,
      "loss": 4.8971,
      "theoretical_loss": 5.539552011933312,
      "tokens_seen": 36634624
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004994383149448345,
      "loss": 4.6474,
      "theoretical_loss": 5.537985677821986,
      "tokens_seen": 36700160
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004994282848545637,
      "loss": 4.8243,
      "theoretical_loss": 5.536422919817495,
      "tokens_seen": 36765696
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004994182547642928,
      "loss": 4.7595,
      "theoretical_loss": 5.5348637234024824,
      "tokens_seen": 36831232
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004994082246740221,
      "loss": 4.8208,
      "theoretical_loss": 5.53330807414422,
      "tokens_seen": 36896768
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004993981945837512,
      "loss": 4.5145,
      "theoretical_loss": 5.5317559576939725,
      "tokens_seen": 36962304
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004993881644934805,
      "loss": 4.5815,
      "theoretical_loss": 5.530207359786353,
      "tokens_seen": 37027840
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004993781344032096,
      "loss": 4.7233,
      "theoretical_loss": 5.5286622662386975,
      "tokens_seen": 37093376
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004993681043129389,
      "loss": 4.7717,
      "theoretical_loss": 5.52712066295044,
      "tokens_seen": 37158912
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000499358074222668,
      "loss": 4.3869,
      "theoretical_loss": 5.525582535902489,
      "tokens_seen": 37224448
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004993480441323972,
      "loss": 4.5429,
      "theoretical_loss": 5.524047871156618,
      "tokens_seen": 37289984
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004993380140421264,
      "loss": 4.8437,
      "theoretical_loss": 5.52251665485486,
      "tokens_seen": 37355520
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004993279839518556,
      "loss": 4.7926,
      "theoretical_loss": 5.520988873218897,
      "tokens_seen": 37421056
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004993179538615848,
      "loss": 4.8248,
      "theoretical_loss": 5.519464512549478,
      "tokens_seen": 37486592
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004993079237713139,
      "loss": 4.7691,
      "theoretical_loss": 5.5179435592258095,
      "tokens_seen": 37552128
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004992978936810431,
      "loss": 4.6871,
      "theoretical_loss": 5.516425999704987,
      "tokens_seen": 37617664
    },
    {
      "epoch": 0.01,
      "objective/train/docs_used": 139553,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 4.058813571929932,
      "objective/train/theoretical_loss": 5.514911820521407,
      "objective/train/tokens_used": 58143200,
      "theoretical_loss": 5.514911820521407,
      "tokens_seen": 37683200
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004992878635907723,
      "loss": 4.5181,
      "theoretical_loss": 5.514911820521407,
      "tokens_seen": 37683200
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004992778335005015,
      "loss": 4.4909,
      "theoretical_loss": 5.5134010082861895,
      "tokens_seen": 37748736
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004992678034102307,
      "loss": 4.5355,
      "theoretical_loss": 5.511893549686616,
      "tokens_seen": 37814272
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004992577733199598,
      "loss": 4.7593,
      "theoretical_loss": 5.51038943148556,
      "tokens_seen": 37879808
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000499247743229689,
      "loss": 4.6539,
      "theoretical_loss": 5.508888640520928,
      "tokens_seen": 37945344
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004992377131394183,
      "loss": 4.5281,
      "theoretical_loss": 5.50739116370511,
      "tokens_seen": 38010880
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004992276830491475,
      "loss": 4.4719,
      "theoretical_loss": 5.505896988024423,
      "tokens_seen": 38076416
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004992176529588767,
      "loss": 4.8288,
      "theoretical_loss": 5.5044061005385725,
      "tokens_seen": 38141952
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004992076228686059,
      "loss": 4.376,
      "theoretical_loss": 5.502918488380116,
      "tokens_seen": 38207488
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000499197592778335,
      "loss": 4.5187,
      "theoretical_loss": 5.501434138753918,
      "tokens_seen": 38273024
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004991875626880642,
      "loss": 4.5366,
      "theoretical_loss": 5.499953038936635,
      "tokens_seen": 38338560
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004991775325977934,
      "loss": 4.5867,
      "theoretical_loss": 5.498475176276176,
      "tokens_seen": 38404096
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004991675025075226,
      "loss": 4.5793,
      "theoretical_loss": 5.497000538191195,
      "tokens_seen": 38469632
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004991574724172518,
      "loss": 4.4392,
      "theoretical_loss": 5.495529112170568,
      "tokens_seen": 38535168
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000499147442326981,
      "loss": 4.6516,
      "theoretical_loss": 5.494060885772887,
      "tokens_seen": 38600704
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004991374122367101,
      "loss": 4.5034,
      "theoretical_loss": 5.492595846625951,
      "tokens_seen": 38666240
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004991273821464393,
      "loss": 4.8256,
      "theoretical_loss": 5.491133982426266,
      "tokens_seen": 38731776
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004991173520561685,
      "loss": 5.0062,
      "theoretical_loss": 5.489675280938547,
      "tokens_seen": 38797312
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004991073219658977,
      "loss": 4.6335,
      "theoretical_loss": 5.488219729995227,
      "tokens_seen": 38862848
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000499097291875627,
      "loss": 4.2841,
      "theoretical_loss": 5.486767317495966,
      "tokens_seen": 38928384
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000499087261785356,
      "loss": 4.8828,
      "theoretical_loss": 5.48531803140717,
      "tokens_seen": 38993920
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004990772316950853,
      "loss": 4.6358,
      "theoretical_loss": 5.483871859761511,
      "tokens_seen": 39059456
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004990672016048144,
      "loss": 4.7315,
      "theoretical_loss": 5.482428790657449,
      "tokens_seen": 39124992
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004990571715145437,
      "loss": 4.8301,
      "theoretical_loss": 5.480988812258763,
      "tokens_seen": 39190528
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004990471414242729,
      "loss": 4.6629,
      "theoretical_loss": 5.479551912794086,
      "tokens_seen": 39256064
    },
    {
      "epoch": 0.01,
      "objective/train/docs_used": 141085,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 4.329990386962891,
      "objective/train/theoretical_loss": 5.478118080556438,
      "objective/train/tokens_used": 59781600,
      "theoretical_loss": 5.478118080556438,
      "tokens_seen": 39321600
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000499037111334002,
      "loss": 4.4817,
      "theoretical_loss": 5.478118080556438,
      "tokens_seen": 39321600
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004990270812437312,
      "loss": 4.7964,
      "theoretical_loss": 5.476687303902768,
      "tokens_seen": 39387136
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004990170511534604,
      "loss": 4.6476,
      "theoretical_loss": 5.475259571253502,
      "tokens_seen": 39452672
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004990070210631896,
      "loss": 4.5687,
      "theoretical_loss": 5.473834871092089,
      "tokens_seen": 39518208
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004989969909729188,
      "loss": 4.6718,
      "theoretical_loss": 5.4724131919645576,
      "tokens_seen": 39583744
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000498986960882648,
      "loss": 4.5519,
      "theoretical_loss": 5.470994522479069,
      "tokens_seen": 39649280
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004989769307923771,
      "loss": 4.8962,
      "theoretical_loss": 5.4695788513054815,
      "tokens_seen": 39714816
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004989669007021063,
      "loss": 4.7691,
      "theoretical_loss": 5.468166167174912,
      "tokens_seen": 39780352
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004989568706118355,
      "loss": 4.8426,
      "theoretical_loss": 5.466756458879306,
      "tokens_seen": 39845888
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004989468405215647,
      "loss": 4.3942,
      "theoretical_loss": 5.465349715271013,
      "tokens_seen": 39911424
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004989368104312939,
      "loss": 4.5906,
      "theoretical_loss": 5.463945925262355,
      "tokens_seen": 39976960
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000498926780341023,
      "loss": 4.7871,
      "theoretical_loss": 5.462545077825214,
      "tokens_seen": 40042496
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004989167502507523,
      "loss": 4.3864,
      "theoretical_loss": 5.461147161990611,
      "tokens_seen": 40108032
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004989067201604814,
      "loss": 4.5497,
      "theoretical_loss": 5.459752166848292,
      "tokens_seen": 40173568
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004988966900702107,
      "loss": 4.45,
      "theoretical_loss": 5.458360081546321,
      "tokens_seen": 40239104
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004988866599799398,
      "loss": 4.5501,
      "theoretical_loss": 5.456970895290674,
      "tokens_seen": 40304640
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004988766298896691,
      "loss": 4.6729,
      "theoretical_loss": 5.455584597344835,
      "tokens_seen": 40370176
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004988665997993982,
      "loss": 4.7367,
      "theoretical_loss": 5.454201177029395,
      "tokens_seen": 40435712
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004988565697091274,
      "loss": 4.5589,
      "theoretical_loss": 5.452820623721662,
      "tokens_seen": 40501248
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004988465396188566,
      "loss": 4.6359,
      "theoretical_loss": 5.45144292685526,
      "tokens_seen": 40566784
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004988365095285858,
      "loss": 4.5998,
      "theoretical_loss": 5.450068075919752,
      "tokens_seen": 40632320
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000498826479438315,
      "loss": 4.6231,
      "theoretical_loss": 5.44869606046024,
      "tokens_seen": 40697856
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004988164493480441,
      "loss": 4.5891,
      "theoretical_loss": 5.447326870076996,
      "tokens_seen": 40763392
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004988064192577733,
      "loss": 4.7368,
      "theoretical_loss": 5.445960494425072,
      "tokens_seen": 40828928
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004987963891675025,
      "loss": 4.4854,
      "theoretical_loss": 5.444596923213931,
      "tokens_seen": 40894464
    },
    {
      "epoch": 0.01,
      "objective/train/docs_used": 141824,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 4.5050435066223145,
      "objective/train/theoretical_loss": 5.443236146207074,
      "objective/train/tokens_used": 61420000,
      "theoretical_loss": 5.443236146207074,
      "tokens_seen": 40960000
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004987863590772317,
      "loss": 4.6512,
      "theoretical_loss": 5.443236146207074,
      "tokens_seen": 40960000
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004987763289869609,
      "loss": 4.5578,
      "theoretical_loss": 5.441878153221662,
      "tokens_seen": 41025536
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00049876629889669,
      "loss": 4.7668,
      "theoretical_loss": 5.440522934128164,
      "tokens_seen": 41091072
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004987562688064192,
      "loss": 4.523,
      "theoretical_loss": 5.439170478849976,
      "tokens_seen": 41156608
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004987462387161484,
      "loss": 4.6254,
      "theoretical_loss": 5.437820777363078,
      "tokens_seen": 41222144
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004987362086258777,
      "loss": 4.6625,
      "theoretical_loss": 5.4364738196956655,
      "tokens_seen": 41287680
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004987261785356068,
      "loss": 4.6117,
      "theoretical_loss": 5.435129595927794,
      "tokens_seen": 41353216
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004987161484453361,
      "loss": 4.7588,
      "theoretical_loss": 5.433788096191039,
      "tokens_seen": 41418752
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004987061183550651,
      "loss": 4.6583,
      "theoretical_loss": 5.432449310668134,
      "tokens_seen": 41484288
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004986960882647944,
      "loss": 4.5594,
      "theoretical_loss": 5.4311132295926345,
      "tokens_seen": 41549824
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004986860581745236,
      "loss": 4.682,
      "theoretical_loss": 5.42977984324857,
      "tokens_seen": 41615360
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004986760280842528,
      "loss": 4.5311,
      "theoretical_loss": 5.428449141970107,
      "tokens_seen": 41680896
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000498665997993982,
      "loss": 4.6285,
      "theoretical_loss": 5.427121116141212,
      "tokens_seen": 41746432
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004986559679037111,
      "loss": 4.4254,
      "theoretical_loss": 5.42579575619531,
      "tokens_seen": 41811968
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004986459378134403,
      "loss": 4.3521,
      "theoretical_loss": 5.424473052614967,
      "tokens_seen": 41877504
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004986359077231695,
      "loss": 4.2651,
      "theoretical_loss": 5.423152995931552,
      "tokens_seen": 41943040
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004986258776328987,
      "loss": 4.4635,
      "theoretical_loss": 5.421835576724906,
      "tokens_seen": 42008576
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004986158475426279,
      "loss": 4.4402,
      "theoretical_loss": 5.420520785623031,
      "tokens_seen": 42074112
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000498605817452357,
      "loss": 4.4318,
      "theoretical_loss": 5.4192086133017625,
      "tokens_seen": 42139648
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004985957873620862,
      "loss": 4.6869,
      "theoretical_loss": 5.417899050484451,
      "tokens_seen": 42205184
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004985857572718154,
      "loss": 4.5541,
      "theoretical_loss": 5.416592087941646,
      "tokens_seen": 42270720
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004985757271815446,
      "loss": 4.6914,
      "theoretical_loss": 5.415287716490787,
      "tokens_seen": 42336256
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004985656970912738,
      "loss": 4.4525,
      "theoretical_loss": 5.413985926995892,
      "tokens_seen": 42401792
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004985556670010031,
      "loss": 4.5715,
      "theoretical_loss": 5.412686710367245,
      "tokens_seen": 42467328
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004985456369107321,
      "loss": 4.4437,
      "theoretical_loss": 5.411390057561097,
      "tokens_seen": 42532864
    },
    {
      "epoch": 0.01,
      "objective/train/docs_used": 143445,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 4.786459922790527,
      "objective/train/theoretical_loss": 5.410095959579362,
      "objective/train/tokens_used": 63058400,
      "theoretical_loss": 5.410095959579362,
      "tokens_seen": 42598400
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004985356068204614,
      "loss": 4.5962,
      "theoretical_loss": 5.410095959579362,
      "tokens_seen": 42598400
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004985255767301905,
      "loss": 4.64,
      "theoretical_loss": 5.408804407469308,
      "tokens_seen": 42663936
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004985155466399198,
      "loss": 4.639,
      "theoretical_loss": 5.407515392323276,
      "tokens_seen": 42729472
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000498505516549649,
      "loss": 4.5622,
      "theoretical_loss": 5.406228905278368,
      "tokens_seen": 42795008
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004984954864593782,
      "loss": 4.7461,
      "theoretical_loss": 5.404944937516161,
      "tokens_seen": 42860544
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004984854563691073,
      "loss": 4.6478,
      "theoretical_loss": 5.403663480262418,
      "tokens_seen": 42926080
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004984754262788365,
      "loss": 4.8657,
      "theoretical_loss": 5.402384524786797,
      "tokens_seen": 42991616
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004984653961885657,
      "loss": 4.6469,
      "theoretical_loss": 5.401108062402562,
      "tokens_seen": 43057152
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004984553660982949,
      "loss": 4.3658,
      "theoretical_loss": 5.399834084466306,
      "tokens_seen": 43122688
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004984453360080241,
      "loss": 4.6816,
      "theoretical_loss": 5.398562582377666,
      "tokens_seen": 43188224
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004984353059177532,
      "loss": 4.3855,
      "theoretical_loss": 5.397293547579041,
      "tokens_seen": 43253760
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004984252758274825,
      "loss": 4.568,
      "theoretical_loss": 5.396026971555319,
      "tokens_seen": 43319296
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004984152457372116,
      "loss": 4.4991,
      "theoretical_loss": 5.394762845833601,
      "tokens_seen": 43384832
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004984052156469409,
      "loss": 4.583,
      "theoretical_loss": 5.393501161982926,
      "tokens_seen": 43450368
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00049839518555667,
      "loss": 4.2179,
      "theoretical_loss": 5.392241911614005,
      "tokens_seen": 43515904
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004983851554663993,
      "loss": 4.5224,
      "theoretical_loss": 5.390985086378949,
      "tokens_seen": 43581440
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004983751253761284,
      "loss": 4.7938,
      "theoretical_loss": 5.389730677971002,
      "tokens_seen": 43646976
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004983650952858576,
      "loss": 4.4534,
      "theoretical_loss": 5.388478678124285,
      "tokens_seen": 43712512
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004983550651955868,
      "loss": 4.2336,
      "theoretical_loss": 5.387229078613521,
      "tokens_seen": 43778048
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000498345035105316,
      "loss": 4.6225,
      "theoretical_loss": 5.385981871253785,
      "tokens_seen": 43843584
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004983350050150452,
      "loss": 4.4481,
      "theoretical_loss": 5.384737047900243,
      "tokens_seen": 43909120
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004983249749247743,
      "loss": 4.5426,
      "theoretical_loss": 5.3834946004478965,
      "tokens_seen": 43974656
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004983149448345035,
      "loss": 4.722,
      "theoretical_loss": 5.382254520831328,
      "tokens_seen": 44040192
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004983049147442327,
      "loss": 4.4346,
      "theoretical_loss": 5.381016801024449,
      "tokens_seen": 44105728
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004982948846539619,
      "loss": 4.5081,
      "theoretical_loss": 5.379781433040252,
      "tokens_seen": 44171264
    },
    {
      "epoch": 0.01,
      "objective/train/docs_used": 144147,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 4.279332637786865,
      "objective/train/theoretical_loss": 5.378548408930558,
      "objective/train/tokens_used": 64696800,
      "theoretical_loss": 5.378548408930558,
      "tokens_seen": 44236800
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004982848545636911,
      "loss": 4.6137,
      "theoretical_loss": 5.378548408930558,
      "tokens_seen": 44236800
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004982748244734202,
      "loss": 4.4353,
      "theoretical_loss": 5.377317720785777,
      "tokens_seen": 44302336
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004982647943831494,
      "loss": 4.4919,
      "theoretical_loss": 5.37608936073466,
      "tokens_seen": 44367872
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004982547642928786,
      "loss": 4.3873,
      "theoretical_loss": 5.374863320944057,
      "tokens_seen": 44433408
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004982447342026079,
      "loss": 4.5235,
      "theoretical_loss": 5.373639593618675,
      "tokens_seen": 44498944
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000498234704112337,
      "loss": 4.5877,
      "theoretical_loss": 5.372418171000847,
      "tokens_seen": 44564480
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004982246740220663,
      "loss": 4.3387,
      "theoretical_loss": 5.371199045370283,
      "tokens_seen": 44630016
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004982146439317953,
      "loss": 4.4968,
      "theoretical_loss": 5.369982209043851,
      "tokens_seen": 44695552
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004982046138415246,
      "loss": 4.5509,
      "theoretical_loss": 5.368767654375327,
      "tokens_seen": 44761088
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004981945837512538,
      "loss": 4.6583,
      "theoretical_loss": 5.367555373755179,
      "tokens_seen": 44826624
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000498184553660983,
      "loss": 4.2686,
      "theoretical_loss": 5.366345359610327,
      "tokens_seen": 44892160
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004981745235707122,
      "loss": 4.5189,
      "theoretical_loss": 5.365137604403923,
      "tokens_seen": 44957696
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004981644934804413,
      "loss": 4.7356,
      "theoretical_loss": 5.363932100635117,
      "tokens_seen": 45023232
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004981544633901705,
      "loss": 4.5647,
      "theoretical_loss": 5.362728840838843,
      "tokens_seen": 45088768
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004981444332998997,
      "loss": 4.4656,
      "theoretical_loss": 5.361527817585586,
      "tokens_seen": 45154304
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004981344032096289,
      "loss": 4.6026,
      "theoretical_loss": 5.360329023481169,
      "tokens_seen": 45219840
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004981243731193581,
      "loss": 4.4202,
      "theoretical_loss": 5.359132451166534,
      "tokens_seen": 45285376
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004981143430290873,
      "loss": 4.5494,
      "theoretical_loss": 5.357938093317518,
      "tokens_seen": 45350912
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004981043129388164,
      "loss": 4.3413,
      "theoretical_loss": 5.356745942644645,
      "tokens_seen": 45416448
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004980942828485456,
      "loss": 4.2947,
      "theoretical_loss": 5.355555991892905,
      "tokens_seen": 45481984
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004980842527582748,
      "loss": 4.7378,
      "theoretical_loss": 5.35436823384155,
      "tokens_seen": 45547520
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000498074222668004,
      "loss": 4.5112,
      "theoretical_loss": 5.353182661303873,
      "tokens_seen": 45613056
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004980641925777333,
      "loss": 4.6471,
      "theoretical_loss": 5.35199926712701,
      "tokens_seen": 45678592
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004980541624874623,
      "loss": 4.4652,
      "theoretical_loss": 5.350818044191721,
      "tokens_seen": 45744128
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004980441323971916,
      "loss": 4.3502,
      "theoretical_loss": 5.349638985412193,
      "tokens_seen": 45809664
    },
    {
      "epoch": 0.01,
      "objective/train/docs_used": 144842,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 4.692098617553711,
      "objective/train/theoretical_loss": 5.348462083735834,
      "objective/train/tokens_used": 66335200,
      "theoretical_loss": 5.348462083735834,
      "tokens_seen": 45875200
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004980341023069207,
      "loss": 4.4348,
      "theoretical_loss": 5.348462083735834,
      "tokens_seen": 45875200
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00049802407221665,
      "loss": 4.45,
      "theoretical_loss": 5.347287332143064,
      "tokens_seen": 45940736
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004980140421263792,
      "loss": 4.4526,
      "theoretical_loss": 5.346114723647119,
      "tokens_seen": 46006272
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004980040120361084,
      "loss": 4.5448,
      "theoretical_loss": 5.344944251293852,
      "tokens_seen": 46071808
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004979939819458375,
      "loss": 4.3139,
      "theoretical_loss": 5.343775908161532,
      "tokens_seen": 46137344
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004979839518555667,
      "loss": 4.5183,
      "theoretical_loss": 5.342609687360644,
      "tokens_seen": 46202880
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004979739217652959,
      "loss": 4.2806,
      "theoretical_loss": 5.341445582033705,
      "tokens_seen": 46268416
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004979638916750251,
      "loss": 4.6082,
      "theoretical_loss": 5.3402835853550545,
      "tokens_seen": 46333952
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004979538615847543,
      "loss": 4.286,
      "theoretical_loss": 5.339123690530673,
      "tokens_seen": 46399488
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004979438314944834,
      "loss": 4.4594,
      "theoretical_loss": 5.337965890797989,
      "tokens_seen": 46465024
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004979338014042126,
      "loss": 4.4121,
      "theoretical_loss": 5.336810179425685,
      "tokens_seen": 46530560
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004979237713139418,
      "loss": 4.3507,
      "theoretical_loss": 5.335656549713516,
      "tokens_seen": 46596096
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000497913741223671,
      "loss": 4.4211,
      "theoretical_loss": 5.334504994992115,
      "tokens_seen": 46661632
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004979037111334002,
      "loss": 4.548,
      "theoretical_loss": 5.333355508622814,
      "tokens_seen": 46727168
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004978936810431293,
      "loss": 4.4084,
      "theoretical_loss": 5.332208083997459,
      "tokens_seen": 46792704
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004978836509528586,
      "loss": 4.3348,
      "theoretical_loss": 5.33106271453822,
      "tokens_seen": 46858240
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004978736208625877,
      "loss": 4.3399,
      "theoretical_loss": 5.329919393697422,
      "tokens_seen": 46923776
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000497863590772317,
      "loss": 4.2743,
      "theoretical_loss": 5.328778114957351,
      "tokens_seen": 46989312
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004978535606820461,
      "loss": 4.3191,
      "theoretical_loss": 5.327638871830089,
      "tokens_seen": 47054848
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004978435305917754,
      "loss": 4.3063,
      "theoretical_loss": 5.326501657857326,
      "tokens_seen": 47120384
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004978335005015045,
      "loss": 4.7146,
      "theoretical_loss": 5.32536646661019,
      "tokens_seen": 47185920
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004978234704112337,
      "loss": 4.5358,
      "theoretical_loss": 5.324233291689069,
      "tokens_seen": 47251456
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004978134403209629,
      "loss": 4.3223,
      "theoretical_loss": 5.323102126723439,
      "tokens_seen": 47316992
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004978034102306921,
      "loss": 4.3505,
      "theoretical_loss": 5.321972965371691,
      "tokens_seen": 47382528
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004977933801404213,
      "loss": 4.515,
      "theoretical_loss": 5.320845801320959,
      "tokens_seen": 47448064
    },
    {
      "epoch": 0.01,
      "objective/train/docs_used": 146139,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 4.262459754943848,
      "objective/train/theoretical_loss": 5.319720628286955,
      "objective/train/tokens_used": 67973600,
      "theoretical_loss": 5.319720628286955,
      "tokens_seen": 47513600
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004977833500501504,
      "loss": 4.315,
      "theoretical_loss": 5.319720628286955,
      "tokens_seen": 47513600
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004977733199598796,
      "loss": 4.3331,
      "theoretical_loss": 5.318597440013795,
      "tokens_seen": 47579136
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004977632898696088,
      "loss": 4.458,
      "theoretical_loss": 5.317476230273831,
      "tokens_seen": 47644672
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000497753259779338,
      "loss": 4.3781,
      "theoretical_loss": 5.316356992867491,
      "tokens_seen": 47710208
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004977432296890672,
      "loss": 4.3273,
      "theoretical_loss": 5.31523972162311,
      "tokens_seen": 47775744
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004977331995987965,
      "loss": 4.613,
      "theoretical_loss": 5.314124410396767,
      "tokens_seen": 47841280
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004977231695085255,
      "loss": 4.3982,
      "theoretical_loss": 5.31301105307212,
      "tokens_seen": 47906816
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004977131394182548,
      "loss": 4.3032,
      "theoretical_loss": 5.311899643560251,
      "tokens_seen": 47972352
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000497703109327984,
      "loss": 4.449,
      "theoretical_loss": 5.310790175799497,
      "tokens_seen": 48037888
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004976930792377132,
      "loss": 4.4828,
      "theoretical_loss": 5.3096826437553,
      "tokens_seen": 48103424
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004976830491474424,
      "loss": 3.9312,
      "theoretical_loss": 5.308577041420046,
      "tokens_seen": 48168960
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004976730190571715,
      "loss": 4.3421,
      "theoretical_loss": 5.3074733628129005,
      "tokens_seen": 48234496
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004976629889669007,
      "loss": 4.5473,
      "theoretical_loss": 5.3063716019796665,
      "tokens_seen": 48300032
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004976529588766299,
      "loss": 4.5363,
      "theoretical_loss": 5.305271752992619,
      "tokens_seen": 48365568
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004976429287863591,
      "loss": 4.5261,
      "theoretical_loss": 5.304173809950358,
      "tokens_seen": 48431104
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004976328986960883,
      "loss": 4.2883,
      "theoretical_loss": 5.303077766977653,
      "tokens_seen": 48496640
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004976228686058175,
      "loss": 4.7128,
      "theoretical_loss": 5.3019836182252895,
      "tokens_seen": 48562176
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004976128385155466,
      "loss": 4.4607,
      "theoretical_loss": 5.300891357869929,
      "tokens_seen": 48627712
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004976028084252758,
      "loss": 4.5213,
      "theoretical_loss": 5.299800980113945,
      "tokens_seen": 48693248
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000497592778335005,
      "loss": 4.3922,
      "theoretical_loss": 5.298712479185288,
      "tokens_seen": 48758784
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004975827482447342,
      "loss": 4.3163,
      "theoretical_loss": 5.297625849337331,
      "tokens_seen": 48824320
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004975727181544635,
      "loss": 4.2912,
      "theoretical_loss": 5.296541084848727,
      "tokens_seen": 48889856
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004975626880641925,
      "loss": 4.47,
      "theoretical_loss": 5.295458180023262,
      "tokens_seen": 48955392
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004975526579739218,
      "loss": 4.3223,
      "theoretical_loss": 5.294377129189715,
      "tokens_seen": 49020928
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004975426278836509,
      "loss": 4.2547,
      "theoretical_loss": 5.293297926701706,
      "tokens_seen": 49086464
    },
    {
      "epoch": 0.01,
      "objective/train/docs_used": 146831,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 4.401537895202637,
      "objective/train/theoretical_loss": 5.292220566937567,
      "objective/train/tokens_used": 69612000,
      "theoretical_loss": 5.292220566937567,
      "tokens_seen": 49152000
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004975325977933802,
      "loss": 4.4595,
      "theoretical_loss": 5.292220566937567,
      "tokens_seen": 49152000
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004975225677031094,
      "loss": 4.2178,
      "theoretical_loss": 5.29114504430019,
      "tokens_seen": 49217536
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004975125376128386,
      "loss": 4.2229,
      "theoretical_loss": 5.290071353216895,
      "tokens_seen": 49283072
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004975025075225677,
      "loss": 4.3304,
      "theoretical_loss": 5.288999488139284,
      "tokens_seen": 49348608
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004974924774322969,
      "loss": 4.2719,
      "theoretical_loss": 5.28792944354311,
      "tokens_seen": 49414144
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004974824473420261,
      "loss": 4.4284,
      "theoretical_loss": 5.286861213928137,
      "tokens_seen": 49479680
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004974724172517553,
      "loss": 4.5986,
      "theoretical_loss": 5.285794793817999,
      "tokens_seen": 49545216
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004974623871614845,
      "loss": 4.2011,
      "theoretical_loss": 5.284730177760077,
      "tokens_seen": 49610752
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004974523570712136,
      "loss": 4.3674,
      "theoretical_loss": 5.283667360325351,
      "tokens_seen": 49676288
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004974423269809428,
      "loss": 4.5679,
      "theoretical_loss": 5.2826063361082785,
      "tokens_seen": 49741824
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000497432296890672,
      "loss": 4.6443,
      "theoretical_loss": 5.281547099726654,
      "tokens_seen": 49807360
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004974222668004012,
      "loss": 4.2155,
      "theoretical_loss": 5.280489645821483,
      "tokens_seen": 49872896
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004974122367101304,
      "loss": 4.2667,
      "theoretical_loss": 5.279433969056848,
      "tokens_seen": 49938432
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004974022066198595,
      "loss": 4.4801,
      "theoretical_loss": 5.278380064119782,
      "tokens_seen": 50003968
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004973921765295888,
      "loss": 4.2357,
      "theoretical_loss": 5.277327925720137,
      "tokens_seen": 50069504
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004973821464393179,
      "loss": 4.7003,
      "theoretical_loss": 5.276277548590457,
      "tokens_seen": 50135040
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004973721163490472,
      "loss": 4.1875,
      "theoretical_loss": 5.275228927485855,
      "tokens_seen": 50200576
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004973620862587763,
      "loss": 4.3904,
      "theoretical_loss": 5.2741820571838804,
      "tokens_seen": 50266112
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004973520561685056,
      "loss": 4.3429,
      "theoretical_loss": 5.273136932484399,
      "tokens_seen": 50331648
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004973420260782347,
      "loss": 4.6097,
      "theoretical_loss": 5.272093548209467,
      "tokens_seen": 50397184
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004973319959879639,
      "loss": 4.1553,
      "theoretical_loss": 5.271051899203207,
      "tokens_seen": 50462720
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004973219658976931,
      "loss": 4.3946,
      "theoretical_loss": 5.270011980331685,
      "tokens_seen": 50528256
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004973119358074223,
      "loss": 4.1563,
      "theoretical_loss": 5.268973786482794,
      "tokens_seen": 50593792
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004973019057171515,
      "loss": 4.4526,
      "theoretical_loss": 5.267937312566123,
      "tokens_seen": 50659328
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004972918756268806,
      "loss": 4.3346,
      "theoretical_loss": 5.266902553512847,
      "tokens_seen": 50724864
    },
    {
      "epoch": 0.02,
      "objective/train/docs_used": 148451,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.5394937992095947,
      "objective/train/theoretical_loss": 5.265869504275602,
      "objective/train/tokens_used": 71250400,
      "theoretical_loss": 5.265869504275602,
      "tokens_seen": 50790400
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004972818455366098,
      "loss": 4.2104,
      "theoretical_loss": 5.265869504275602,
      "tokens_seen": 50790400
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000497271815446339,
      "loss": 4.6599,
      "theoretical_loss": 5.264838159828369,
      "tokens_seen": 50855936
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004972617853560682,
      "loss": 4.5403,
      "theoretical_loss": 5.263808515166355,
      "tokens_seen": 50921472
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004972517552657974,
      "loss": 3.8993,
      "theoretical_loss": 5.262780565305875,
      "tokens_seen": 50987008
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004972417251755266,
      "loss": 4.4009,
      "theoretical_loss": 5.261754305284241,
      "tokens_seen": 51052544
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004972316950852557,
      "loss": 4.4019,
      "theoretical_loss": 5.260729730159641,
      "tokens_seen": 51118080
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004972216649949849,
      "loss": 4.191,
      "theoretical_loss": 5.259706835011027,
      "tokens_seen": 51183616
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004972116349047142,
      "loss": 4.3533,
      "theoretical_loss": 5.2586856149380035,
      "tokens_seen": 51249152
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004972016048144433,
      "loss": 4.1959,
      "theoretical_loss": 5.257666065060709,
      "tokens_seen": 51314688
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004971915747241726,
      "loss": 4.5251,
      "theoretical_loss": 5.256648180519708,
      "tokens_seen": 51380224
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004971815446339017,
      "loss": 4.348,
      "theoretical_loss": 5.255631956475881,
      "tokens_seen": 51445760
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004971715145436309,
      "loss": 4.4193,
      "theoretical_loss": 5.25461738811031,
      "tokens_seen": 51511296
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004971614844533601,
      "loss": 4.2312,
      "theoretical_loss": 5.25360447062417,
      "tokens_seen": 51576832
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004971514543630893,
      "loss": 4.6115,
      "theoretical_loss": 5.252593199238619,
      "tokens_seen": 51642368
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004971414242728185,
      "loss": 4.2891,
      "theoretical_loss": 5.2515835691946915,
      "tokens_seen": 51707904
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004971313941825477,
      "loss": 4.0652,
      "theoretical_loss": 5.2505755757531904,
      "tokens_seen": 51773440
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004971213640922768,
      "loss": 4.5308,
      "theoretical_loss": 5.24956921419458,
      "tokens_seen": 51838976
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000497111334002006,
      "loss": 4.2823,
      "theoretical_loss": 5.248564479818876,
      "tokens_seen": 51904512
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004971013039117352,
      "loss": 4.3883,
      "theoretical_loss": 5.247561367945544,
      "tokens_seen": 51970048
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004970912738214644,
      "loss": 4.3249,
      "theoretical_loss": 5.246559873913396,
      "tokens_seen": 52035584
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004970812437311936,
      "loss": 4.3358,
      "theoretical_loss": 5.245559993080484,
      "tokens_seen": 52101120
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004970712136409227,
      "loss": 4.0621,
      "theoretical_loss": 5.24456172082399,
      "tokens_seen": 52166656
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004970611835506519,
      "loss": 4.5677,
      "theoretical_loss": 5.243565052540136,
      "tokens_seen": 52232192
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004970511534603811,
      "loss": 4.4829,
      "theoretical_loss": 5.242569983644074,
      "tokens_seen": 52297728
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004970411233701103,
      "loss": 4.1669,
      "theoretical_loss": 5.241576509569784,
      "tokens_seen": 52363264
    },
    {
      "epoch": 0.02,
      "objective/train/docs_used": 149172,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 4.183032989501953,
      "objective/train/theoretical_loss": 5.240584625769978,
      "objective/train/tokens_used": 72888800,
      "theoretical_loss": 5.240584625769978,
      "tokens_seen": 52428800
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004970310932798396,
      "loss": 4.3346,
      "theoretical_loss": 5.240584625769978,
      "tokens_seen": 52428800
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004970210631895686,
      "loss": 4.3921,
      "theoretical_loss": 5.239594327715992,
      "tokens_seen": 52494336
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004970110330992979,
      "loss": 4.4333,
      "theoretical_loss": 5.238605610897698,
      "tokens_seen": 52559872
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004970010030090271,
      "loss": 4.2883,
      "theoretical_loss": 5.237618470823394,
      "tokens_seen": 52625408
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004969909729187563,
      "loss": 4.1672,
      "theoretical_loss": 5.2366329030197125,
      "tokens_seen": 52690944
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004969809428284855,
      "loss": 4.335,
      "theoretical_loss": 5.235648903031521,
      "tokens_seen": 52756480
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004969709127382147,
      "loss": 4.1939,
      "theoretical_loss": 5.2346664664218245,
      "tokens_seen": 52822016
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004969608826479438,
      "loss": 4.2423,
      "theoretical_loss": 5.233685588771669,
      "tokens_seen": 52887552
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000496950852557673,
      "loss": 4.1699,
      "theoretical_loss": 5.232706265680049,
      "tokens_seen": 52953088
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004969408224674022,
      "loss": 4.3159,
      "theoretical_loss": 5.231728492763811,
      "tokens_seen": 53018624
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004969307923771314,
      "loss": 3.9781,
      "theoretical_loss": 5.230752265657554,
      "tokens_seen": 53084160
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004969207622868606,
      "loss": 4.0656,
      "theoretical_loss": 5.229777580013545,
      "tokens_seen": 53149696
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004969107321965897,
      "loss": 4.6733,
      "theoretical_loss": 5.228804431501619,
      "tokens_seen": 53215232
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000496900702106319,
      "loss": 4.4798,
      "theoretical_loss": 5.227832815809087,
      "tokens_seen": 53280768
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004968906720160481,
      "loss": 4.4063,
      "theoretical_loss": 5.226862728640651,
      "tokens_seen": 53346304
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004968806419257774,
      "loss": 4.3503,
      "theoretical_loss": 5.2258941657183,
      "tokens_seen": 53411840
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004968706118355065,
      "loss": 4.346,
      "theoretical_loss": 5.2249271227812315,
      "tokens_seen": 53477376
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004968605817452358,
      "loss": 4.3345,
      "theoretical_loss": 5.223961595585755,
      "tokens_seen": 53542912
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004968505516549649,
      "loss": 4.2838,
      "theoretical_loss": 5.222997579905204,
      "tokens_seen": 53608448
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004968405215646941,
      "loss": 3.9395,
      "theoretical_loss": 5.222035071529845,
      "tokens_seen": 53673984
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004968304914744233,
      "loss": 4.2813,
      "theoretical_loss": 5.2210740662667945,
      "tokens_seen": 53739520
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004968204613841525,
      "loss": 4.2565,
      "theoretical_loss": 5.220114559939923,
      "tokens_seen": 53805056
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004968104312938817,
      "loss": 4.2532,
      "theoretical_loss": 5.219156548389775,
      "tokens_seen": 53870592
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004968004012036108,
      "loss": 4.1641,
      "theoretical_loss": 5.218200027473481,
      "tokens_seen": 53936128
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.00049679037111334,
      "loss": 4.2325,
      "theoretical_loss": 5.217244993064664,
      "tokens_seen": 54001664
    },
    {
      "epoch": 0.02,
      "objective/train/docs_used": 150441,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 4.059125900268555,
      "objective/train/theoretical_loss": 5.216291441053366,
      "objective/train/tokens_used": 74527200,
      "theoretical_loss": 5.216291441053366,
      "tokens_seen": 54067200
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004967803410230692,
      "loss": 4.3793,
      "theoretical_loss": 5.216291441053366,
      "tokens_seen": 54067200
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004967703109327984,
      "loss": 4.3139,
      "theoretical_loss": 5.215339367345955,
      "tokens_seen": 54132736
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004967602808425276,
      "loss": 4.4466,
      "theoretical_loss": 5.214388767865036,
      "tokens_seen": 54198272
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004967502507522568,
      "loss": 4.2178,
      "theoretical_loss": 5.2134396385493815,
      "tokens_seen": 54263808
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004967402206619859,
      "loss": 4.3888,
      "theoretical_loss": 5.212491975353835,
      "tokens_seen": 54329344
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004967301905717151,
      "loss": 4.4649,
      "theoretical_loss": 5.211545774249233,
      "tokens_seen": 54394880
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004967201604814444,
      "loss": 4.5714,
      "theoretical_loss": 5.210601031222324,
      "tokens_seen": 54460416
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004967101303911735,
      "loss": 4.3355,
      "theoretical_loss": 5.209657742275683,
      "tokens_seen": 54525952
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004967001003009028,
      "loss": 4.2998,
      "theoretical_loss": 5.208715903427631,
      "tokens_seen": 54591488
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000496690070210632,
      "loss": 4.4595,
      "theoretical_loss": 5.207775510712159,
      "tokens_seen": 54657024
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004966800401203611,
      "loss": 4.4266,
      "theoretical_loss": 5.2068365601788384,
      "tokens_seen": 54722560
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004966700100300903,
      "loss": 4.3538,
      "theoretical_loss": 5.205899047892753,
      "tokens_seen": 54788096
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004966599799398195,
      "loss": 4.2137,
      "theoretical_loss": 5.2049629699344075,
      "tokens_seen": 54853632
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004966499498495487,
      "loss": 4.4532,
      "theoretical_loss": 5.204028322399658,
      "tokens_seen": 54919168
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004966399197592779,
      "loss": 4.365,
      "theoretical_loss": 5.203095101399628,
      "tokens_seen": 54984704
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000496629889669007,
      "loss": 4.4775,
      "theoretical_loss": 5.202163303060633,
      "tokens_seen": 55050240
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004966198595787362,
      "loss": 4.6992,
      "theoretical_loss": 5.201232923524104,
      "tokens_seen": 55115776
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004966098294884654,
      "loss": 4.293,
      "theoretical_loss": 5.20030395894651,
      "tokens_seen": 55181312
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004965997993981946,
      "loss": 4.3133,
      "theoretical_loss": 5.199376405499277,
      "tokens_seen": 55246848
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004965897693079238,
      "loss": 4.4571,
      "theoretical_loss": 5.198450259368721,
      "tokens_seen": 55312384
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004965797392176529,
      "loss": 4.3994,
      "theoretical_loss": 5.197525516755965,
      "tokens_seen": 55377920
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004965697091273821,
      "loss": 3.9741,
      "theoretical_loss": 5.196602173876867,
      "tokens_seen": 55443456
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004965596790371113,
      "loss": 4.0186,
      "theoretical_loss": 5.195680226961947,
      "tokens_seen": 55508992
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004965496489468405,
      "loss": 4.2808,
      "theoretical_loss": 5.194759672256309,
      "tokens_seen": 55574528
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004965396188565698,
      "loss": 4.4242,
      "theoretical_loss": 5.19384050601957,
      "tokens_seen": 55640064
    },
    {
      "epoch": 0.02,
      "objective/train/docs_used": 151052,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.9295265674591064,
      "objective/train/theoretical_loss": 5.192922724525789,
      "objective/train/tokens_used": 76165600,
      "theoretical_loss": 5.192922724525789,
      "tokens_seen": 55705600
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004965295887662988,
      "loss": 4.1666,
      "theoretical_loss": 5.192922724525789,
      "tokens_seen": 55705600
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004965195586760281,
      "loss": 4.4752,
      "theoretical_loss": 5.19200632406339,
      "tokens_seen": 55771136
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004965095285857573,
      "loss": 4.162,
      "theoretical_loss": 5.19109130093509,
      "tokens_seen": 55836672
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004964994984954865,
      "loss": 4.268,
      "theoretical_loss": 5.190177651457833,
      "tokens_seen": 55902208
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004964894684052157,
      "loss": 4.2668,
      "theoretical_loss": 5.189265371962712,
      "tokens_seen": 55967744
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004964794383149449,
      "loss": 4.1351,
      "theoretical_loss": 5.188354458794902,
      "tokens_seen": 56033280
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000496469408224674,
      "loss": 4.3194,
      "theoretical_loss": 5.187444908313586,
      "tokens_seen": 56098816
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004964593781344032,
      "loss": 4.3667,
      "theoretical_loss": 5.186536716891892,
      "tokens_seen": 56164352
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004964493480441324,
      "loss": 4.3239,
      "theoretical_loss": 5.185629880916814,
      "tokens_seen": 56229888
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004964393179538616,
      "loss": 4.1802,
      "theoretical_loss": 5.18472439678915,
      "tokens_seen": 56295424
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004964292878635908,
      "loss": 4.1018,
      "theoretical_loss": 5.18382026092343,
      "tokens_seen": 56360960
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.00049641925777332,
      "loss": 4.2962,
      "theoretical_loss": 5.182917469747851,
      "tokens_seen": 56426496
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004964092276830491,
      "loss": 3.9407,
      "theoretical_loss": 5.182016019704204,
      "tokens_seen": 56492032
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004963991975927783,
      "loss": 4.468,
      "theoretical_loss": 5.1811159072478095,
      "tokens_seen": 56557568
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004963891675025075,
      "loss": 4.2165,
      "theoretical_loss": 5.180217128847451,
      "tokens_seen": 56623104
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004963791374122367,
      "loss": 4.3633,
      "theoretical_loss": 5.17931968098531,
      "tokens_seen": 56688640
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004963691073219659,
      "loss": 4.2837,
      "theoretical_loss": 5.178423560156894,
      "tokens_seen": 56754176
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004963590772316951,
      "loss": 4.5775,
      "theoretical_loss": 5.177528762870973,
      "tokens_seen": 56819712
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004963490471414242,
      "loss": 4.2386,
      "theoretical_loss": 5.176635285649521,
      "tokens_seen": 56885248
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004963390170511535,
      "loss": 4.4926,
      "theoretical_loss": 5.175743125027638,
      "tokens_seen": 56950784
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004963289869608827,
      "loss": 4.2936,
      "theoretical_loss": 5.174852277553498,
      "tokens_seen": 57016320
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004963189568706119,
      "loss": 4.3052,
      "theoretical_loss": 5.173962739788276,
      "tokens_seen": 57081856
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000496308926780341,
      "loss": 4.2385,
      "theoretical_loss": 5.17307450830609,
      "tokens_seen": 57147392
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004962988966900702,
      "loss": 4.0804,
      "theoretical_loss": 5.172187579693933,
      "tokens_seen": 57212928
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004962888665997994,
      "loss": 4.2481,
      "theoretical_loss": 5.1713019505516105,
      "tokens_seen": 57278464
    },
    {
      "epoch": 0.02,
      "objective/train/docs_used": 152615,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 4.15744161605835,
      "objective/train/theoretical_loss": 5.170417617491682,
      "objective/train/tokens_used": 77804000,
      "theoretical_loss": 5.170417617491682,
      "tokens_seen": 57344000
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004962788365095286,
      "loss": 4.1901,
      "theoretical_loss": 5.170417617491682,
      "tokens_seen": 57344000
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004962688064192578,
      "loss": 4.4107,
      "theoretical_loss": 5.169534577139395,
      "tokens_seen": 57409536
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000496258776328987,
      "loss": 4.4229,
      "theoretical_loss": 5.168652826132623,
      "tokens_seen": 57475072
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004962487462387161,
      "loss": 4.2645,
      "theoretical_loss": 5.167772361121805,
      "tokens_seen": 57540608
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004962387161484453,
      "loss": 4.0974,
      "theoretical_loss": 5.166893178769884,
      "tokens_seen": 57606144
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004962286860581746,
      "loss": 4.5159,
      "theoretical_loss": 5.1660152757522475,
      "tokens_seen": 57671680
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004962186559679037,
      "loss": 4.3628,
      "theoretical_loss": 5.165138648756665,
      "tokens_seen": 57737216
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000496208625877633,
      "loss": 3.9702,
      "theoretical_loss": 5.164263294483226,
      "tokens_seen": 57802752
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004961985957873621,
      "loss": 4.1025,
      "theoretical_loss": 5.163389209644287,
      "tokens_seen": 57868288
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004961885656970913,
      "loss": 4.0294,
      "theoretical_loss": 5.162516390964408,
      "tokens_seen": 57933824
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004961785356068205,
      "loss": 4.155,
      "theoretical_loss": 5.1616448351802875,
      "tokens_seen": 57999360
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004961685055165497,
      "loss": 4.1206,
      "theoretical_loss": 5.160774539040716,
      "tokens_seen": 58064896
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004961584754262789,
      "loss": 4.2619,
      "theoretical_loss": 5.159905499306511,
      "tokens_seen": 58130432
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004961484453360081,
      "loss": 4.2084,
      "theoretical_loss": 5.159037712750455,
      "tokens_seen": 58195968
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004961384152457372,
      "loss": 4.3638,
      "theoretical_loss": 5.158171176157245,
      "tokens_seen": 58261504
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004961283851554664,
      "loss": 4.4218,
      "theoretical_loss": 5.157305886323435,
      "tokens_seen": 58327040
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004961183550651956,
      "loss": 4.4587,
      "theoretical_loss": 5.156441840057371,
      "tokens_seen": 58392576
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004961083249749248,
      "loss": 4.1511,
      "theoretical_loss": 5.155579034179144,
      "tokens_seen": 58458112
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000496098294884654,
      "loss": 3.9374,
      "theoretical_loss": 5.15471746552053,
      "tokens_seen": 58523648
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004960882647943831,
      "loss": 4.1491,
      "theoretical_loss": 5.153857130924929,
      "tokens_seen": 58589184
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004960782347041123,
      "loss": 4.3473,
      "theoretical_loss": 5.1529980272473175,
      "tokens_seen": 58654720
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004960682046138415,
      "loss": 4.2407,
      "theoretical_loss": 5.152140151354191,
      "tokens_seen": 58720256
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004960581745235707,
      "loss": 4.2574,
      "theoretical_loss": 5.151283500123505,
      "tokens_seen": 58785792
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004960481444333,
      "loss": 4.196,
      "theoretical_loss": 5.150428070444621,
      "tokens_seen": 58851328
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000496038114343029,
      "loss": 4.1719,
      "theoretical_loss": 5.149573859218261,
      "tokens_seen": 58916864
    },
    {
      "epoch": 0.02,
      "objective/train/docs_used": 153245,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 4.41132926940918,
      "objective/train/theoretical_loss": 5.1487208633564405,
      "objective/train/tokens_used": 79442400,
      "theoretical_loss": 5.1487208633564405,
      "tokens_seen": 58982400
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004960280842527583,
      "loss": 4.2262,
      "theoretical_loss": 5.1487208633564405,
      "tokens_seen": 58982400
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004960180541624875,
      "loss": 4.3165,
      "theoretical_loss": 5.147869079782423,
      "tokens_seen": 59047936
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004960080240722167,
      "loss": 4.1407,
      "theoretical_loss": 5.147018505430666,
      "tokens_seen": 59113472
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004959979939819459,
      "loss": 4.2968,
      "theoretical_loss": 5.146169137246765,
      "tokens_seen": 59179008
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004959879638916751,
      "loss": 4.3721,
      "theoretical_loss": 5.145320972187402,
      "tokens_seen": 59244544
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004959779338014042,
      "loss": 3.9537,
      "theoretical_loss": 5.144474007220293,
      "tokens_seen": 59310080
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004959679037111334,
      "loss": 4.0455,
      "theoretical_loss": 5.143628239324139,
      "tokens_seen": 59375616
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004959578736208626,
      "loss": 4.1954,
      "theoretical_loss": 5.142783665488567,
      "tokens_seen": 59441152
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004959478435305918,
      "loss": 4.4969,
      "theoretical_loss": 5.1419402827140885,
      "tokens_seen": 59506688
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000495937813440321,
      "loss": 4.0657,
      "theoretical_loss": 5.141098088012036,
      "tokens_seen": 59572224
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004959277833500501,
      "loss": 4.0912,
      "theoretical_loss": 5.140257078404524,
      "tokens_seen": 59637760
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004959177532597793,
      "loss": 4.2994,
      "theoretical_loss": 5.13941725092439,
      "tokens_seen": 59703296
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004959077231695085,
      "loss": 4.0762,
      "theoretical_loss": 5.138578602615146,
      "tokens_seen": 59768832
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004958976930792377,
      "loss": 4.3081,
      "theoretical_loss": 5.137741130530934,
      "tokens_seen": 59834368
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004958876629889669,
      "loss": 4.1723,
      "theoretical_loss": 5.1369048317364685,
      "tokens_seen": 59899904
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000495877632898696,
      "loss": 4.1926,
      "theoretical_loss": 5.13606970330699,
      "tokens_seen": 59965440
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004958676028084253,
      "loss": 4.3296,
      "theoretical_loss": 5.135235742328217,
      "tokens_seen": 60030976
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004958575727181544,
      "loss": 4.2648,
      "theoretical_loss": 5.134402945896297,
      "tokens_seen": 60096512
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004958475426278837,
      "loss": 4.0622,
      "theoretical_loss": 5.133571311117755,
      "tokens_seen": 60162048
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004958375125376129,
      "loss": 4.1772,
      "theoretical_loss": 5.132740835109448,
      "tokens_seen": 60227584
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004958274824473421,
      "loss": 3.9454,
      "theoretical_loss": 5.131911514998518,
      "tokens_seen": 60293120
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004958174523570712,
      "loss": 4.3073,
      "theoretical_loss": 5.131083347922338,
      "tokens_seen": 60358656
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004958074222668004,
      "loss": 4.0057,
      "theoretical_loss": 5.130256331028474,
      "tokens_seen": 60424192
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004957973921765296,
      "loss": 4.3214,
      "theoretical_loss": 5.129430461474628,
      "tokens_seen": 60489728
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004957873620862588,
      "loss": 4.1685,
      "theoretical_loss": 5.128605736428597,
      "tokens_seen": 60555264
    },
    {
      "epoch": 0.02,
      "objective/train/docs_used": 154602,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 4.488969802856445,
      "objective/train/theoretical_loss": 5.127782153068225,
      "objective/train/tokens_used": 81080800,
      "theoretical_loss": 5.127782153068225,
      "tokens_seen": 60620800
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000495777331995988,
      "loss": 4.1876,
      "theoretical_loss": 5.127782153068225,
      "tokens_seen": 60620800
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004957673019057172,
      "loss": 4.1921,
      "theoretical_loss": 5.126959708581356,
      "tokens_seen": 60686336
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004957572718154463,
      "loss": 4.2615,
      "theoretical_loss": 5.1261384001657895,
      "tokens_seen": 60751872
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004957472417251755,
      "loss": 4.156,
      "theoretical_loss": 5.125318225029231,
      "tokens_seen": 60817408
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004957372116349047,
      "loss": 4.1728,
      "theoretical_loss": 5.124499180389249,
      "tokens_seen": 60882944
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004957271815446339,
      "loss": 4.2523,
      "theoretical_loss": 5.12368126347323,
      "tokens_seen": 60948480
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004957171514543631,
      "loss": 4.1074,
      "theoretical_loss": 5.122864471518334,
      "tokens_seen": 61014016
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004957071213640923,
      "loss": 4.2443,
      "theoretical_loss": 5.122048801771443,
      "tokens_seen": 61079552
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004956970912738214,
      "loss": 3.8773,
      "theoretical_loss": 5.121234251489128,
      "tokens_seen": 61145088
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004956870611835507,
      "loss": 4.0361,
      "theoretical_loss": 5.120420817937591,
      "tokens_seen": 61210624
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004956770310932798,
      "loss": 3.9077,
      "theoretical_loss": 5.119608498392633,
      "tokens_seen": 61276160
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004956670010030091,
      "loss": 3.9712,
      "theoretical_loss": 5.118797290139605,
      "tokens_seen": 61341696
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004956569709127383,
      "loss": 4.1847,
      "theoretical_loss": 5.117987190473361,
      "tokens_seen": 61407232
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004956469408224674,
      "loss": 3.9935,
      "theoretical_loss": 5.1171781966982195,
      "tokens_seen": 61472768
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004956369107321966,
      "loss": 3.932,
      "theoretical_loss": 5.116370306127921,
      "tokens_seen": 61538304
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004956268806419258,
      "loss": 4.3194,
      "theoretical_loss": 5.11556351608558,
      "tokens_seen": 61603840
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000495616850551655,
      "loss": 4.322,
      "theoretical_loss": 5.114757823903647,
      "tokens_seen": 61669376
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004956068204613842,
      "loss": 4.0708,
      "theoretical_loss": 5.113953226923864,
      "tokens_seen": 61734912
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004955967903711133,
      "loss": 4.3149,
      "theoretical_loss": 5.113149722497221,
      "tokens_seen": 61800448
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004955867602808425,
      "loss": 4.3968,
      "theoretical_loss": 5.112347307983919,
      "tokens_seen": 61865984
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004955767301905717,
      "loss": 4.1143,
      "theoretical_loss": 5.111545980753322,
      "tokens_seen": 61931520
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004955667001003009,
      "loss": 3.9537,
      "theoretical_loss": 5.110745738183919,
      "tokens_seen": 61997056
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004955566700100301,
      "loss": 4.1058,
      "theoretical_loss": 5.109946577663284,
      "tokens_seen": 62062592
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004955466399197592,
      "loss": 4.3783,
      "theoretical_loss": 5.109148496588032,
      "tokens_seen": 62128128
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004955366098294884,
      "loss": 4.1335,
      "theoretical_loss": 5.108351492363779,
      "tokens_seen": 62193664
    },
    {
      "epoch": 0.02,
      "objective/train/docs_used": 155165,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.9781312942504883,
      "objective/train/theoretical_loss": 5.107555562405102,
      "objective/train/tokens_used": 82719200,
      "theoretical_loss": 5.107555562405102,
      "tokens_seen": 62259200
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004955265797392177,
      "loss": 4.1683,
      "theoretical_loss": 5.107555562405102,
      "tokens_seen": 62259200
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004955165496489468,
      "loss": 4.3202,
      "theoretical_loss": 5.106760704135499,
      "tokens_seen": 62324736
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004955065195586761,
      "loss": 4.1865,
      "theoretical_loss": 5.105966914987349,
      "tokens_seen": 62390272
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004954964894684052,
      "loss": 4.0606,
      "theoretical_loss": 5.1051741924018685,
      "tokens_seen": 62455808
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004954864593781344,
      "loss": 4.1826,
      "theoretical_loss": 5.10438253382908,
      "tokens_seen": 62521344
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004954764292878636,
      "loss": 4.123,
      "theoretical_loss": 5.103591936727762,
      "tokens_seen": 62586880
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004954663991975928,
      "loss": 3.9976,
      "theoretical_loss": 5.102802398565418,
      "tokens_seen": 62652416
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000495456369107322,
      "loss": 4.1302,
      "theoretical_loss": 5.102013916818235,
      "tokens_seen": 62717952
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004954463390170512,
      "loss": 4.1543,
      "theoretical_loss": 5.101226488971042,
      "tokens_seen": 62783488
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004954363089267803,
      "loss": 3.8957,
      "theoretical_loss": 5.100440112517276,
      "tokens_seen": 62849024
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004954262788365095,
      "loss": 4.1473,
      "theoretical_loss": 5.09965478495894,
      "tokens_seen": 62914560
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004954162487462387,
      "loss": 4.3325,
      "theoretical_loss": 5.098870503806567,
      "tokens_seen": 62980096
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004954062186559679,
      "loss": 4.268,
      "theoretical_loss": 5.09808726657918,
      "tokens_seen": 63045632
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004953961885656971,
      "loss": 4.096,
      "theoretical_loss": 5.097305070804255,
      "tokens_seen": 63111168
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004953861584754263,
      "loss": 4.0601,
      "theoretical_loss": 5.096523914017688,
      "tokens_seen": 63176704
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004953761283851555,
      "loss": 3.6987,
      "theoretical_loss": 5.095743793763747,
      "tokens_seen": 63242240
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004953660982948846,
      "loss": 4.1201,
      "theoretical_loss": 5.094964707595047,
      "tokens_seen": 63307776
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004953560682046139,
      "loss": 3.6626,
      "theoretical_loss": 5.094186653072505,
      "tokens_seen": 63373312
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004953460381143431,
      "loss": 4.0223,
      "theoretical_loss": 5.093409627765306,
      "tokens_seen": 63438848
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004953360080240723,
      "loss": 4.2156,
      "theoretical_loss": 5.092633629250866,
      "tokens_seen": 63504384
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004953259779338014,
      "loss": 3.9654,
      "theoretical_loss": 5.091858655114796,
      "tokens_seen": 63569920
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004953159478435306,
      "loss": 4.0676,
      "theoretical_loss": 5.091084702950868,
      "tokens_seen": 63635456
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004953059177532598,
      "loss": 4.0842,
      "theoretical_loss": 5.090311770360971,
      "tokens_seen": 63700992
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000495295887662989,
      "loss": 4.2643,
      "theoretical_loss": 5.089539854955088,
      "tokens_seen": 63766528
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004952858575727182,
      "loss": 3.9339,
      "theoretical_loss": 5.088768954351249,
      "tokens_seen": 63832064
    },
    {
      "epoch": 0.02,
      "objective/train/docs_used": 156647,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.6908023357391357,
      "objective/train/theoretical_loss": 5.087999066175502,
      "objective/train/tokens_used": 84357600,
      "theoretical_loss": 5.087999066175502,
      "tokens_seen": 63897600
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004952758274824474,
      "loss": 3.8197,
      "theoretical_loss": 5.087999066175502,
      "tokens_seen": 63897600
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004952657973921765,
      "loss": 4.0257,
      "theoretical_loss": 5.0872301880618735,
      "tokens_seen": 63963136
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004952557673019057,
      "loss": 4.062,
      "theoretical_loss": 5.086462317652341,
      "tokens_seen": 64028672
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004952457372116349,
      "loss": 4.1515,
      "theoretical_loss": 5.085695452596788,
      "tokens_seen": 64094208
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004952357071213641,
      "loss": 4.1296,
      "theoretical_loss": 5.084929590552976,
      "tokens_seen": 64159744
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004952256770310933,
      "loss": 3.8944,
      "theoretical_loss": 5.0841647291865115,
      "tokens_seen": 64225280
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004952156469408225,
      "loss": 4.2949,
      "theoretical_loss": 5.083400866170806,
      "tokens_seen": 64290816
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004952056168505516,
      "loss": 3.9266,
      "theoretical_loss": 5.082637999187046,
      "tokens_seen": 64356352
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004951955867602809,
      "loss": 4.0743,
      "theoretical_loss": 5.081876125924159,
      "tokens_seen": 64421888
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.00049518555667001,
      "loss": 3.9946,
      "theoretical_loss": 5.0811152440787755,
      "tokens_seen": 64487424
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004951755265797393,
      "loss": 4.1916,
      "theoretical_loss": 5.0803553513552036,
      "tokens_seen": 64552960
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004951654964894685,
      "loss": 4.2992,
      "theoretical_loss": 5.079596445465386,
      "tokens_seen": 64618496
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004951554663991976,
      "loss": 3.9929,
      "theoretical_loss": 5.078838524128878,
      "tokens_seen": 64684032
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004951454363089268,
      "loss": 3.7549,
      "theoretical_loss": 5.078081585072802,
      "tokens_seen": 64749568
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000495135406218656,
      "loss": 4.0781,
      "theoretical_loss": 5.077325626031826,
      "tokens_seen": 64815104
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004951253761283852,
      "loss": 3.9946,
      "theoretical_loss": 5.076570644748123,
      "tokens_seen": 64880640
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004951153460381144,
      "loss": 3.966,
      "theoretical_loss": 5.075816638971341,
      "tokens_seen": 64946176
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004951053159478435,
      "loss": 4.3961,
      "theoretical_loss": 5.075063606458576,
      "tokens_seen": 65011712
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004950952858575727,
      "loss": 4.1845,
      "theoretical_loss": 5.074311544974331,
      "tokens_seen": 65077248
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004950852557673019,
      "loss": 4.0004,
      "theoretical_loss": 5.07356045229049,
      "tokens_seen": 65142784
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004950752256770311,
      "loss": 3.7156,
      "theoretical_loss": 5.072810326186285,
      "tokens_seen": 65208320
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004950651955867603,
      "loss": 4.0024,
      "theoretical_loss": 5.072061164448261,
      "tokens_seen": 65273856
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004950551654964894,
      "loss": 4.0715,
      "theoretical_loss": 5.071312964870252,
      "tokens_seen": 65339392
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004950451354062186,
      "loss": 4.07,
      "theoretical_loss": 5.070565725253344,
      "tokens_seen": 65404928
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004950351053159479,
      "loss": 4.1316,
      "theoretical_loss": 5.069819443405842,
      "tokens_seen": 65470464
    },
    {
      "epoch": 0.02,
      "objective/train/docs_used": 157489,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 4.333700180053711,
      "objective/train/theoretical_loss": 5.069074117143246,
      "objective/train/tokens_used": 85996000,
      "theoretical_loss": 5.069074117143246,
      "tokens_seen": 65536000
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000495025075225677,
      "loss": 3.9793,
      "theoretical_loss": 5.069074117143246,
      "tokens_seen": 65536000
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004950150451354063,
      "loss": 3.8486,
      "theoretical_loss": 5.068329744288216,
      "tokens_seen": 65601536
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004950050150451354,
      "loss": 4.252,
      "theoretical_loss": 5.067586322670541,
      "tokens_seen": 65667072
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004949949849548646,
      "loss": 4.079,
      "theoretical_loss": 5.0668438501271105,
      "tokens_seen": 65732608
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004949849548645938,
      "loss": 4.09,
      "theoretical_loss": 5.066102324501883,
      "tokens_seen": 65798144
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000494974924774323,
      "loss": 4.1249,
      "theoretical_loss": 5.065361743645855,
      "tokens_seen": 65863680
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004949648946840522,
      "loss": 3.8237,
      "theoretical_loss": 5.064622105417033,
      "tokens_seen": 65929216
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004949548645937814,
      "loss": 4.0333,
      "theoretical_loss": 5.063883407680405,
      "tokens_seen": 65994752
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004949448345035105,
      "loss": 4.0507,
      "theoretical_loss": 5.063145648307904,
      "tokens_seen": 66060288
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004949348044132397,
      "loss": 3.5886,
      "theoretical_loss": 5.062408825178388,
      "tokens_seen": 66125824
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004949247743229689,
      "loss": 3.9934,
      "theoretical_loss": 5.061672936177604,
      "tokens_seen": 66191360
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004949147442326981,
      "loss": 3.8668,
      "theoretical_loss": 5.06093797919816,
      "tokens_seen": 66256896
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004949047141424273,
      "loss": 4.1734,
      "theoretical_loss": 5.060203952139497,
      "tokens_seen": 66322432
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004948946840521565,
      "loss": 3.8566,
      "theoretical_loss": 5.059470852907861,
      "tokens_seen": 66387968
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004948846539618856,
      "loss": 3.9639,
      "theoretical_loss": 5.0587386794162725,
      "tokens_seen": 66453504
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004948746238716148,
      "loss": 3.9601,
      "theoretical_loss": 5.058007429584498,
      "tokens_seen": 66519040
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000494864593781344,
      "loss": 3.7703,
      "theoretical_loss": 5.057277101339023,
      "tokens_seen": 66584576
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004948545636910733,
      "loss": 4.193,
      "theoretical_loss": 5.056547692613021,
      "tokens_seen": 66650112
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004948445336008024,
      "loss": 3.8359,
      "theoretical_loss": 5.055819201346331,
      "tokens_seen": 66715648
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004948345035105316,
      "loss": 4.019,
      "theoretical_loss": 5.055091625485421,
      "tokens_seen": 66781184
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004948244734202607,
      "loss": 3.6005,
      "theoretical_loss": 5.054364962983367,
      "tokens_seen": 66846720
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.00049481444332999,
      "loss": 4.065,
      "theoretical_loss": 5.053639211799824,
      "tokens_seen": 66912256
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004948044132397192,
      "loss": 4.0457,
      "theoretical_loss": 5.052914369900997,
      "tokens_seen": 66977792
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004947943831494484,
      "loss": 3.8273,
      "theoretical_loss": 5.052190435259614,
      "tokens_seen": 67043328
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004947843530591776,
      "loss": 3.9551,
      "theoretical_loss": 5.051467405854897,
      "tokens_seen": 67108864
    },
    {
      "epoch": 0.02,
      "objective/train/docs_used": 158845,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.314136505126953,
      "objective/train/theoretical_loss": 5.05074527967254,
      "objective/train/tokens_used": 87634400,
      "theoretical_loss": 5.05074527967254,
      "tokens_seen": 67174400
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004947743229689067,
      "loss": 3.8036,
      "theoretical_loss": 5.05074527967254,
      "tokens_seen": 67174400
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004947642928786359,
      "loss": 4.2425,
      "theoretical_loss": 5.050024054704677,
      "tokens_seen": 67239936
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004947542627883651,
      "loss": 3.6902,
      "theoretical_loss": 5.049303728949859,
      "tokens_seen": 67305472
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004947442326980943,
      "loss": 4.0812,
      "theoretical_loss": 5.048584300413019,
      "tokens_seen": 67371008
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004947342026078235,
      "loss": 3.9309,
      "theoretical_loss": 5.04786576710546,
      "tokens_seen": 67436544
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004947241725175527,
      "loss": 4.0231,
      "theoretical_loss": 5.0471481270448155,
      "tokens_seen": 67502080
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004947141424272818,
      "loss": 3.7707,
      "theoretical_loss": 5.046431378255027,
      "tokens_seen": 67567616
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004947041123370111,
      "loss": 4.1525,
      "theoretical_loss": 5.045715518766322,
      "tokens_seen": 67633152
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004946940822467402,
      "loss": 3.7804,
      "theoretical_loss": 5.0450005466151815,
      "tokens_seen": 67698688
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004946840521564695,
      "loss": 3.9759,
      "theoretical_loss": 5.044286459844319,
      "tokens_seen": 67764224
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004946740220661987,
      "loss": 4.0479,
      "theoretical_loss": 5.043573256502652,
      "tokens_seen": 67829760
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004946639919759278,
      "loss": 4.1937,
      "theoretical_loss": 5.0428609346452795,
      "tokens_seen": 67895296
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000494653961885657,
      "loss": 4.1032,
      "theoretical_loss": 5.042149492333452,
      "tokens_seen": 67960832
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004946439317953862,
      "loss": 3.7257,
      "theoretical_loss": 5.041438927634549,
      "tokens_seen": 68026368
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004946339017051154,
      "loss": 4.3465,
      "theoretical_loss": 5.040729238622053,
      "tokens_seen": 68091904
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004946238716148446,
      "loss": 3.7946,
      "theoretical_loss": 5.040020423375525,
      "tokens_seen": 68157440
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004946138415245737,
      "loss": 3.8581,
      "theoretical_loss": 5.039312479980579,
      "tokens_seen": 68222976
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004946038114343029,
      "loss": 3.9283,
      "theoretical_loss": 5.038605406528857,
      "tokens_seen": 68288512
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004945937813440321,
      "loss": 3.9674,
      "theoretical_loss": 5.037899201118005,
      "tokens_seen": 68354048
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004945837512537613,
      "loss": 3.9114,
      "theoretical_loss": 5.037193861851646,
      "tokens_seen": 68419584
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004945737211634905,
      "loss": 4.0903,
      "theoretical_loss": 5.03648938683936,
      "tokens_seen": 68485120
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004945636910732196,
      "loss": 4.0617,
      "theoretical_loss": 5.035785774196654,
      "tokens_seen": 68550656
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004945536609829488,
      "loss": 3.9602,
      "theoretical_loss": 5.035083022044944,
      "tokens_seen": 68616192
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004945436308926781,
      "loss": 3.7551,
      "theoretical_loss": 5.034381128511525,
      "tokens_seen": 68681728
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004945336008024072,
      "loss": 3.728,
      "theoretical_loss": 5.0336800917295506,
      "tokens_seen": 68747264
    },
    {
      "epoch": 0.02,
      "objective/train/docs_used": 159466,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.931746244430542,
      "objective/train/theoretical_loss": 5.032979909838007,
      "objective/train/tokens_used": 89272800,
      "theoretical_loss": 5.032979909838007,
      "tokens_seen": 68812800
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004945235707121365,
      "loss": 4.1672,
      "theoretical_loss": 5.032979909838007,
      "tokens_seen": 68812800
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004945135406218656,
      "loss": 3.7009,
      "theoretical_loss": 5.032280580981691,
      "tokens_seen": 68878336
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004945035105315948,
      "loss": 3.9125,
      "theoretical_loss": 5.031582103311187,
      "tokens_seen": 68943872
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000494493480441324,
      "loss": 3.8562,
      "theoretical_loss": 5.030884474982842,
      "tokens_seen": 69009408
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004944834503510532,
      "loss": 4.0789,
      "theoretical_loss": 5.030187694158739,
      "tokens_seen": 69074944
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004944734202607824,
      "loss": 3.6415,
      "theoretical_loss": 5.02949175900668,
      "tokens_seen": 69140480
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004944633901705116,
      "loss": 3.9927,
      "theoretical_loss": 5.028796667700159,
      "tokens_seen": 69206016
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004944533600802407,
      "loss": 3.6965,
      "theoretical_loss": 5.0281024184183405,
      "tokens_seen": 69271552
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004944433299899699,
      "loss": 3.9956,
      "theoretical_loss": 5.0274090093460355,
      "tokens_seen": 69337088
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004944332998996991,
      "loss": 4.0111,
      "theoretical_loss": 5.026716438673677,
      "tokens_seen": 69402624
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004944232698094283,
      "loss": 3.8318,
      "theoretical_loss": 5.0260247045973045,
      "tokens_seen": 69468160
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004944132397191575,
      "loss": 4.0343,
      "theoretical_loss": 5.02533380531853,
      "tokens_seen": 69533696
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004944032096288867,
      "loss": 3.6072,
      "theoretical_loss": 5.024643739044526,
      "tokens_seen": 69599232
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004943931795386158,
      "loss": 4.0637,
      "theoretical_loss": 5.023954503987998,
      "tokens_seen": 69664768
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000494383149448345,
      "loss": 3.9123,
      "theoretical_loss": 5.023266098367161,
      "tokens_seen": 69730304
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004943731193580742,
      "loss": 3.7934,
      "theoretical_loss": 5.022578520405721,
      "tokens_seen": 69795840
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004943630892678035,
      "loss": 3.8608,
      "theoretical_loss": 5.0218917683328534,
      "tokens_seen": 69861376
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004943530591775326,
      "loss": 3.8826,
      "theoretical_loss": 5.021205840383175,
      "tokens_seen": 69926912
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004943430290872618,
      "loss": 3.7947,
      "theoretical_loss": 5.020520734796728,
      "tokens_seen": 69992448
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004943329989969909,
      "loss": 3.8164,
      "theoretical_loss": 5.019836449818957,
      "tokens_seen": 70057984
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004943229689067202,
      "loss": 3.6969,
      "theoretical_loss": 5.019152983700687,
      "tokens_seen": 70123520
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004943129388164494,
      "loss": 3.8535,
      "theoretical_loss": 5.018470334698101,
      "tokens_seen": 70189056
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004943029087261786,
      "loss": 3.9834,
      "theoretical_loss": 5.01778850107272,
      "tokens_seen": 70254592
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004942928786359078,
      "loss": 3.7347,
      "theoretical_loss": 5.017107481091379,
      "tokens_seen": 70320128
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004942828485456369,
      "loss": 3.6688,
      "theoretical_loss": 5.016427273026212,
      "tokens_seen": 70385664
    },
    {
      "epoch": 0.02,
      "objective/train/docs_used": 160845,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.9590885639190674,
      "objective/train/theoretical_loss": 5.015747875154622,
      "objective/train/tokens_used": 90911200,
      "theoretical_loss": 5.015747875154622,
      "tokens_seen": 70451200
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004942728184553661,
      "loss": 4.0648,
      "theoretical_loss": 5.015747875154622,
      "tokens_seen": 70451200
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004942627883650953,
      "loss": 3.7767,
      "theoretical_loss": 5.015069285759269,
      "tokens_seen": 70516736
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004942527582748245,
      "loss": 4.1511,
      "theoretical_loss": 5.01439150312804,
      "tokens_seen": 70582272
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004942427281845537,
      "loss": 3.8699,
      "theoretical_loss": 5.0137145255540405,
      "tokens_seen": 70647808
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004942326980942828,
      "loss": 3.8057,
      "theoretical_loss": 5.013038351335559,
      "tokens_seen": 70713344
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000494222668004012,
      "loss": 3.8407,
      "theoretical_loss": 5.012362978776057,
      "tokens_seen": 70778880
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004942126379137412,
      "loss": 3.7493,
      "theoretical_loss": 5.011688406184147,
      "tokens_seen": 70844416
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004942026078234704,
      "loss": 3.9297,
      "theoretical_loss": 5.011014631873566,
      "tokens_seen": 70909952
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004941925777331996,
      "loss": 4.0194,
      "theoretical_loss": 5.010341654163167,
      "tokens_seen": 70975488
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004941825476429289,
      "loss": 3.9314,
      "theoretical_loss": 5.009669471376882,
      "tokens_seen": 71041024
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004941725175526579,
      "loss": 3.553,
      "theoretical_loss": 5.008998081843721,
      "tokens_seen": 71106560
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004941624874623872,
      "loss": 4.1107,
      "theoretical_loss": 5.008327483897736,
      "tokens_seen": 71172096
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004941524573721163,
      "loss": 3.8063,
      "theoretical_loss": 5.00765767587801,
      "tokens_seen": 71237632
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004941424272818456,
      "loss": 3.7438,
      "theoretical_loss": 5.006988656128635,
      "tokens_seen": 71303168
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004941323971915748,
      "loss": 3.6967,
      "theoretical_loss": 5.006320422998691,
      "tokens_seen": 71368704
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004941223671013039,
      "loss": 3.9589,
      "theoretical_loss": 5.00565297484223,
      "tokens_seen": 71434240
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004941123370110331,
      "loss": 3.9641,
      "theoretical_loss": 5.004986310018252,
      "tokens_seen": 71499776
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004941023069207623,
      "loss": 4.093,
      "theoretical_loss": 5.004320426890686,
      "tokens_seen": 71565312
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004940922768304915,
      "loss": 3.8581,
      "theoretical_loss": 5.003655323828376,
      "tokens_seen": 71630848
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004940822467402207,
      "loss": 3.9614,
      "theoretical_loss": 5.002990999205057,
      "tokens_seen": 71696384
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004940722166499498,
      "loss": 3.9838,
      "theoretical_loss": 5.002327451399335,
      "tokens_seen": 71761920
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000494062186559679,
      "loss": 4.1628,
      "theoretical_loss": 5.001664678794671,
      "tokens_seen": 71827456
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004940521564694082,
      "loss": 3.9386,
      "theoretical_loss": 5.001002679779363,
      "tokens_seen": 71892992
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004940421263791374,
      "loss": 3.9657,
      "theoretical_loss": 5.0003414527465235,
      "tokens_seen": 71958528
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004940320962888666,
      "loss": 4.0775,
      "theoretical_loss": 4.99968099609406,
      "tokens_seen": 72024064
    },
    {
      "epoch": 0.02,
      "objective/train/docs_used": 161502,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 4.11187744140625,
      "objective/train/theoretical_loss": 4.999021308224664,
      "objective/train/tokens_used": 92549600,
      "theoretical_loss": 4.999021308224664,
      "tokens_seen": 72089600
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004940220661985958,
      "loss": 3.9657,
      "theoretical_loss": 4.999021308224664,
      "tokens_seen": 72089600
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004940120361083249,
      "loss": 3.9299,
      "theoretical_loss": 4.998362387545782,
      "tokens_seen": 72155136
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004940020060180542,
      "loss": 3.833,
      "theoretical_loss": 4.997704232469606,
      "tokens_seen": 72220672
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004939919759277834,
      "loss": 3.6457,
      "theoretical_loss": 4.997046841413049,
      "tokens_seen": 72286208
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004939819458375126,
      "loss": 4.0112,
      "theoretical_loss": 4.996390212797728,
      "tokens_seen": 72351744
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004939719157472418,
      "loss": 3.6914,
      "theoretical_loss": 4.995734345049949,
      "tokens_seen": 72417280
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000493961885656971,
      "loss": 3.8694,
      "theoretical_loss": 4.995079236600686,
      "tokens_seen": 72482816
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004939518555667001,
      "loss": 3.9413,
      "theoretical_loss": 4.994424885885564,
      "tokens_seen": 72548352
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004939418254764293,
      "loss": 4.0597,
      "theoretical_loss": 4.993771291344839,
      "tokens_seen": 72613888
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004939317953861585,
      "loss": 4.0831,
      "theoretical_loss": 4.993118451423381,
      "tokens_seen": 72679424
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004939217652958877,
      "loss": 3.8775,
      "theoretical_loss": 4.992466364570659,
      "tokens_seen": 72744960
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004939117352056169,
      "loss": 4.0318,
      "theoretical_loss": 4.991815029240721,
      "tokens_seen": 72810496
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000493901705115346,
      "loss": 3.9009,
      "theoretical_loss": 4.991164443892175,
      "tokens_seen": 72876032
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004938916750250752,
      "loss": 3.8903,
      "theoretical_loss": 4.990514606988173,
      "tokens_seen": 72941568
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004938816449348044,
      "loss": 3.8758,
      "theoretical_loss": 4.989865516996396,
      "tokens_seen": 73007104
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004938716148445337,
      "loss": 3.6876,
      "theoretical_loss": 4.98921717238903,
      "tokens_seen": 73072640
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004938615847542628,
      "loss": 3.9345,
      "theoretical_loss": 4.988569571642756,
      "tokens_seen": 73138176
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000493851554663992,
      "loss": 3.6821,
      "theoretical_loss": 4.98792271323873,
      "tokens_seen": 73203712
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004938415245737211,
      "loss": 3.9708,
      "theoretical_loss": 4.9872765956625615,
      "tokens_seen": 73269248
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004938314944834504,
      "loss": 3.8872,
      "theoretical_loss": 4.9866312174043035,
      "tokens_seen": 73334784
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004938214643931796,
      "loss": 3.799,
      "theoretical_loss": 4.9859865769584335,
      "tokens_seen": 73400320
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004938114343029088,
      "loss": 3.9449,
      "theoretical_loss": 4.9853426728238315,
      "tokens_seen": 73465856
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000493801404212638,
      "loss": 3.5261,
      "theoretical_loss": 4.984699503503771,
      "tokens_seen": 73531392
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004937913741223671,
      "loss": 3.9521,
      "theoretical_loss": 4.984057067505898,
      "tokens_seen": 73596928
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004937813440320963,
      "loss": 3.977,
      "theoretical_loss": 4.9834153633422105,
      "tokens_seen": 73662464
    },
    {
      "epoch": 0.02,
      "objective/train/docs_used": 162491,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.735835313796997,
      "objective/train/theoretical_loss": 4.982774389529053,
      "objective/train/tokens_used": 94188000,
      "theoretical_loss": 4.982774389529053,
      "tokens_seen": 73728000
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004937713139418255,
      "loss": 3.751,
      "theoretical_loss": 4.982774389529053,
      "tokens_seen": 73728000
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004937612838515547,
      "loss": 3.8968,
      "theoretical_loss": 4.9821341445870875,
      "tokens_seen": 73793536
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004937512537612839,
      "loss": 3.5854,
      "theoretical_loss": 4.981494627041286,
      "tokens_seen": 73859072
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000493741223671013,
      "loss": 3.5525,
      "theoretical_loss": 4.98085583542091,
      "tokens_seen": 73924608
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004937311935807422,
      "loss": 3.9853,
      "theoretical_loss": 4.980217768259496,
      "tokens_seen": 73990144
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004937211634904714,
      "loss": 3.9015,
      "theoretical_loss": 4.979580424094836,
      "tokens_seen": 74055680
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004937111334002006,
      "loss": 4.0376,
      "theoretical_loss": 4.978943801468967,
      "tokens_seen": 74121216
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004937011033099298,
      "loss": 3.5555,
      "theoretical_loss": 4.978307898928149,
      "tokens_seen": 74186752
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004936910732196591,
      "loss": 3.9641,
      "theoretical_loss": 4.977672715022855,
      "tokens_seen": 74252288
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004936810431293881,
      "loss": 3.8699,
      "theoretical_loss": 4.97703824830775,
      "tokens_seen": 74317824
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004936710130391174,
      "loss": 3.9669,
      "theoretical_loss": 4.976404497341676,
      "tokens_seen": 74383360
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004936609829488465,
      "loss": 3.6727,
      "theoretical_loss": 4.975771460687641,
      "tokens_seen": 74448896
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004936509528585758,
      "loss": 3.4212,
      "theoretical_loss": 4.975139136912794,
      "tokens_seen": 74514432
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000493640922768305,
      "loss": 3.8349,
      "theoretical_loss": 4.974507524588424,
      "tokens_seen": 74579968
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004936308926780341,
      "loss": 3.9229,
      "theoretical_loss": 4.973876622289927,
      "tokens_seen": 74645504
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004936208625877633,
      "loss": 3.4796,
      "theoretical_loss": 4.973246428596802,
      "tokens_seen": 74711040
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004936108324974925,
      "loss": 3.7328,
      "theoretical_loss": 4.972616942092634,
      "tokens_seen": 74776576
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004936008024072217,
      "loss": 4.1417,
      "theoretical_loss": 4.971988161365077,
      "tokens_seen": 74842112
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004935907723169509,
      "loss": 3.5812,
      "theoretical_loss": 4.9713600850058395,
      "tokens_seen": 74907648
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.00049358074222668,
      "loss": 3.661,
      "theoretical_loss": 4.970732711610667,
      "tokens_seen": 74973184
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004935707121364092,
      "loss": 3.9342,
      "theoretical_loss": 4.97010603977933,
      "tokens_seen": 75038720
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004935606820461384,
      "loss": 3.7614,
      "theoretical_loss": 4.96948006811561,
      "tokens_seen": 75104256
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004935506519558676,
      "loss": 3.6541,
      "theoretical_loss": 4.968854795227281,
      "tokens_seen": 75169792
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004935406218655968,
      "loss": 3.8975,
      "theoretical_loss": 4.968230219726093,
      "tokens_seen": 75235328
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000493530591775326,
      "loss": 3.8834,
      "theoretical_loss": 4.967606340227765,
      "tokens_seen": 75300864
    },
    {
      "epoch": 0.02,
      "objective/train/docs_used": 163075,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.548557996749878,
      "objective/train/theoretical_loss": 4.966983155351962,
      "objective/train/tokens_used": 95826400,
      "theoretical_loss": 4.966983155351962,
      "tokens_seen": 75366400
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004935205616850551,
      "loss": 3.6221,
      "theoretical_loss": 4.966983155351962,
      "tokens_seen": 75366400
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004935105315947844,
      "loss": 4.1199,
      "theoretical_loss": 4.966360663722287,
      "tokens_seen": 75431936
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004935005015045135,
      "loss": 3.9567,
      "theoretical_loss": 4.96573886396626,
      "tokens_seen": 75497472
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004934904714142428,
      "loss": 4.2001,
      "theoretical_loss": 4.965117754715307,
      "tokens_seen": 75563008
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004934804413239719,
      "loss": 3.8199,
      "theoretical_loss": 4.964497334604748,
      "tokens_seen": 75628544
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004934704112337011,
      "loss": 3.8548,
      "theoretical_loss": 4.963877602273776,
      "tokens_seen": 75694080
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004934603811434303,
      "loss": 4.0167,
      "theoretical_loss": 4.963258556365449,
      "tokens_seen": 75759616
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004934503510531595,
      "loss": 3.5764,
      "theoretical_loss": 4.962640195526673,
      "tokens_seen": 75825152
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004934403209628887,
      "loss": 3.9003,
      "theoretical_loss": 4.962022518408183,
      "tokens_seen": 75890688
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004934302908726179,
      "loss": 3.9603,
      "theoretical_loss": 4.96140552366454,
      "tokens_seen": 75956224
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000493420260782347,
      "loss": 3.6928,
      "theoretical_loss": 4.9607892099541075,
      "tokens_seen": 76021760
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004934102306920762,
      "loss": 3.9968,
      "theoretical_loss": 4.9601735759390415,
      "tokens_seen": 76087296
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004934002006018054,
      "loss": 3.5467,
      "theoretical_loss": 4.959558620285274,
      "tokens_seen": 76152832
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004933901705115346,
      "loss": 3.4065,
      "theoretical_loss": 4.958944341662502,
      "tokens_seen": 76218368
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004933801404212638,
      "loss": 4.0583,
      "theoretical_loss": 4.958330738744172,
      "tokens_seen": 76283904
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000493370110330993,
      "loss": 3.7728,
      "theoretical_loss": 4.957717810207466,
      "tokens_seen": 76349440
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004933600802407221,
      "loss": 3.7919,
      "theoretical_loss": 4.957105554733289,
      "tokens_seen": 76414976
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004933500501504513,
      "loss": 3.9052,
      "theoretical_loss": 4.956493971006253,
      "tokens_seen": 76480512
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004933400200601805,
      "loss": 3.7208,
      "theoretical_loss": 4.955883057714669,
      "tokens_seen": 76546048
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004933299899699098,
      "loss": 3.7532,
      "theoretical_loss": 4.955272813550524,
      "tokens_seen": 76611584
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004933199598796389,
      "loss": 3.8077,
      "theoretical_loss": 4.954663237209477,
      "tokens_seen": 76677120
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004933099297893682,
      "loss": 3.8766,
      "theoretical_loss": 4.954054327390841,
      "tokens_seen": 76742656
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004932998996990972,
      "loss": 3.9885,
      "theoretical_loss": 4.9534460827975675,
      "tokens_seen": 76808192
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004932898696088265,
      "loss": 3.7149,
      "theoretical_loss": 4.952838502136241,
      "tokens_seen": 76873728
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004932798395185557,
      "loss": 3.8799,
      "theoretical_loss": 4.952231584117056,
      "tokens_seen": 76939264
    },
    {
      "epoch": 0.02,
      "objective/train/docs_used": 164442,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.9509365558624268,
      "objective/train/theoretical_loss": 4.951625327453812,
      "objective/train/tokens_used": 97464800,
      "theoretical_loss": 4.951625327453812,
      "tokens_seen": 77004800
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004932698094282849,
      "loss": 3.6636,
      "theoretical_loss": 4.951625327453812,
      "tokens_seen": 77004800
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004932597793380141,
      "loss": 3.8467,
      "theoretical_loss": 4.951019730863894,
      "tokens_seen": 77070336
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004932497492477432,
      "loss": 3.7791,
      "theoretical_loss": 4.950414793068266,
      "tokens_seen": 77135872
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004932397191574724,
      "loss": 3.6553,
      "theoretical_loss": 4.94981051279145,
      "tokens_seen": 77201408
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004932296890672016,
      "loss": 4.0068,
      "theoretical_loss": 4.94920688876152,
      "tokens_seen": 77266944
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004932196589769308,
      "loss": 3.4971,
      "theoretical_loss": 4.948603919710088,
      "tokens_seen": 77332480
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.00049320962888666,
      "loss": 3.5554,
      "theoretical_loss": 4.948001604372287,
      "tokens_seen": 77398016
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004931995987963893,
      "loss": 3.8427,
      "theoretical_loss": 4.947399941486762,
      "tokens_seen": 77463552
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004931895687061183,
      "loss": 3.7901,
      "theoretical_loss": 4.946798929795658,
      "tokens_seen": 77529088
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004931795386158476,
      "loss": 3.6006,
      "theoretical_loss": 4.946198568044602,
      "tokens_seen": 77594624
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004931695085255767,
      "loss": 3.7489,
      "theoretical_loss": 4.945598854982698,
      "tokens_seen": 77660160
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000493159478435306,
      "loss": 3.6273,
      "theoretical_loss": 4.944999789362508,
      "tokens_seen": 77725696
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004931494483450352,
      "loss": 3.9755,
      "theoretical_loss": 4.944401369940043,
      "tokens_seen": 77791232
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004931394182547643,
      "loss": 3.8917,
      "theoretical_loss": 4.9438035954747495,
      "tokens_seen": 77856768
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004931293881644935,
      "loss": 3.9238,
      "theoretical_loss": 4.9432064647294975,
      "tokens_seen": 77922304
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004931193580742227,
      "loss": 3.8091,
      "theoretical_loss": 4.942609976470566,
      "tokens_seen": 77987840
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004931093279839519,
      "loss": 3.6847,
      "theoretical_loss": 4.942014129467637,
      "tokens_seen": 78053376
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004930992978936811,
      "loss": 3.7321,
      "theoretical_loss": 4.941418922493774,
      "tokens_seen": 78118912
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004930892678034102,
      "loss": 3.9275,
      "theoretical_loss": 4.940824354325419,
      "tokens_seen": 78184448
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004930792377131394,
      "loss": 3.5937,
      "theoretical_loss": 4.940230423742372,
      "tokens_seen": 78249984
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004930692076228686,
      "loss": 3.9973,
      "theoretical_loss": 4.939637129527789,
      "tokens_seen": 78315520
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004930591775325978,
      "loss": 3.6964,
      "theoretical_loss": 4.939044470468156,
      "tokens_seen": 78381056
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000493049147442327,
      "loss": 3.8794,
      "theoretical_loss": 4.938452445353294,
      "tokens_seen": 78446592
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004930391173520562,
      "loss": 3.6823,
      "theoretical_loss": 4.937861052976332,
      "tokens_seen": 78512128
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004930290872617853,
      "loss": 4.0123,
      "theoretical_loss": 4.937270292133704,
      "tokens_seen": 78577664
    },
    {
      "epoch": 0.02,
      "objective/train/docs_used": 165209,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.9666755199432373,
      "objective/train/theoretical_loss": 4.9366801616251355,
      "objective/train/tokens_used": 99103200,
      "theoretical_loss": 4.9366801616251355,
      "tokens_seen": 78643200
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004930190571715146,
      "loss": 3.7987,
      "theoretical_loss": 4.9366801616251355,
      "tokens_seen": 78643200
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004930090270812437,
      "loss": 3.7133,
      "theoretical_loss": 4.93609066025363,
      "tokens_seen": 78708736
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000492998996990973,
      "loss": 3.8611,
      "theoretical_loss": 4.935501786825457,
      "tokens_seen": 78774272
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004929889669007021,
      "loss": 3.7493,
      "theoretical_loss": 4.934913540150143,
      "tokens_seen": 78839808
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004929789368104313,
      "loss": 3.8265,
      "theoretical_loss": 4.934325919040461,
      "tokens_seen": 78905344
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004929689067201605,
      "loss": 3.7146,
      "theoretical_loss": 4.933738922312413,
      "tokens_seen": 78970880
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004929588766298897,
      "loss": 3.8796,
      "theoretical_loss": 4.933152548785222,
      "tokens_seen": 79036416
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004929488465396189,
      "loss": 3.6061,
      "theoretical_loss": 4.932566797281324,
      "tokens_seen": 79101952
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004929388164493481,
      "loss": 3.8318,
      "theoretical_loss": 4.931981666626351,
      "tokens_seen": 79167488
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004929287863590773,
      "loss": 3.8867,
      "theoretical_loss": 4.931397155649121,
      "tokens_seen": 79233024
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004929187562688064,
      "loss": 3.6513,
      "theoretical_loss": 4.930813263181631,
      "tokens_seen": 79298560
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004929087261785356,
      "loss": 4.1337,
      "theoretical_loss": 4.93022998805904,
      "tokens_seen": 79364096
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004928986960882648,
      "loss": 3.7265,
      "theoretical_loss": 4.929647329119659,
      "tokens_seen": 79429632
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000492888665997994,
      "loss": 3.6387,
      "theoretical_loss": 4.9290652852049455,
      "tokens_seen": 79495168
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004928786359077232,
      "loss": 3.569,
      "theoretical_loss": 4.928483855159485,
      "tokens_seen": 79560704
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004928686058174523,
      "loss": 4.0239,
      "theoretical_loss": 4.927903037830983,
      "tokens_seen": 79626240
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004928585757271815,
      "loss": 3.9572,
      "theoretical_loss": 4.9273228320702565,
      "tokens_seen": 79691776
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004928485456369107,
      "loss": 3.7791,
      "theoretical_loss": 4.926743236731218,
      "tokens_seen": 79757312
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.00049283851554664,
      "loss": 3.8708,
      "theoretical_loss": 4.926164250670868,
      "tokens_seen": 79822848
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004928284854563691,
      "loss": 3.6422,
      "theoretical_loss": 4.925585872749284,
      "tokens_seen": 79888384
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004928184553660984,
      "loss": 3.9702,
      "theoretical_loss": 4.925008101829608,
      "tokens_seen": 79953920
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004928084252758274,
      "loss": 3.745,
      "theoretical_loss": 4.9244309367780374,
      "tokens_seen": 80019456
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004927983951855567,
      "loss": 3.6731,
      "theoretical_loss": 4.923854376463816,
      "tokens_seen": 80084992
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004927883650952859,
      "loss": 3.8668,
      "theoretical_loss": 4.923278419759217,
      "tokens_seen": 80150528
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004927783350050151,
      "loss": 3.8823,
      "theoretical_loss": 4.92270306553954,
      "tokens_seen": 80216064
    },
    {
      "epoch": 0.02,
      "objective/train/docs_used": 165995,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.573359489440918,
      "objective/train/theoretical_loss": 4.922128312683096,
      "objective/train/tokens_used": 100741600,
      "theoretical_loss": 4.922128312683096,
      "tokens_seen": 80281600
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004927683049147443,
      "loss": 3.6868,
      "theoretical_loss": 4.922128312683096,
      "tokens_seen": 80281600
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004927582748244734,
      "loss": 3.8374,
      "theoretical_loss": 4.921554160071194,
      "tokens_seen": 80347136
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004927482447342026,
      "loss": 3.6288,
      "theoretical_loss": 4.920980606588142,
      "tokens_seen": 80412672
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004927382146439318,
      "loss": 3.7754,
      "theoretical_loss": 4.920407651121222,
      "tokens_seen": 80478208
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000492728184553661,
      "loss": 3.869,
      "theoretical_loss": 4.919835292560689,
      "tokens_seen": 80543744
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004927181544633902,
      "loss": 3.8912,
      "theoretical_loss": 4.919263529799759,
      "tokens_seen": 80609280
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004927081243731193,
      "loss": 3.8343,
      "theoretical_loss": 4.918692361734598,
      "tokens_seen": 80674816
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004926980942828485,
      "loss": 3.8784,
      "theoretical_loss": 4.91812178726431,
      "tokens_seen": 80740352
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004926880641925777,
      "loss": 3.6742,
      "theoretical_loss": 4.917551805290929,
      "tokens_seen": 80805888
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004926780341023069,
      "loss": 3.7316,
      "theoretical_loss": 4.916982414719408,
      "tokens_seen": 80871424
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004926680040120361,
      "loss": 3.9762,
      "theoretical_loss": 4.9164136144576105,
      "tokens_seen": 80936960
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004926579739217654,
      "loss": 3.5511,
      "theoretical_loss": 4.915845403416299,
      "tokens_seen": 81002496
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004926479438314944,
      "loss": 4.2042,
      "theoretical_loss": 4.915277780509124,
      "tokens_seen": 81068032
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004926379137412237,
      "loss": 3.7369,
      "theoretical_loss": 4.914710744652614,
      "tokens_seen": 81133568
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004926278836509528,
      "loss": 3.7712,
      "theoretical_loss": 4.914144294766169,
      "tokens_seen": 81199104
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004926178535606821,
      "loss": 3.7433,
      "theoretical_loss": 4.913578429772047,
      "tokens_seen": 81264640
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004926078234704113,
      "loss": 3.6981,
      "theoretical_loss": 4.913013148595355,
      "tokens_seen": 81330176
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004925977933801404,
      "loss": 3.6646,
      "theoretical_loss": 4.912448450164041,
      "tokens_seen": 81395712
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004925877632898696,
      "loss": 3.775,
      "theoretical_loss": 4.91188433340888,
      "tokens_seen": 81461248
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004925777331995988,
      "loss": 3.7487,
      "theoretical_loss": 4.911320797263471,
      "tokens_seen": 81526784
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000492567703109328,
      "loss": 3.6504,
      "theoretical_loss": 4.910757840664219,
      "tokens_seen": 81592320
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004925576730190572,
      "loss": 3.6375,
      "theoretical_loss": 4.910195462550334,
      "tokens_seen": 81657856
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004925476429287864,
      "loss": 3.6929,
      "theoretical_loss": 4.909633661863811,
      "tokens_seen": 81723392
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004925376128385155,
      "loss": 3.7826,
      "theoretical_loss": 4.909072437549434,
      "tokens_seen": 81788928
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004925275827482447,
      "loss": 3.561,
      "theoretical_loss": 4.908511788554753,
      "tokens_seen": 81854464
    },
    {
      "epoch": 0.02,
      "objective/train/docs_used": 167243,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 4.2206902503967285,
      "objective/train/theoretical_loss": 4.907951713830082,
      "objective/train/tokens_used": 102380000,
      "theoretical_loss": 4.907951713830082,
      "tokens_seen": 81920000
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004925175526579739,
      "loss": 4.1068,
      "theoretical_loss": 4.907951713830082,
      "tokens_seen": 81920000
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004925075225677031,
      "loss": 3.8568,
      "theoretical_loss": 4.907392212328489,
      "tokens_seen": 81985536
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004924974924774323,
      "loss": 3.5104,
      "theoretical_loss": 4.906833283005785,
      "tokens_seen": 82051072
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004924874623871615,
      "loss": 3.5485,
      "theoretical_loss": 4.906274924820515,
      "tokens_seen": 82116608
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004924774322968907,
      "loss": 4.0778,
      "theoretical_loss": 4.90571713673395,
      "tokens_seen": 82182144
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004924674022066199,
      "loss": 3.9291,
      "theoretical_loss": 4.905159917710073,
      "tokens_seen": 82247680
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004924573721163491,
      "loss": 3.6175,
      "theoretical_loss": 4.904603266715578,
      "tokens_seen": 82313216
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004924473420260783,
      "loss": 3.8755,
      "theoretical_loss": 4.904047182719854,
      "tokens_seen": 82378752
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004924373119358075,
      "loss": 3.7729,
      "theoretical_loss": 4.903491664694977,
      "tokens_seen": 82444288
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004924272818455366,
      "loss": 3.8916,
      "theoretical_loss": 4.902936711615702,
      "tokens_seen": 82509824
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004924172517552658,
      "loss": 3.7129,
      "theoretical_loss": 4.902382322459456,
      "tokens_seen": 82575360
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000492407221664995,
      "loss": 3.5962,
      "theoretical_loss": 4.901828496206322,
      "tokens_seen": 82640896
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004923971915747242,
      "loss": 3.7346,
      "theoretical_loss": 4.90127523183904,
      "tokens_seen": 82706432
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004923871614844534,
      "loss": 3.8739,
      "theoretical_loss": 4.900722528342988,
      "tokens_seen": 82771968
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004923771313941825,
      "loss": 3.7026,
      "theoretical_loss": 4.900170384706181,
      "tokens_seen": 82837504
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004923671013039117,
      "loss": 3.6768,
      "theoretical_loss": 4.899618799919256,
      "tokens_seen": 82903040
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004923570712136409,
      "loss": 3.5168,
      "theoretical_loss": 4.899067772975469,
      "tokens_seen": 82968576
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004923470411233702,
      "loss": 3.357,
      "theoretical_loss": 4.898517302870679,
      "tokens_seen": 83034112
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004923370110330993,
      "loss": 3.6077,
      "theoretical_loss": 4.897967388603346,
      "tokens_seen": 83099648
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004923269809428286,
      "loss": 3.879,
      "theoretical_loss": 4.897418029174519,
      "tokens_seen": 83165184
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004923169508525576,
      "loss": 3.5125,
      "theoretical_loss": 4.896869223587828,
      "tokens_seen": 83230720
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004923069207622869,
      "loss": 3.7582,
      "theoretical_loss": 4.896320970849472,
      "tokens_seen": 83296256
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004922968906720161,
      "loss": 3.9325,
      "theoretical_loss": 4.895773269968219,
      "tokens_seen": 83361792
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004922868605817453,
      "loss": 3.6416,
      "theoretical_loss": 4.895226119955386,
      "tokens_seen": 83427328
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004922768304914745,
      "loss": 3.7782,
      "theoretical_loss": 4.894679519824841,
      "tokens_seen": 83492864
    },
    {
      "epoch": 0.03,
      "objective/train/docs_used": 167845,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.8555917739868164,
      "objective/train/theoretical_loss": 4.894133468592984,
      "objective/train/tokens_used": 104018400,
      "theoretical_loss": 4.894133468592984,
      "tokens_seen": 83558400
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004922668004012036,
      "loss": 3.9438,
      "theoretical_loss": 4.894133468592984,
      "tokens_seen": 83558400
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004922567703109328,
      "loss": 3.5061,
      "theoretical_loss": 4.8935879652787495,
      "tokens_seen": 83623936
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000492246740220662,
      "loss": 3.6214,
      "theoretical_loss": 4.893043008903591,
      "tokens_seen": 83689472
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004922367101303912,
      "loss": 3.9424,
      "theoretical_loss": 4.892498598491473,
      "tokens_seen": 83755008
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004922266800401204,
      "loss": 3.4731,
      "theoretical_loss": 4.891954733068863,
      "tokens_seen": 83820544
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004922166499498495,
      "loss": 3.7306,
      "theoretical_loss": 4.891411411664727,
      "tokens_seen": 83886080
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004922066198595787,
      "loss": 3.8514,
      "theoretical_loss": 4.890868633310515,
      "tokens_seen": 83951616
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004921965897693079,
      "loss": 3.7332,
      "theoretical_loss": 4.890326397040158,
      "tokens_seen": 84017152
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004921865596790371,
      "loss": 3.6843,
      "theoretical_loss": 4.889784701890056,
      "tokens_seen": 84082688
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004921765295887663,
      "loss": 3.5876,
      "theoretical_loss": 4.8892435468990705,
      "tokens_seen": 84148224
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004921664994984956,
      "loss": 3.6961,
      "theoretical_loss": 4.88870293110852,
      "tokens_seen": 84213760
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004921564694082246,
      "loss": 3.6685,
      "theoretical_loss": 4.888162853562166,
      "tokens_seen": 84279296
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004921464393179539,
      "loss": 3.4346,
      "theoretical_loss": 4.88762331330621,
      "tokens_seen": 84344832
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000492136409227683,
      "loss": 3.5563,
      "theoretical_loss": 4.88708430938928,
      "tokens_seen": 84410368
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004921263791374123,
      "loss": 3.7265,
      "theoretical_loss": 4.8865458408624285,
      "tokens_seen": 84475904
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004921163490471415,
      "loss": 3.5372,
      "theoretical_loss": 4.8860079067791204,
      "tokens_seen": 84541440
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004921063189568706,
      "loss": 3.4829,
      "theoretical_loss": 4.885470506195227,
      "tokens_seen": 84606976
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004920962888665998,
      "loss": 3.533,
      "theoretical_loss": 4.884933638169014,
      "tokens_seen": 84672512
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000492086258776329,
      "loss": 3.9744,
      "theoretical_loss": 4.88439730176114,
      "tokens_seen": 84738048
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004920762286860582,
      "loss": 3.674,
      "theoretical_loss": 4.883861496034644,
      "tokens_seen": 84803584
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004920661985957874,
      "loss": 3.9751,
      "theoretical_loss": 4.88332622005494,
      "tokens_seen": 84869120
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004920561685055166,
      "loss": 3.6296,
      "theoretical_loss": 4.8827914728898065,
      "tokens_seen": 84934656
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004920461384152457,
      "loss": 3.6182,
      "theoretical_loss": 4.88225725360938,
      "tokens_seen": 85000192
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004920361083249749,
      "loss": 3.7477,
      "theoretical_loss": 4.881723561286149,
      "tokens_seen": 85065728
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004920260782347041,
      "loss": 3.6688,
      "theoretical_loss": 4.881190394994943,
      "tokens_seen": 85131264
    },
    {
      "epoch": 0.03,
      "objective/train/docs_used": 169231,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 4.038014888763428,
      "objective/train/theoretical_loss": 4.880657753812926,
      "objective/train/tokens_used": 105656800,
      "theoretical_loss": 4.880657753812926,
      "tokens_seen": 85196800
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004920160481444333,
      "loss": 3.8814,
      "theoretical_loss": 4.880657753812926,
      "tokens_seen": 85196800
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004920060180541625,
      "loss": 3.6396,
      "theoretical_loss": 4.880125636819594,
      "tokens_seen": 85262336
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004919959879638916,
      "loss": 3.7983,
      "theoretical_loss": 4.879594043096755,
      "tokens_seen": 85327872
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004919859578736209,
      "loss": 3.8557,
      "theoretical_loss": 4.879062971728534,
      "tokens_seen": 85393408
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00049197592778335,
      "loss": 3.5785,
      "theoretical_loss": 4.87853242180136,
      "tokens_seen": 85458944
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004919658976930793,
      "loss": 3.4825,
      "theoretical_loss": 4.878002392403959,
      "tokens_seen": 85524480
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004919558676028084,
      "loss": 3.5934,
      "theoretical_loss": 4.877472882627343,
      "tokens_seen": 85590016
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004919458375125377,
      "loss": 3.7652,
      "theoretical_loss": 4.8769438915648085,
      "tokens_seen": 85655552
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004919358074222668,
      "loss": 3.5099,
      "theoretical_loss": 4.876415418311928,
      "tokens_seen": 85721088
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000491925777331996,
      "loss": 3.4595,
      "theoretical_loss": 4.875887461966537,
      "tokens_seen": 85786624
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004919157472417252,
      "loss": 3.7013,
      "theoretical_loss": 4.875360021628733,
      "tokens_seen": 85852160
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004919057171514544,
      "loss": 3.8989,
      "theoretical_loss": 4.874833096400865,
      "tokens_seen": 85917696
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004918956870611836,
      "loss": 3.8334,
      "theoretical_loss": 4.874306685387525,
      "tokens_seen": 85983232
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004918856569709127,
      "loss": 3.4766,
      "theoretical_loss": 4.873780787695547,
      "tokens_seen": 86048768
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004918756268806419,
      "loss": 3.6543,
      "theoretical_loss": 4.87325540243399,
      "tokens_seen": 86114304
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004918655967903711,
      "loss": 3.5469,
      "theoretical_loss": 4.872730528714139,
      "tokens_seen": 86179840
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004918555667001003,
      "loss": 3.7906,
      "theoretical_loss": 4.872206165649493,
      "tokens_seen": 86245376
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004918455366098295,
      "loss": 3.7537,
      "theoretical_loss": 4.871682312355761,
      "tokens_seen": 86310912
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004918355065195586,
      "loss": 3.8324,
      "theoretical_loss": 4.871158967950852,
      "tokens_seen": 86376448
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004918254764292878,
      "loss": 3.5791,
      "theoretical_loss": 4.870636131554869,
      "tokens_seen": 86441984
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000491815446339017,
      "loss": 3.6966,
      "theoretical_loss": 4.8701138022901045,
      "tokens_seen": 86507520
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004918054162487463,
      "loss": 3.89,
      "theoretical_loss": 4.869591979281028,
      "tokens_seen": 86573056
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004917953861584754,
      "loss": 3.9098,
      "theoretical_loss": 4.8690706616542805,
      "tokens_seen": 86638592
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004917853560682047,
      "loss": 3.6221,
      "theoretical_loss": 4.868549848538675,
      "tokens_seen": 86704128
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004917753259779337,
      "loss": 3.7127,
      "theoretical_loss": 4.868029539065176,
      "tokens_seen": 86769664
    },
    {
      "epoch": 0.03,
      "objective/train/docs_used": 169899,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.6036012172698975,
      "objective/train/theoretical_loss": 4.867509732366907,
      "objective/train/tokens_used": 107295200,
      "theoretical_loss": 4.867509732366907,
      "tokens_seen": 86835200
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000491765295887663,
      "loss": 3.6755,
      "theoretical_loss": 4.867509732366907,
      "tokens_seen": 86835200
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004917552657973922,
      "loss": 3.8447,
      "theoretical_loss": 4.866990427579129,
      "tokens_seen": 86900736
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004917452357071214,
      "loss": 3.8378,
      "theoretical_loss": 4.866471623839248,
      "tokens_seen": 86966272
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004917352056168506,
      "loss": 3.7024,
      "theoretical_loss": 4.8659533202867955,
      "tokens_seen": 87031808
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004917251755265797,
      "loss": 3.7791,
      "theoretical_loss": 4.86543551606343,
      "tokens_seen": 87097344
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004917151454363089,
      "loss": 3.7572,
      "theoretical_loss": 4.864918210312927,
      "tokens_seen": 87162880
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004917051153460381,
      "loss": 3.4891,
      "theoretical_loss": 4.864401402181173,
      "tokens_seen": 87228416
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004916950852557673,
      "loss": 3.7334,
      "theoretical_loss": 4.863885090816158,
      "tokens_seen": 87293952
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004916850551654965,
      "loss": 3.818,
      "theoretical_loss": 4.863369275367968,
      "tokens_seen": 87359488
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004916750250752258,
      "loss": 3.6889,
      "theoretical_loss": 4.862853954988781,
      "tokens_seen": 87425024
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004916649949849548,
      "loss": 3.6571,
      "theoretical_loss": 4.862339128832857,
      "tokens_seen": 87490560
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004916549648946841,
      "loss": 3.6037,
      "theoretical_loss": 4.861824796056533,
      "tokens_seen": 87556096
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004916449348044132,
      "loss": 4.1055,
      "theoretical_loss": 4.861310955818219,
      "tokens_seen": 87621632
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004916349047141425,
      "loss": 3.9186,
      "theoretical_loss": 4.860797607278385,
      "tokens_seen": 87687168
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004916248746238717,
      "loss": 3.6732,
      "theoretical_loss": 4.86028474959956,
      "tokens_seen": 87752704
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004916148445336008,
      "loss": 3.6389,
      "theoretical_loss": 4.859772381946323,
      "tokens_seen": 87818240
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00049160481444333,
      "loss": 3.4738,
      "theoretical_loss": 4.859260503485298,
      "tokens_seen": 87883776
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004915947843530592,
      "loss": 3.6106,
      "theoretical_loss": 4.858749113385144,
      "tokens_seen": 87949312
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004915847542627884,
      "loss": 3.8513,
      "theoretical_loss": 4.858238210816554,
      "tokens_seen": 88014848
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004915747241725176,
      "loss": 3.5283,
      "theoretical_loss": 4.8577277949522415,
      "tokens_seen": 88080384
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004915646940822468,
      "loss": 3.6318,
      "theoretical_loss": 4.857217864966943,
      "tokens_seen": 88145920
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004915546639919759,
      "loss": 3.9574,
      "theoretical_loss": 4.856708420037402,
      "tokens_seen": 88211456
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004915446339017051,
      "loss": 3.6005,
      "theoretical_loss": 4.8561994593423705,
      "tokens_seen": 88276992
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004915346038114343,
      "loss": 3.6561,
      "theoretical_loss": 4.8556909820625975,
      "tokens_seen": 88342528
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004915245737211635,
      "loss": 3.5624,
      "theoretical_loss": 4.855182987380823,
      "tokens_seen": 88408064
    },
    {
      "epoch": 0.03,
      "objective/train/docs_used": 170786,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.6651480197906494,
      "objective/train/theoretical_loss": 4.854675474481779,
      "objective/train/tokens_used": 108933600,
      "theoretical_loss": 4.854675474481779,
      "tokens_seen": 88473600
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004915145436308927,
      "loss": 3.5245,
      "theoretical_loss": 4.854675474481779,
      "tokens_seen": 88473600
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004915045135406218,
      "loss": 3.9545,
      "theoretical_loss": 4.8541684425521705,
      "tokens_seen": 88539136
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004914944834503511,
      "loss": 3.7957,
      "theoretical_loss": 4.85366189078068,
      "tokens_seen": 88604672
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004914844533600802,
      "loss": 3.9079,
      "theoretical_loss": 4.853155818357957,
      "tokens_seen": 88670208
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004914744232698095,
      "loss": 3.9184,
      "theoretical_loss": 4.852650224476609,
      "tokens_seen": 88735744
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004914643931795386,
      "loss": 3.731,
      "theoretical_loss": 4.852145108331205,
      "tokens_seen": 88801280
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004914543630892679,
      "loss": 3.7566,
      "theoretical_loss": 4.851640469118255,
      "tokens_seen": 88866816
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000491444332998997,
      "loss": 3.6482,
      "theoretical_loss": 4.851136306036219,
      "tokens_seen": 88932352
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004914343029087262,
      "loss": 3.7222,
      "theoretical_loss": 4.850632618285486,
      "tokens_seen": 88997888
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004914242728184554,
      "loss": 3.6696,
      "theoretical_loss": 4.850129405068383,
      "tokens_seen": 89063424
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004914142427281846,
      "loss": 3.8089,
      "theoretical_loss": 4.849626665589156,
      "tokens_seen": 89128960
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004914042126379138,
      "loss": 3.5295,
      "theoretical_loss": 4.849124399053969,
      "tokens_seen": 89194496
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004913941825476429,
      "loss": 3.6736,
      "theoretical_loss": 4.8486226046709024,
      "tokens_seen": 89260032
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004913841524573721,
      "loss": 3.5729,
      "theoretical_loss": 4.8481212816499415,
      "tokens_seen": 89325568
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004913741223671013,
      "loss": 3.7998,
      "theoretical_loss": 4.847620429202967,
      "tokens_seen": 89391104
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004913640922768305,
      "loss": 3.4795,
      "theoretical_loss": 4.847120046543763,
      "tokens_seen": 89456640
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004913540621865597,
      "loss": 3.8715,
      "theoretical_loss": 4.846620132887992,
      "tokens_seen": 89522176
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004913440320962888,
      "loss": 3.7386,
      "theoretical_loss": 4.8461206874532055,
      "tokens_seen": 89587712
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000491334002006018,
      "loss": 3.6659,
      "theoretical_loss": 4.845621709458831,
      "tokens_seen": 89653248
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004913239719157472,
      "loss": 3.4551,
      "theoretical_loss": 4.845123198126162,
      "tokens_seen": 89718784
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004913139418254765,
      "loss": 3.6237,
      "theoretical_loss": 4.844625152678364,
      "tokens_seen": 89784320
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004913039117352056,
      "loss": 3.8513,
      "theoretical_loss": 4.844127572340455,
      "tokens_seen": 89849856
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004912938816449349,
      "loss": 3.6724,
      "theoretical_loss": 4.84363045633931,
      "tokens_seen": 89915392
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004912838515546639,
      "loss": 3.8743,
      "theoretical_loss": 4.843133803903651,
      "tokens_seen": 89980928
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004912738214643932,
      "loss": 3.6873,
      "theoretical_loss": 4.84263761426404,
      "tokens_seen": 90046464
    },
    {
      "epoch": 0.03,
      "objective/train/docs_used": 171435,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.2105295658111572,
      "objective/train/theoretical_loss": 4.842141886652876,
      "objective/train/tokens_used": 110572000,
      "theoretical_loss": 4.842141886652876,
      "tokens_seen": 90112000
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004912637913741224,
      "loss": 3.2592,
      "theoretical_loss": 4.842141886652876,
      "tokens_seen": 90112000
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004912537612838516,
      "loss": 3.6091,
      "theoretical_loss": 4.841646620304388,
      "tokens_seen": 90177536
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004912437311935808,
      "loss": 3.7829,
      "theoretical_loss": 4.841151814454632,
      "tokens_seen": 90243072
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00049123370110331,
      "loss": 3.8082,
      "theoretical_loss": 4.840657468341476,
      "tokens_seen": 90308608
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004912236710130391,
      "loss": 3.5862,
      "theoretical_loss": 4.84016358120461,
      "tokens_seen": 90374144
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004912136409227683,
      "loss": 3.523,
      "theoretical_loss": 4.839670152285526,
      "tokens_seen": 90439680
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004912036108324975,
      "loss": 3.3392,
      "theoretical_loss": 4.8391771808275195,
      "tokens_seen": 90505216
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004911935807422267,
      "loss": 3.6086,
      "theoretical_loss": 4.838684666075682,
      "tokens_seen": 90570752
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004911835506519559,
      "loss": 3.7421,
      "theoretical_loss": 4.838192607276896,
      "tokens_seen": 90636288
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000491173520561685,
      "loss": 3.549,
      "theoretical_loss": 4.837701003679829,
      "tokens_seen": 90701824
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004911634904714142,
      "loss": 3.6988,
      "theoretical_loss": 4.8372098545349305,
      "tokens_seen": 90767360
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004911534603811434,
      "loss": 3.5152,
      "theoretical_loss": 4.836719159094422,
      "tokens_seen": 90832896
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004911434302908726,
      "loss": 3.5065,
      "theoretical_loss": 4.836228916612292,
      "tokens_seen": 90898432
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004911334002006019,
      "loss": 3.8109,
      "theoretical_loss": 4.835739126344298,
      "tokens_seen": 90963968
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004911233701103309,
      "loss": 3.9112,
      "theoretical_loss": 4.8352497875479505,
      "tokens_seen": 91029504
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004911133400200602,
      "loss": 3.8912,
      "theoretical_loss": 4.834760899482514,
      "tokens_seen": 91095040
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004911033099297893,
      "loss": 3.9841,
      "theoretical_loss": 4.834272461409001,
      "tokens_seen": 91160576
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004910932798395186,
      "loss": 3.6817,
      "theoretical_loss": 4.833784472590165,
      "tokens_seen": 91226112
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004910832497492478,
      "loss": 3.795,
      "theoretical_loss": 4.833296932290495,
      "tokens_seen": 91291648
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000491073219658977,
      "loss": 3.8488,
      "theoretical_loss": 4.832809839776213,
      "tokens_seen": 91357184
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004910631895687061,
      "loss": 4.0801,
      "theoretical_loss": 4.832323194315265,
      "tokens_seen": 91422720
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004910531594784353,
      "loss": 3.8667,
      "theoretical_loss": 4.831836995177319,
      "tokens_seen": 91488256
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004910431293881645,
      "loss": 3.849,
      "theoretical_loss": 4.831351241633756,
      "tokens_seen": 91553792
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004910330992978937,
      "loss": 3.8149,
      "theoretical_loss": 4.8308659329576695,
      "tokens_seen": 91619328
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004910230692076229,
      "loss": 3.7799,
      "theoretical_loss": 4.830381068423856,
      "tokens_seen": 91684864
    },
    {
      "epoch": 0.03,
      "objective/train/docs_used": 171740,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.2981529235839844,
      "objective/train/theoretical_loss": 4.8298966473088125,
      "objective/train/tokens_used": 112210400,
      "theoretical_loss": 4.8298966473088125,
      "tokens_seen": 91750400
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000491013039117352,
      "loss": 3.4557,
      "theoretical_loss": 4.8298966473088125,
      "tokens_seen": 91750400
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004910030090270812,
      "loss": 4.0207,
      "theoretical_loss": 4.829412668890729,
      "tokens_seen": 91815936
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004909929789368104,
      "loss": 3.8245,
      "theoretical_loss": 4.8289291324494865,
      "tokens_seen": 91881472
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004909829488465397,
      "loss": 3.7635,
      "theoretical_loss": 4.828446037266647,
      "tokens_seen": 91947008
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004909729187562688,
      "loss": 3.6794,
      "theoretical_loss": 4.827963382625454,
      "tokens_seen": 92012544
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004909628886659981,
      "loss": 3.7506,
      "theoretical_loss": 4.827481167810825,
      "tokens_seen": 92078080
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004909528585757272,
      "loss": 4.0619,
      "theoretical_loss": 4.826999392109344,
      "tokens_seen": 92143616
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004909428284854564,
      "loss": 3.8418,
      "theoretical_loss": 4.826518054809259,
      "tokens_seen": 92209152
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004909327983951856,
      "loss": 3.809,
      "theoretical_loss": 4.826037155200478,
      "tokens_seen": 92274688
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004909227683049148,
      "loss": 3.9797,
      "theoretical_loss": 4.825556692574562,
      "tokens_seen": 92340224
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000490912738214644,
      "loss": 3.8061,
      "theoretical_loss": 4.825076666224717,
      "tokens_seen": 92405760
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004909027081243731,
      "loss": 3.7777,
      "theoretical_loss": 4.824597075445799,
      "tokens_seen": 92471296
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004908926780341023,
      "loss": 3.8168,
      "theoretical_loss": 4.824117919534297,
      "tokens_seen": 92536832
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004908826479438315,
      "loss": 3.8435,
      "theoretical_loss": 4.823639197788334,
      "tokens_seen": 92602368
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004908726178535607,
      "loss": 3.5213,
      "theoretical_loss": 4.823160909507665,
      "tokens_seen": 92667904
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004908625877632899,
      "loss": 3.6417,
      "theoretical_loss": 4.822683053993664,
      "tokens_seen": 92733440
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000490852557673019,
      "loss": 3.4858,
      "theoretical_loss": 4.822205630549329,
      "tokens_seen": 92798976
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004908425275827482,
      "loss": 3.7943,
      "theoretical_loss": 4.821728638479267,
      "tokens_seen": 92864512
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004908324974924774,
      "loss": 3.7895,
      "theoretical_loss": 4.821252077089696,
      "tokens_seen": 92930048
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004908224674022067,
      "loss": 3.753,
      "theoretical_loss": 4.820775945688437,
      "tokens_seen": 92995584
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004908124373119358,
      "loss": 3.9838,
      "theoretical_loss": 4.820300243584913,
      "tokens_seen": 93061120
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004908024072216651,
      "loss": 3.8889,
      "theoretical_loss": 4.819824970090138,
      "tokens_seen": 93126656
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004907923771313941,
      "loss": 3.8958,
      "theoretical_loss": 4.819350124516717,
      "tokens_seen": 93192192
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004907823470411234,
      "loss": 3.6089,
      "theoretical_loss": 4.818875706178841,
      "tokens_seen": 93257728
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004907723169508526,
      "loss": 3.9225,
      "theoretical_loss": 4.818401714392279,
      "tokens_seen": 93323264
    },
    {
      "epoch": 0.03,
      "objective/train/docs_used": 171740,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.6480627059936523,
      "objective/train/theoretical_loss": 4.817928148474378,
      "objective/train/tokens_used": 113848800,
      "theoretical_loss": 4.817928148474378,
      "tokens_seen": 93388800
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004907622868605818,
      "loss": 3.544,
      "theoretical_loss": 4.817928148474378,
      "tokens_seen": 93388800
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000490752256770311,
      "loss": 3.8504,
      "theoretical_loss": 4.817455007744052,
      "tokens_seen": 93454336
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004907422266800401,
      "loss": 3.6831,
      "theoretical_loss": 4.816982291521785,
      "tokens_seen": 93519872
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004907321965897693,
      "loss": 3.7159,
      "theoretical_loss": 4.816509999129618,
      "tokens_seen": 93585408
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004907221664994985,
      "loss": 3.8475,
      "theoretical_loss": 4.816038129891151,
      "tokens_seen": 93650944
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004907121364092277,
      "loss": 3.867,
      "theoretical_loss": 4.815566683131536,
      "tokens_seen": 93716480
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004907021063189569,
      "loss": 3.7402,
      "theoretical_loss": 4.815095658177472,
      "tokens_seen": 93782016
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000490692076228686,
      "loss": 3.6066,
      "theoretical_loss": 4.814625054357199,
      "tokens_seen": 93847552
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004906820461384152,
      "loss": 3.7494,
      "theoretical_loss": 4.814154871000497,
      "tokens_seen": 93913088
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004906720160481444,
      "loss": 3.996,
      "theoretical_loss": 4.813685107438679,
      "tokens_seen": 93978624
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004906619859578736,
      "loss": 3.7818,
      "theoretical_loss": 4.813215763004585,
      "tokens_seen": 94044160
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004906519558676028,
      "loss": 3.8077,
      "theoretical_loss": 4.812746837032582,
      "tokens_seen": 94109696
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004906419257773321,
      "loss": 3.8845,
      "theoretical_loss": 4.812278328858554,
      "tokens_seen": 94175232
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004906318956870611,
      "loss": 3.8249,
      "theoretical_loss": 4.811810237819904,
      "tokens_seen": 94240768
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004906218655967904,
      "loss": 3.8511,
      "theoretical_loss": 4.81134256325554,
      "tokens_seen": 94306304
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004906118355065195,
      "loss": 3.5025,
      "theoretical_loss": 4.810875304505881,
      "tokens_seen": 94371840
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004906018054162488,
      "loss": 3.9814,
      "theoretical_loss": 4.810408460912846,
      "tokens_seen": 94437376
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000490591775325978,
      "loss": 3.7851,
      "theoretical_loss": 4.809942031819853,
      "tokens_seen": 94502912
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004905817452357072,
      "loss": 3.7676,
      "theoretical_loss": 4.809476016571809,
      "tokens_seen": 94568448
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004905717151454363,
      "loss": 3.8481,
      "theoretical_loss": 4.809010414515113,
      "tokens_seen": 94633984
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004905616850551655,
      "loss": 3.8722,
      "theoretical_loss": 4.808545224997644,
      "tokens_seen": 94699520
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004905516549648947,
      "loss": 4.0819,
      "theoretical_loss": 4.808080447368766,
      "tokens_seen": 94765056
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004905416248746239,
      "loss": 3.7927,
      "theoretical_loss": 4.807616080979315,
      "tokens_seen": 94830592
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004905315947843531,
      "loss": 3.8727,
      "theoretical_loss": 4.807152125181597,
      "tokens_seen": 94896128
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004905215646940822,
      "loss": 3.7942,
      "theoretical_loss": 4.806688579329387,
      "tokens_seen": 94961664
    },
    {
      "epoch": 0.03,
      "objective/train/docs_used": 173310,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.714308500289917,
      "objective/train/theoretical_loss": 4.8062254427779205,
      "objective/train/tokens_used": 115487200,
      "theoretical_loss": 4.8062254427779205,
      "tokens_seen": 95027200
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004905115346038114,
      "loss": 3.8544,
      "theoretical_loss": 4.8062254427779205,
      "tokens_seen": 95027200
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004905015045135406,
      "loss": 3.7483,
      "theoretical_loss": 4.80576271488389,
      "tokens_seen": 95092736
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004904914744232698,
      "loss": 3.8588,
      "theoretical_loss": 4.805300395005444,
      "tokens_seen": 95158272
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000490481444332999,
      "loss": 4.0297,
      "theoretical_loss": 4.804838482502181,
      "tokens_seen": 95223808
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004904714142427281,
      "loss": 3.9291,
      "theoretical_loss": 4.8043769767351385,
      "tokens_seen": 95289344
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004904613841524574,
      "loss": 3.7934,
      "theoretical_loss": 4.8039158770668005,
      "tokens_seen": 95354880
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004904513540621865,
      "loss": 3.911,
      "theoretical_loss": 4.803455182861087,
      "tokens_seen": 95420416
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004904413239719158,
      "loss": 3.6903,
      "theoretical_loss": 4.802994893483348,
      "tokens_seen": 95485952
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004904312938816449,
      "loss": 3.3024,
      "theoretical_loss": 4.802535008300364,
      "tokens_seen": 95551488
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004904212637913742,
      "loss": 3.9319,
      "theoretical_loss": 4.802075526680335,
      "tokens_seen": 95617024
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004904112337011033,
      "loss": 3.905,
      "theoretical_loss": 4.801616447992888,
      "tokens_seen": 95682560
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004904012036108325,
      "loss": 3.6539,
      "theoretical_loss": 4.801157771609061,
      "tokens_seen": 95748096
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004903911735205617,
      "loss": 3.9673,
      "theoretical_loss": 4.8006994969013,
      "tokens_seen": 95813632
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004903811434302909,
      "loss": 3.9648,
      "theoretical_loss": 4.800241623243467,
      "tokens_seen": 95879168
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004903711133400201,
      "loss": 3.6807,
      "theoretical_loss": 4.799784150010819,
      "tokens_seen": 95944704
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004903610832497492,
      "loss": 3.5081,
      "theoretical_loss": 4.799327076580017,
      "tokens_seen": 96010240
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004903510531594784,
      "loss": 3.9447,
      "theoretical_loss": 4.798870402329115,
      "tokens_seen": 96075776
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004903410230692076,
      "loss": 3.8515,
      "theoretical_loss": 4.798414126637558,
      "tokens_seen": 96141312
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004903309929789368,
      "loss": 3.6779,
      "theoretical_loss": 4.797958248886179,
      "tokens_seen": 96206848
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000490320962888666,
      "loss": 3.7298,
      "theoretical_loss": 4.797502768457193,
      "tokens_seen": 96272384
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004903109327983952,
      "loss": 3.9204,
      "theoretical_loss": 4.797047684734192,
      "tokens_seen": 96337920
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004903009027081243,
      "loss": 3.7826,
      "theoretical_loss": 4.796592997102147,
      "tokens_seen": 96403456
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004902908726178535,
      "loss": 3.9417,
      "theoretical_loss": 4.796138704947397,
      "tokens_seen": 96468992
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004902808425275828,
      "loss": 3.7212,
      "theoretical_loss": 4.795684807657649,
      "tokens_seen": 96534528
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004902708124373119,
      "loss": 3.97,
      "theoretical_loss": 4.795231304621968,
      "tokens_seen": 96600064
    },
    {
      "epoch": 0.03,
      "objective/train/docs_used": 173936,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.69027042388916,
      "objective/train/theoretical_loss": 4.794778195230787,
      "objective/train/tokens_used": 117125600,
      "theoretical_loss": 4.794778195230787,
      "tokens_seen": 96665600
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004902607823470412,
      "loss": 3.7908,
      "theoretical_loss": 4.794778195230787,
      "tokens_seen": 96665600
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004902507522567703,
      "loss": 3.6508,
      "theoretical_loss": 4.794325478875885,
      "tokens_seen": 96731136
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004902407221664995,
      "loss": 3.8929,
      "theoretical_loss": 4.793873154950399,
      "tokens_seen": 96796672
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004902306920762287,
      "loss": 3.9375,
      "theoretical_loss": 4.793421222848808,
      "tokens_seen": 96862208
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004902206619859579,
      "loss": 3.9788,
      "theoretical_loss": 4.7929696819669365,
      "tokens_seen": 96927744
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004902106318956871,
      "loss": 3.9239,
      "theoretical_loss": 4.792518531701948,
      "tokens_seen": 96993280
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004902006018054163,
      "loss": 3.8721,
      "theoretical_loss": 4.792067771452341,
      "tokens_seen": 97058816
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004901905717151454,
      "loss": 3.6114,
      "theoretical_loss": 4.791617400617948,
      "tokens_seen": 97124352
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004901805416248746,
      "loss": 3.7371,
      "theoretical_loss": 4.791167418599925,
      "tokens_seen": 97189888
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004901705115346038,
      "loss": 3.74,
      "theoretical_loss": 4.790717824800755,
      "tokens_seen": 97255424
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000490160481444333,
      "loss": 3.9032,
      "theoretical_loss": 4.790268618624239,
      "tokens_seen": 97320960
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004901504513540623,
      "loss": 3.6149,
      "theoretical_loss": 4.789819799475499,
      "tokens_seen": 97386496
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004901404212637913,
      "loss": 3.8858,
      "theoretical_loss": 4.789371366760961,
      "tokens_seen": 97452032
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004901303911735206,
      "loss": 3.722,
      "theoretical_loss": 4.788923319888369,
      "tokens_seen": 97517568
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004901203610832497,
      "loss": 3.6514,
      "theoretical_loss": 4.788475658266766,
      "tokens_seen": 97583104
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000490110330992979,
      "loss": 3.6426,
      "theoretical_loss": 4.788028381306497,
      "tokens_seen": 97648640
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004901003009027082,
      "loss": 3.6375,
      "theoretical_loss": 4.787581488419207,
      "tokens_seen": 97714176
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004900902708124374,
      "loss": 3.7361,
      "theoretical_loss": 4.787134979017832,
      "tokens_seen": 97779712
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004900802407221665,
      "loss": 3.676,
      "theoretical_loss": 4.786688852516599,
      "tokens_seen": 97845248
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004900702106318957,
      "loss": 3.9366,
      "theoretical_loss": 4.786243108331024,
      "tokens_seen": 97910784
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004900601805416249,
      "loss": 3.7287,
      "theoretical_loss": 4.7857977458779,
      "tokens_seen": 97976320
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004900501504513541,
      "loss": 3.5148,
      "theoretical_loss": 4.785352764575304,
      "tokens_seen": 98041856
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004900401203610833,
      "loss": 3.8189,
      "theoretical_loss": 4.784908163842585,
      "tokens_seen": 98107392
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004900300902708124,
      "loss": 3.741,
      "theoretical_loss": 4.784463943100367,
      "tokens_seen": 98172928
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004900200601805416,
      "loss": 3.9671,
      "theoretical_loss": 4.7840201017705395,
      "tokens_seen": 98238464
    },
    {
      "epoch": 0.03,
      "objective/train/docs_used": 175137,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.979011058807373,
      "objective/train/theoretical_loss": 4.783576639276257,
      "objective/train/tokens_used": 118764000,
      "theoretical_loss": 4.783576639276257,
      "tokens_seen": 98304000
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004900100300902708,
      "loss": 3.7197,
      "theoretical_loss": 4.783576639276257,
      "tokens_seen": 98304000
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00049,
      "loss": 3.9945,
      "theoretical_loss": 4.783133555041934,
      "tokens_seen": 98369536
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004899899699097292,
      "loss": 3.8699,
      "theoretical_loss": 4.782690848493245,
      "tokens_seen": 98435072
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004899799398194583,
      "loss": 3.7009,
      "theoretical_loss": 4.7822485190571165,
      "tokens_seen": 98500608
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004899699097291876,
      "loss": 3.9709,
      "theoretical_loss": 4.781806566161723,
      "tokens_seen": 98566144
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004899598796389167,
      "loss": 3.8659,
      "theoretical_loss": 4.781364989236488,
      "tokens_seen": 98631680
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000489949849548646,
      "loss": 3.6496,
      "theoretical_loss": 4.78092378771208,
      "tokens_seen": 98697216
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004899398194583751,
      "loss": 3.7254,
      "theoretical_loss": 4.780482961020402,
      "tokens_seen": 98762752
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004899297893681044,
      "loss": 3.7948,
      "theoretical_loss": 4.780042508594596,
      "tokens_seen": 98828288
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004899197592778335,
      "loss": 3.9568,
      "theoretical_loss": 4.779602429869035,
      "tokens_seen": 98893824
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004899097291875627,
      "loss": 3.7481,
      "theoretical_loss": 4.779162724279324,
      "tokens_seen": 98959360
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004898996990972919,
      "loss": 3.755,
      "theoretical_loss": 4.7787233912622895,
      "tokens_seen": 99024896
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004898896690070211,
      "loss": 3.7307,
      "theoretical_loss": 4.778284430255981,
      "tokens_seen": 99090432
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004898796389167503,
      "loss": 3.4225,
      "theoretical_loss": 4.77784584069967,
      "tokens_seen": 99155968
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004898696088264794,
      "loss": 3.8225,
      "theoretical_loss": 4.777407622033838,
      "tokens_seen": 99221504
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004898595787362086,
      "loss": 3.5734,
      "theoretical_loss": 4.776969773700181,
      "tokens_seen": 99287040
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004898495486459378,
      "loss": 3.6576,
      "theoretical_loss": 4.776532295141601,
      "tokens_seen": 99352576
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000489839518555667,
      "loss": 3.916,
      "theoretical_loss": 4.776095185802211,
      "tokens_seen": 99418112
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004898294884653962,
      "loss": 3.4804,
      "theoretical_loss": 4.775658445127318,
      "tokens_seen": 99483648
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004898194583751254,
      "loss": 3.6648,
      "theoretical_loss": 4.775222072563429,
      "tokens_seen": 99549184
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004898094282848545,
      "loss": 4.0798,
      "theoretical_loss": 4.7747860675582485,
      "tokens_seen": 99614720
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004897993981945837,
      "loss": 3.7901,
      "theoretical_loss": 4.77435042956067,
      "tokens_seen": 99680256
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000489789368104313,
      "loss": 3.6989,
      "theoretical_loss": 4.773915158020776,
      "tokens_seen": 99745792
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004897793380140421,
      "loss": 3.6796,
      "theoretical_loss": 4.773480252389831,
      "tokens_seen": 99811328
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004897693079237714,
      "loss": 3.6931,
      "theoretical_loss": 4.773045712120284,
      "tokens_seen": 99876864
    },
    {
      "epoch": 0.03,
      "objective/train/docs_used": 175797,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.162777900695801,
      "objective/train/theoretical_loss": 4.77261153666576,
      "objective/train/tokens_used": 120402400,
      "theoretical_loss": 4.77261153666576,
      "tokens_seen": 99942400
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004897592778335005,
      "loss": 3.6811,
      "theoretical_loss": 4.77261153666576,
      "tokens_seen": 99942400
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004897492477432297,
      "loss": 3.8359,
      "theoretical_loss": 4.772177725481062,
      "tokens_seen": 100007936
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004897392176529589,
      "loss": 4.1084,
      "theoretical_loss": 4.77174427802216,
      "tokens_seen": 100073472
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004897291875626881,
      "loss": 3.5353,
      "theoretical_loss": 4.771311193746191,
      "tokens_seen": 100139008
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004897191574724173,
      "loss": 3.6427,
      "theoretical_loss": 4.770878472111465,
      "tokens_seen": 100204544
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004897091273821465,
      "loss": 3.7959,
      "theoretical_loss": 4.770446112577445,
      "tokens_seen": 100270080
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004896990972918756,
      "loss": 3.7708,
      "theoretical_loss": 4.770014114604756,
      "tokens_seen": 100335616
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004896890672016048,
      "loss": 3.698,
      "theoretical_loss": 4.769582477655177,
      "tokens_seen": 100401152
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000489679037111334,
      "loss": 3.8029,
      "theoretical_loss": 4.769151201191641,
      "tokens_seen": 100466688
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004896690070210632,
      "loss": 3.488,
      "theoretical_loss": 4.768720284678228,
      "tokens_seen": 100532224
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004896589769307924,
      "loss": 3.7797,
      "theoretical_loss": 4.768289727580161,
      "tokens_seen": 100597760
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004896489468405215,
      "loss": 3.4447,
      "theoretical_loss": 4.767859529363809,
      "tokens_seen": 100663296
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004896389167502507,
      "loss": 3.7995,
      "theoretical_loss": 4.767429689496682,
      "tokens_seen": 100728832
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004896288866599799,
      "loss": 3.8211,
      "theoretical_loss": 4.767000207447417,
      "tokens_seen": 100794368
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004896188565697091,
      "loss": 3.5972,
      "theoretical_loss": 4.766571082685794,
      "tokens_seen": 100859904
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004896088264794384,
      "loss": 3.8521,
      "theoretical_loss": 4.766142314682716,
      "tokens_seen": 100925440
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004895987963891674,
      "loss": 3.7264,
      "theoretical_loss": 4.765713902910214,
      "tokens_seen": 100990976
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004895887662988967,
      "loss": 3.702,
      "theoretical_loss": 4.765285846841444,
      "tokens_seen": 101056512
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004895787362086259,
      "loss": 3.8745,
      "theoretical_loss": 4.76485814595068,
      "tokens_seen": 101122048
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004895687061183551,
      "loss": 3.8407,
      "theoretical_loss": 4.764430799713314,
      "tokens_seen": 101187584
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004895586760280843,
      "loss": 3.5995,
      "theoretical_loss": 4.764003807605853,
      "tokens_seen": 101253120
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004895486459378135,
      "loss": 3.6566,
      "theoretical_loss": 4.763577169105912,
      "tokens_seen": 101318656
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004895386158475426,
      "loss": 3.7968,
      "theoretical_loss": 4.763150883692218,
      "tokens_seen": 101384192
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004895285857572718,
      "loss": 3.6979,
      "theoretical_loss": 4.762724950844598,
      "tokens_seen": 101449728
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000489518555667001,
      "loss": 3.7237,
      "theoretical_loss": 4.762299370043984,
      "tokens_seen": 101515264
    },
    {
      "epoch": 0.03,
      "objective/train/docs_used": 179449,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.7288663387298584,
      "objective/train/theoretical_loss": 4.761874140772408,
      "objective/train/tokens_used": 122040800,
      "theoretical_loss": 4.761874140772408,
      "tokens_seen": 101580800
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004895085255767302,
      "loss": 3.8993,
      "theoretical_loss": 4.761874140772408,
      "tokens_seen": 101580800
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004894984954864594,
      "loss": 3.523,
      "theoretical_loss": 4.761449262512993,
      "tokens_seen": 101646336
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004894884653961885,
      "loss": 3.4939,
      "theoretical_loss": 4.761024734749958,
      "tokens_seen": 101711872
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004894784353059178,
      "loss": 3.8311,
      "theoretical_loss": 4.76060055696861,
      "tokens_seen": 101777408
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004894684052156469,
      "loss": 3.8815,
      "theoretical_loss": 4.760176728655345,
      "tokens_seen": 101842944
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004894583751253762,
      "loss": 3.8344,
      "theoretical_loss": 4.75975324929764,
      "tokens_seen": 101908480
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004894483450351053,
      "loss": 3.7128,
      "theoretical_loss": 4.759330118384053,
      "tokens_seen": 101974016
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004894383149448346,
      "loss": 3.7163,
      "theoretical_loss": 4.758907335404221,
      "tokens_seen": 102039552
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004894282848545637,
      "loss": 3.5908,
      "theoretical_loss": 4.758484899848854,
      "tokens_seen": 102105088
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004894182547642929,
      "loss": 3.8424,
      "theoretical_loss": 4.7580628112097365,
      "tokens_seen": 102170624
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004894082246740221,
      "loss": 3.6798,
      "theoretical_loss": 4.7576410689797175,
      "tokens_seen": 102236160
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004893981945837513,
      "loss": 3.6937,
      "theoretical_loss": 4.757219672652717,
      "tokens_seen": 102301696
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004893881644934805,
      "loss": 3.6042,
      "theoretical_loss": 4.756798621723712,
      "tokens_seen": 102367232
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004893781344032096,
      "loss": 3.8252,
      "theoretical_loss": 4.756377915688748,
      "tokens_seen": 102432768
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004893681043129388,
      "loss": 3.6281,
      "theoretical_loss": 4.755957554044917,
      "tokens_seen": 102498304
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000489358074222668,
      "loss": 3.634,
      "theoretical_loss": 4.755537536290373,
      "tokens_seen": 102563840
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004893480441323972,
      "loss": 3.5845,
      "theoretical_loss": 4.755117861924321,
      "tokens_seen": 102629376
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004893380140421264,
      "loss": 3.7616,
      "theoretical_loss": 4.754698530447009,
      "tokens_seen": 102694912
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004893279839518556,
      "loss": 3.8346,
      "theoretical_loss": 4.754279541359738,
      "tokens_seen": 102760448
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004893179538615847,
      "loss": 3.4694,
      "theoretical_loss": 4.753860894164845,
      "tokens_seen": 102825984
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004893079237713139,
      "loss": 3.8897,
      "theoretical_loss": 4.75344258836571,
      "tokens_seen": 102891520
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004892978936810432,
      "loss": 3.5267,
      "theoretical_loss": 4.753024623466752,
      "tokens_seen": 102957056
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004892878635907723,
      "loss": 3.4275,
      "theoretical_loss": 4.752606998973421,
      "tokens_seen": 103022592
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004892778335005016,
      "loss": 3.5392,
      "theoretical_loss": 4.752189714392202,
      "tokens_seen": 103088128
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004892678034102307,
      "loss": 3.6427,
      "theoretical_loss": 4.7517727692306035,
      "tokens_seen": 103153664
    },
    {
      "epoch": 0.03,
      "objective/train/docs_used": 184726,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9697039127349854,
      "objective/train/theoretical_loss": 4.751356162997164,
      "objective/train/tokens_used": 123679200,
      "theoretical_loss": 4.751356162997164,
      "tokens_seen": 103219200
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004892577733199599,
      "loss": 3.525,
      "theoretical_loss": 4.751356162997164,
      "tokens_seen": 103219200
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004892477432296891,
      "loss": 3.6341,
      "theoretical_loss": 4.750939895201443,
      "tokens_seen": 103284736
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004892377131394183,
      "loss": 3.6469,
      "theoretical_loss": 4.750523965354024,
      "tokens_seen": 103350272
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004892276830491475,
      "loss": 3.4156,
      "theoretical_loss": 4.750108372966501,
      "tokens_seen": 103415808
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004892176529588767,
      "loss": 3.5315,
      "theoretical_loss": 4.749693117551491,
      "tokens_seen": 103481344
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004892076228686058,
      "loss": 3.6321,
      "theoretical_loss": 4.749278198622617,
      "tokens_seen": 103546880
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000489197592778335,
      "loss": 3.54,
      "theoretical_loss": 4.748863615694514,
      "tokens_seen": 103612416
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004891875626880642,
      "loss": 3.6321,
      "theoretical_loss": 4.748449368282822,
      "tokens_seen": 103677952
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004891775325977934,
      "loss": 3.7374,
      "theoretical_loss": 4.748035455904185,
      "tokens_seen": 103743488
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004891675025075226,
      "loss": 3.7389,
      "theoretical_loss": 4.747621878076252,
      "tokens_seen": 103809024
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004891574724172517,
      "loss": 3.8047,
      "theoretical_loss": 4.747208634317664,
      "tokens_seen": 103874560
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004891474423269809,
      "loss": 3.6744,
      "theoretical_loss": 4.746795724148061,
      "tokens_seen": 103940096
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004891374122367101,
      "loss": 3.9085,
      "theoretical_loss": 4.746383147088078,
      "tokens_seen": 104005632
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004891273821464393,
      "loss": 3.471,
      "theoretical_loss": 4.745970902659338,
      "tokens_seen": 104071168
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004891173520561686,
      "loss": 3.8325,
      "theoretical_loss": 4.745558990384451,
      "tokens_seen": 104136704
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004891073219658976,
      "loss": 3.523,
      "theoretical_loss": 4.7451474097870125,
      "tokens_seen": 104202240
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004890972918756269,
      "loss": 3.5334,
      "theoretical_loss": 4.744736160391602,
      "tokens_seen": 104267776
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004890872617853561,
      "loss": 3.6145,
      "theoretical_loss": 4.744325241723777,
      "tokens_seen": 104333312
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004890772316950853,
      "loss": 3.7004,
      "theoretical_loss": 4.743914653310073,
      "tokens_seen": 104398848
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004890672016048145,
      "loss": 3.674,
      "theoretical_loss": 4.743504394678,
      "tokens_seen": 104464384
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004890571715145437,
      "loss": 3.8081,
      "theoretical_loss": 4.743094465356039,
      "tokens_seen": 104529920
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004890471414242728,
      "loss": 3.7858,
      "theoretical_loss": 4.742684864873641,
      "tokens_seen": 104595456
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000489037111334002,
      "loss": 3.5326,
      "theoretical_loss": 4.742275592761223,
      "tokens_seen": 104660992
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004890270812437312,
      "loss": 3.5429,
      "theoretical_loss": 4.741866648550168,
      "tokens_seen": 104726528
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004890170511534604,
      "loss": 3.4406,
      "theoretical_loss": 4.741458031772817,
      "tokens_seen": 104792064
    },
    {
      "epoch": 0.03,
      "objective/train/docs_used": 187033,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 4.107706069946289,
      "objective/train/theoretical_loss": 4.741049741962473,
      "objective/train/tokens_used": 125317600,
      "theoretical_loss": 4.741049741962473,
      "tokens_seen": 104857600
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004890070210631896,
      "loss": 3.9333,
      "theoretical_loss": 4.741049741962473,
      "tokens_seen": 104857600
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004889969909729187,
      "loss": 3.5559,
      "theoretical_loss": 4.740641778653395,
      "tokens_seen": 104923136
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004889869608826479,
      "loss": 3.4574,
      "theoretical_loss": 4.740234141380794,
      "tokens_seen": 104988672
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004889769307923771,
      "loss": 3.7356,
      "theoretical_loss": 4.739826829680833,
      "tokens_seen": 105054208
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004889669007021063,
      "loss": 3.472,
      "theoretical_loss": 4.739419843090626,
      "tokens_seen": 105119744
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004889568706118355,
      "loss": 3.7092,
      "theoretical_loss": 4.739013181148229,
      "tokens_seen": 105185280
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004889468405215647,
      "loss": 3.8201,
      "theoretical_loss": 4.738606843392644,
      "tokens_seen": 105250816
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004889368104312939,
      "loss": 3.4478,
      "theoretical_loss": 4.738200829363815,
      "tokens_seen": 105316352
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000488926780341023,
      "loss": 3.9938,
      "theoretical_loss": 4.737795138602624,
      "tokens_seen": 105381888
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004889167502507523,
      "loss": 3.7457,
      "theoretical_loss": 4.737389770650887,
      "tokens_seen": 105447424
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004889067201604815,
      "loss": 3.6661,
      "theoretical_loss": 4.736984725051357,
      "tokens_seen": 105512960
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004888966900702107,
      "loss": 3.6371,
      "theoretical_loss": 4.736580001347717,
      "tokens_seen": 105578496
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004888866599799398,
      "loss": 3.7063,
      "theoretical_loss": 4.736175599084576,
      "tokens_seen": 105644032
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000488876629889669,
      "loss": 3.7422,
      "theoretical_loss": 4.735771517807473,
      "tokens_seen": 105709568
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004888665997993982,
      "loss": 3.6926,
      "theoretical_loss": 4.735367757062869,
      "tokens_seen": 105775104
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004888565697091274,
      "loss": 3.616,
      "theoretical_loss": 4.734964316398148,
      "tokens_seen": 105840640
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004888465396188566,
      "loss": 3.5101,
      "theoretical_loss": 4.734561195361609,
      "tokens_seen": 105906176
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004888365095285858,
      "loss": 3.5185,
      "theoretical_loss": 4.734158393502471,
      "tokens_seen": 105971712
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004888264794383149,
      "loss": 3.8055,
      "theoretical_loss": 4.733755910370867,
      "tokens_seen": 106037248
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004888164493480441,
      "loss": 3.6915,
      "theoretical_loss": 4.73335374551784,
      "tokens_seen": 106102784
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004888064192577733,
      "loss": 3.5202,
      "theoretical_loss": 4.732951898495341,
      "tokens_seen": 106168320
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004887963891675025,
      "loss": 3.675,
      "theoretical_loss": 4.7325503688562325,
      "tokens_seen": 106233856
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004887863590772317,
      "loss": 3.4758,
      "theoretical_loss": 4.732149156154276,
      "tokens_seen": 106299392
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000488776328986961,
      "loss": 3.7876,
      "theoretical_loss": 4.731748259944139,
      "tokens_seen": 106364928
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048876629889669,
      "loss": 3.649,
      "theoretical_loss": 4.731347679781386,
      "tokens_seen": 106430464
    },
    {
      "epoch": 0.03,
      "objective/train/docs_used": 192056,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.8580915927886963,
      "objective/train/theoretical_loss": 4.730947415222481,
      "objective/train/tokens_used": 126956000,
      "theoretical_loss": 4.730947415222481,
      "tokens_seen": 106496000
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004887562688064193,
      "loss": 3.6611,
      "theoretical_loss": 4.730947415222481,
      "tokens_seen": 106496000
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004887462387161484,
      "loss": 3.8264,
      "theoretical_loss": 4.730547465824781,
      "tokens_seen": 106561536
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004887362086258777,
      "loss": 3.863,
      "theoretical_loss": 4.730147831146537,
      "tokens_seen": 106627072
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004887261785356069,
      "loss": 3.5344,
      "theoretical_loss": 4.72974851074689,
      "tokens_seen": 106692608
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000488716148445336,
      "loss": 3.4504,
      "theoretical_loss": 4.729349504185867,
      "tokens_seen": 106758144
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004887061183550652,
      "loss": 3.642,
      "theoretical_loss": 4.728950811024383,
      "tokens_seen": 106823680
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004886960882647944,
      "loss": 3.6794,
      "theoretical_loss": 4.7285524308242355,
      "tokens_seen": 106889216
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004886860581745236,
      "loss": 3.2107,
      "theoretical_loss": 4.728154363148102,
      "tokens_seen": 106954752
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004886760280842528,
      "loss": 3.6065,
      "theoretical_loss": 4.72775660755954,
      "tokens_seen": 107020288
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004886659979939819,
      "loss": 3.9955,
      "theoretical_loss": 4.72735916362298,
      "tokens_seen": 107085824
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004886559679037111,
      "loss": 3.6974,
      "theoretical_loss": 4.7269620309037315,
      "tokens_seen": 107151360
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004886459378134403,
      "loss": 3.728,
      "theoretical_loss": 4.726565208967973,
      "tokens_seen": 107216896
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004886359077231695,
      "loss": 3.4342,
      "theoretical_loss": 4.726168697382751,
      "tokens_seen": 107282432
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004886258776328988,
      "loss": 3.2861,
      "theoretical_loss": 4.725772495715983,
      "tokens_seen": 107347968
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004886158475426278,
      "loss": 3.7293,
      "theoretical_loss": 4.725376603536446,
      "tokens_seen": 107413504
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004886058174523571,
      "loss": 3.964,
      "theoretical_loss": 4.724981020413787,
      "tokens_seen": 107479040
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004885957873620863,
      "loss": 3.6411,
      "theoretical_loss": 4.724585745918505,
      "tokens_seen": 107544576
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004885857572718155,
      "loss": 3.6381,
      "theoretical_loss": 4.7241907796219635,
      "tokens_seen": 107610112
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004885757271815447,
      "loss": 3.6016,
      "theoretical_loss": 4.723796121096381,
      "tokens_seen": 107675648
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004885656970912739,
      "loss": 3.3598,
      "theoretical_loss": 4.723401769914824,
      "tokens_seen": 107741184
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000488555667001003,
      "loss": 3.731,
      "theoretical_loss": 4.723007725651219,
      "tokens_seen": 107806720
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004885456369107322,
      "loss": 3.6714,
      "theoretical_loss": 4.722613987880335,
      "tokens_seen": 107872256
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004885356068204614,
      "loss": 3.6898,
      "theoretical_loss": 4.722220556177792,
      "tokens_seen": 107937792
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004885255767301906,
      "loss": 3.609,
      "theoretical_loss": 4.721827430120053,
      "tokens_seen": 108003328
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004885155466399198,
      "loss": 3.4883,
      "theoretical_loss": 4.721434609284424,
      "tokens_seen": 108068864
    },
    {
      "epoch": 0.03,
      "objective/train/docs_used": 197231,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.762702465057373,
      "objective/train/theoretical_loss": 4.721042093249051,
      "objective/train/tokens_used": 128594400,
      "theoretical_loss": 4.721042093249051,
      "tokens_seen": 108134400
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000488505516549649,
      "loss": 3.6925,
      "theoretical_loss": 4.721042093249051,
      "tokens_seen": 108134400
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004884954864593781,
      "loss": 3.4986,
      "theoretical_loss": 4.720649881592919,
      "tokens_seen": 108199936
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004884854563691073,
      "loss": 3.4754,
      "theoretical_loss": 4.7202579738958494,
      "tokens_seen": 108265472
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004884754262788365,
      "loss": 3.3368,
      "theoretical_loss": 4.7198663697384955,
      "tokens_seen": 108331008
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004884653961885657,
      "loss": 3.6466,
      "theoretical_loss": 4.719475068702346,
      "tokens_seen": 108396544
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004884553660982949,
      "loss": 3.5481,
      "theoretical_loss": 4.719084070369714,
      "tokens_seen": 108462080
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004884453360080241,
      "loss": 3.5884,
      "theoretical_loss": 4.718693374323747,
      "tokens_seen": 108527616
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004884353059177532,
      "loss": 3.8041,
      "theoretical_loss": 4.718302980148412,
      "tokens_seen": 108593152
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004884252758274825,
      "loss": 3.8141,
      "theoretical_loss": 4.717912887428501,
      "tokens_seen": 108658688
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004884152457372117,
      "loss": 3.5719,
      "theoretical_loss": 4.717523095749626,
      "tokens_seen": 108724224
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004884052156469409,
      "loss": 3.7324,
      "theoretical_loss": 4.717133604698222,
      "tokens_seen": 108789760
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048839518555667,
      "loss": 3.6457,
      "theoretical_loss": 4.7167444138615355,
      "tokens_seen": 108855296
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004883851554663992,
      "loss": 3.4354,
      "theoretical_loss": 4.716355522827633,
      "tokens_seen": 108920832
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004883751253761284,
      "loss": 3.6252,
      "theoretical_loss": 4.715966931185388,
      "tokens_seen": 108986368
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004883650952858576,
      "loss": 3.591,
      "theoretical_loss": 4.715578638524491,
      "tokens_seen": 109051904
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004883550651955868,
      "loss": 3.6158,
      "theoretical_loss": 4.715190644435435,
      "tokens_seen": 109117440
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000488345035105316,
      "loss": 3.6187,
      "theoretical_loss": 4.714802948509522,
      "tokens_seen": 109182976
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004883350050150451,
      "loss": 3.7455,
      "theoretical_loss": 4.71441555033886,
      "tokens_seen": 109248512
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004883249749247743,
      "loss": 3.712,
      "theoretical_loss": 4.714028449516356,
      "tokens_seen": 109314048
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004883149448345035,
      "loss": 3.7868,
      "theoretical_loss": 4.713641645635718,
      "tokens_seen": 109379584
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004883049147442327,
      "loss": 3.4678,
      "theoretical_loss": 4.713255138291454,
      "tokens_seen": 109445120
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004882948846539619,
      "loss": 3.6966,
      "theoretical_loss": 4.712868927078868,
      "tokens_seen": 109510656
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004882848545636911,
      "loss": 3.7838,
      "theoretical_loss": 4.712483011594056,
      "tokens_seen": 109576192
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004882748244734203,
      "loss": 3.3161,
      "theoretical_loss": 4.7120973914339075,
      "tokens_seen": 109641728
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048826479438314946,
      "loss": 3.8165,
      "theoretical_loss": 4.7117120661961005,
      "tokens_seen": 109707264
    },
    {
      "epoch": 0.03,
      "objective/train/docs_used": 198319,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.8059911727905273,
      "objective/train/theoretical_loss": 4.711327035479103,
      "objective/train/tokens_used": 130232800,
      "theoretical_loss": 4.711327035479103,
      "tokens_seen": 109772800
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048825476429287864,
      "loss": 3.6663,
      "theoretical_loss": 4.711327035479103,
      "tokens_seen": 109772800
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004882447342026078,
      "loss": 3.4004,
      "theoretical_loss": 4.710942298882169,
      "tokens_seen": 109838336
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000488234704112337,
      "loss": 3.5996,
      "theoretical_loss": 4.710557856005335,
      "tokens_seen": 109903872
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048822467402206624,
      "loss": 3.7299,
      "theoretical_loss": 4.710173706449419,
      "tokens_seen": 109969408
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048821464393179536,
      "loss": 3.6217,
      "theoretical_loss": 4.709789849816021,
      "tokens_seen": 110034944
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004882046138415246,
      "loss": 3.8175,
      "theoretical_loss": 4.7094062857075185,
      "tokens_seen": 110100480
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004881945837512537,
      "loss": 3.5329,
      "theoretical_loss": 4.709023013727063,
      "tokens_seen": 110166016
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048818455366098296,
      "loss": 3.5599,
      "theoretical_loss": 4.708640033478584,
      "tokens_seen": 110231552
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048817452357071214,
      "loss": 3.4719,
      "theoretical_loss": 4.708257344566778,
      "tokens_seen": 110297088
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004881644934804413,
      "loss": 3.627,
      "theoretical_loss": 4.7078749465971175,
      "tokens_seen": 110362624
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004881544633901705,
      "loss": 3.7143,
      "theoretical_loss": 4.707492839175837,
      "tokens_seen": 110428160
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048814443329989974,
      "loss": 3.5303,
      "theoretical_loss": 4.707111021909941,
      "tokens_seen": 110493696
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048813440320962887,
      "loss": 3.5631,
      "theoretical_loss": 4.706729494407197,
      "tokens_seen": 110559232
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004881243731193581,
      "loss": 3.4962,
      "theoretical_loss": 4.706348256276138,
      "tokens_seen": 110624768
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048811434302908723,
      "loss": 3.6254,
      "theoretical_loss": 4.705967307126051,
      "tokens_seen": 110690304
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048810431293881646,
      "loss": 3.5333,
      "theoretical_loss": 4.705586646566987,
      "tokens_seen": 110755840
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048809428284854564,
      "loss": 3.386,
      "theoretical_loss": 4.705206274209751,
      "tokens_seen": 110821376
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004880842527582748,
      "loss": 3.6736,
      "theoretical_loss": 4.704826189665905,
      "tokens_seen": 110886912
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000488074222668004,
      "loss": 3.4569,
      "theoretical_loss": 4.704446392547759,
      "tokens_seen": 110952448
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004880641925777332,
      "loss": 3.5925,
      "theoretical_loss": 4.7040668824683785,
      "tokens_seen": 111017984
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048805416248746237,
      "loss": 3.6843,
      "theoretical_loss": 4.7036876590415755,
      "tokens_seen": 111083520
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004880441323971916,
      "loss": 3.8082,
      "theoretical_loss": 4.7033087218819105,
      "tokens_seen": 111149056
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048803410230692073,
      "loss": 3.5422,
      "theoretical_loss": 4.7029300706046895,
      "tokens_seen": 111214592
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048802407221664997,
      "loss": 3.7771,
      "theoretical_loss": 4.702551704825957,
      "tokens_seen": 111280128
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048801404212637915,
      "loss": 3.6929,
      "theoretical_loss": 4.702173624162507,
      "tokens_seen": 111345664
    },
    {
      "epoch": 0.03,
      "objective/train/docs_used": 199018,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.2992162704467773,
      "objective/train/theoretical_loss": 4.701795828231866,
      "objective/train/tokens_used": 131871200,
      "theoretical_loss": 4.701795828231866,
      "tokens_seen": 111411200
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048800401203610833,
      "loss": 3.3269,
      "theoretical_loss": 4.701795828231866,
      "tokens_seen": 111411200
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004879939819458375,
      "loss": 3.4583,
      "theoretical_loss": 4.701418316652299,
      "tokens_seen": 111476736
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004879839518555667,
      "loss": 3.5179,
      "theoretical_loss": 4.701041089042813,
      "tokens_seen": 111542272
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004879739217652959,
      "loss": 3.5889,
      "theoretical_loss": 4.700664145023142,
      "tokens_seen": 111607808
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004879638916750251,
      "loss": 3.6855,
      "theoretical_loss": 4.700287484213753,
      "tokens_seen": 111673344
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004879538615847543,
      "loss": 3.6633,
      "theoretical_loss": 4.699911106235849,
      "tokens_seen": 111738880
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048794383149448347,
      "loss": 3.3769,
      "theoretical_loss": 4.6995350107113545,
      "tokens_seen": 111804416
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048793380140421265,
      "loss": 3.4899,
      "theoretical_loss": 4.699159197262922,
      "tokens_seen": 111869952
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048792377131394183,
      "loss": 3.4922,
      "theoretical_loss": 4.698783665513934,
      "tokens_seen": 111935488
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048791374122367107,
      "loss": 3.6508,
      "theoretical_loss": 4.698408415088491,
      "tokens_seen": 112001024
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004879037111334002,
      "loss": 3.644,
      "theoretical_loss": 4.698033445611415,
      "tokens_seen": 112066560
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048789368104312943,
      "loss": 3.6637,
      "theoretical_loss": 4.6976587567082495,
      "tokens_seen": 112132096
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048788365095285856,
      "loss": 3.3232,
      "theoretical_loss": 4.697284348005253,
      "tokens_seen": 112197632
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004878736208625878,
      "loss": 3.6476,
      "theoretical_loss": 4.696910219129402,
      "tokens_seen": 112263168
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048786359077231697,
      "loss": 3.7732,
      "theoretical_loss": 4.696536369708386,
      "tokens_seen": 112328704
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048785356068204615,
      "loss": 3.5794,
      "theoretical_loss": 4.696162799370606,
      "tokens_seen": 112394240
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048784353059177533,
      "loss": 3.5167,
      "theoretical_loss": 4.695789507745176,
      "tokens_seen": 112459776
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048783350050150457,
      "loss": 3.508,
      "theoretical_loss": 4.695416494461917,
      "tokens_seen": 112525312
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004878234704112337,
      "loss": 3.6201,
      "theoretical_loss": 4.695043759151353,
      "tokens_seen": 112590848
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048781344032096293,
      "loss": 3.5772,
      "theoretical_loss": 4.694671301444722,
      "tokens_seen": 112656384
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048780341023069206,
      "loss": 3.511,
      "theoretical_loss": 4.694299120973957,
      "tokens_seen": 112721920
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004877933801404213,
      "loss": 3.4497,
      "theoretical_loss": 4.693927217371698,
      "tokens_seen": 112787456
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004877833500501505,
      "loss": 3.6642,
      "theoretical_loss": 4.693555590271282,
      "tokens_seen": 112852992
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048777331995987966,
      "loss": 3.1849,
      "theoretical_loss": 4.693184239306744,
      "tokens_seen": 112918528
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048776328986960884,
      "loss": 3.7865,
      "theoretical_loss": 4.692813164112819,
      "tokens_seen": 112984064
    },
    {
      "epoch": 0.03,
      "objective/train/docs_used": 200274,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.979872226715088,
      "objective/train/theoretical_loss": 4.692442364324931,
      "objective/train/tokens_used": 133509600,
      "theoretical_loss": 4.692442364324931,
      "tokens_seen": 113049600
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000487753259779338,
      "loss": 3.2723,
      "theoretical_loss": 4.692442364324931,
      "tokens_seen": 113049600
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004877432296890672,
      "loss": 3.5444,
      "theoretical_loss": 4.692071839579201,
      "tokens_seen": 113115136
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048773319959879644,
      "loss": 3.4821,
      "theoretical_loss": 4.6917015895124425,
      "tokens_seen": 113180672
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048772316950852556,
      "loss": 3.8326,
      "theoretical_loss": 4.691331613762153,
      "tokens_seen": 113246208
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004877131394182548,
      "loss": 3.4582,
      "theoretical_loss": 4.690961911966523,
      "tokens_seen": 113311744
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004877031093279839,
      "loss": 3.3548,
      "theoretical_loss": 4.690592483764427,
      "tokens_seen": 113377280
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048769307923771316,
      "loss": 3.4375,
      "theoretical_loss": 4.690223328795424,
      "tokens_seen": 113442816
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048768304914744234,
      "loss": 3.5515,
      "theoretical_loss": 4.689854446699757,
      "tokens_seen": 113508352
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004876730190571715,
      "loss": 3.5217,
      "theoretical_loss": 4.689485837118347,
      "tokens_seen": 113573888
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004876629889669007,
      "loss": 3.5703,
      "theoretical_loss": 4.689117499692798,
      "tokens_seen": 113639424
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048765295887662994,
      "loss": 3.568,
      "theoretical_loss": 4.688749434065389,
      "tokens_seen": 113704960
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048764292878635907,
      "loss": 3.5214,
      "theoretical_loss": 4.688381639879076,
      "tokens_seen": 113770496
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004876328986960883,
      "loss": 3.7132,
      "theoretical_loss": 4.68801411677749,
      "tokens_seen": 113836032
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048762286860581743,
      "loss": 3.6747,
      "theoretical_loss": 4.687646864404934,
      "tokens_seen": 113901568
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048761283851554666,
      "loss": 3.7008,
      "theoretical_loss": 4.687279882406381,
      "tokens_seen": 113967104
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048760280842527584,
      "loss": 3.6326,
      "theoretical_loss": 4.686913170427477,
      "tokens_seen": 114032640
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000487592778335005,
      "loss": 3.5939,
      "theoretical_loss": 4.68654672811453,
      "tokens_seen": 114098176
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004875827482447342,
      "loss": 3.5751,
      "theoretical_loss": 4.68618055511452,
      "tokens_seen": 114163712
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004875727181544634,
      "loss": 3.4383,
      "theoretical_loss": 4.685814651075088,
      "tokens_seen": 114229248
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048756268806419257,
      "loss": 3.484,
      "theoretical_loss": 4.685449015644537,
      "tokens_seen": 114294784
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004875526579739218,
      "loss": 3.5631,
      "theoretical_loss": 4.685083648471835,
      "tokens_seen": 114360320
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048754262788365093,
      "loss": 3.592,
      "theoretical_loss": 4.684718549206607,
      "tokens_seen": 114425856
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048753259779338017,
      "loss": 3.7434,
      "theoretical_loss": 4.6843537174991345,
      "tokens_seen": 114491392
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048752256770310935,
      "loss": 3.5416,
      "theoretical_loss": 4.6839891530003595,
      "tokens_seen": 114556928
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048751253761283853,
      "loss": 3.8109,
      "theoretical_loss": 4.683624855361876,
      "tokens_seen": 114622464
    },
    {
      "epoch": 0.03,
      "objective/train/docs_used": 200959,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 4.169064998626709,
      "objective/train/theoretical_loss": 4.68326082423593,
      "objective/train/tokens_used": 135148000,
      "theoretical_loss": 4.68326082423593,
      "tokens_seen": 114688000
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004875025075225677,
      "loss": 3.5971,
      "theoretical_loss": 4.68326082423593,
      "tokens_seen": 114688000
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004874924774322969,
      "loss": 3.6054,
      "theoretical_loss": 4.682897059275422,
      "tokens_seen": 114753536
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048748244734202607,
      "loss": 3.5122,
      "theoretical_loss": 4.682533560133901,
      "tokens_seen": 114819072
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004874724172517553,
      "loss": 3.5253,
      "theoretical_loss": 4.682170326465565,
      "tokens_seen": 114884608
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048746238716148443,
      "loss": 3.4639,
      "theoretical_loss": 4.681807357925257,
      "tokens_seen": 114950144
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048745235707121367,
      "loss": 3.4482,
      "theoretical_loss": 4.681444654168468,
      "tokens_seen": 115015680
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004874423269809428,
      "loss": 3.4827,
      "theoretical_loss": 4.68108221485133,
      "tokens_seen": 115081216
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048743229689067203,
      "loss": 3.5963,
      "theoretical_loss": 4.680720039630617,
      "tokens_seen": 115146752
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004874222668004012,
      "loss": 3.604,
      "theoretical_loss": 4.680358128163747,
      "tokens_seen": 115212288
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004874122367101304,
      "loss": 3.575,
      "theoretical_loss": 4.679996480108773,
      "tokens_seen": 115277824
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004874022066198596,
      "loss": 3.6315,
      "theoretical_loss": 4.6796350951243895,
      "tokens_seen": 115343360
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048739217652958876,
      "loss": 3.5988,
      "theoretical_loss": 4.679273972869922,
      "tokens_seen": 115408896
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048738214643931794,
      "loss": 3.544,
      "theoretical_loss": 4.678913113005333,
      "tokens_seen": 115474432
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048737211634904717,
      "loss": 3.434,
      "theoretical_loss": 4.6785525151912175,
      "tokens_seen": 115539968
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004873620862587763,
      "loss": 3.7679,
      "theoretical_loss": 4.678192179088802,
      "tokens_seen": 115605504
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048735205616850553,
      "loss": 3.3392,
      "theoretical_loss": 4.6778321043599425,
      "tokens_seen": 115671040
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004873420260782347,
      "loss": 3.6271,
      "theoretical_loss": 4.677472290667122,
      "tokens_seen": 115736576
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004873319959879639,
      "loss": 3.5469,
      "theoretical_loss": 4.677112737673453,
      "tokens_seen": 115802112
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004873219658976931,
      "loss": 3.7895,
      "theoretical_loss": 4.676753445042669,
      "tokens_seen": 115867648
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048731193580742226,
      "loss": 3.395,
      "theoretical_loss": 4.676394412439132,
      "tokens_seen": 115933184
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048730190571715144,
      "loss": 3.4239,
      "theoretical_loss": 4.6760356395278215,
      "tokens_seen": 115998720
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004872918756268807,
      "loss": 3.8464,
      "theoretical_loss": 4.675677125974339,
      "tokens_seen": 116064256
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004872818455366098,
      "loss": 3.6943,
      "theoretical_loss": 4.675318871444908,
      "tokens_seen": 116129792
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048727181544633904,
      "loss": 3.2368,
      "theoretical_loss": 4.674960875606366,
      "tokens_seen": 116195328
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048726178535606816,
      "loss": 3.5752,
      "theoretical_loss": 4.674603138126168,
      "tokens_seen": 116260864
    },
    {
      "epoch": 0.04,
      "objective/train/docs_used": 202277,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.6310184001922607,
      "objective/train/theoretical_loss": 4.674245658672382,
      "objective/train/tokens_used": 136786400,
      "theoretical_loss": 4.674245658672382,
      "tokens_seen": 116326400
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004872517552657974,
      "loss": 3.3857,
      "theoretical_loss": 4.674245658672382,
      "tokens_seen": 116326400
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004872417251755266,
      "loss": 3.6198,
      "theoretical_loss": 4.673888436913694,
      "tokens_seen": 116391936
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048723169508525576,
      "loss": 3.5595,
      "theoretical_loss": 4.673531472519397,
      "tokens_seen": 116457472
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000487221664994985,
      "loss": 3.6411,
      "theoretical_loss": 4.673174765159393,
      "tokens_seen": 116523008
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004872116349047141,
      "loss": 3.5595,
      "theoretical_loss": 4.672818314504198,
      "tokens_seen": 116588544
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048720160481444336,
      "loss": 3.5692,
      "theoretical_loss": 4.6724621202249335,
      "tokens_seen": 116654080
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048719157472417254,
      "loss": 3.3932,
      "theoretical_loss": 4.672106181993324,
      "tokens_seen": 116719616
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004871815446339017,
      "loss": 3.5326,
      "theoretical_loss": 4.6717504994817,
      "tokens_seen": 116785152
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004871715145436309,
      "loss": 3.2663,
      "theoretical_loss": 4.671395072362996,
      "tokens_seen": 116850688
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048716148445336014,
      "loss": 3.565,
      "theoretical_loss": 4.671039900310747,
      "tokens_seen": 116916224
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048715145436308927,
      "loss": 3.582,
      "theoretical_loss": 4.670684982999088,
      "tokens_seen": 116981760
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004871414242728185,
      "loss": 3.6837,
      "theoretical_loss": 4.670330320102753,
      "tokens_seen": 117047296
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048713139418254763,
      "loss": 3.6938,
      "theoretical_loss": 4.669975911297072,
      "tokens_seen": 117112832
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048712136409227686,
      "loss": 3.5719,
      "theoretical_loss": 4.669621756257971,
      "tokens_seen": 117178368
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048711133400200604,
      "loss": 3.6868,
      "theoretical_loss": 4.669267854661973,
      "tokens_seen": 117243904
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004871013039117352,
      "loss": 3.3895,
      "theoretical_loss": 4.668914206186189,
      "tokens_seen": 117309440
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004870912738214644,
      "loss": 3.4743,
      "theoretical_loss": 4.6685608105083265,
      "tokens_seen": 117374976
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004870812437311936,
      "loss": 3.4699,
      "theoretical_loss": 4.66820766730668,
      "tokens_seen": 117440512
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048707121364092277,
      "loss": 3.727,
      "theoretical_loss": 4.667854776260132,
      "tokens_seen": 117506048
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000487061183550652,
      "loss": 3.5116,
      "theoretical_loss": 4.667502137048155,
      "tokens_seen": 117571584
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048705115346038113,
      "loss": 3.4195,
      "theoretical_loss": 4.667149749350805,
      "tokens_seen": 117637120
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048704112337011037,
      "loss": 3.6593,
      "theoretical_loss": 4.666797612848723,
      "tokens_seen": 117702656
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048703109327983955,
      "loss": 3.3363,
      "theoretical_loss": 4.666445727223134,
      "tokens_seen": 117768192
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048702106318956873,
      "loss": 3.5724,
      "theoretical_loss": 4.666094092155843,
      "tokens_seen": 117833728
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004870110330992979,
      "loss": 3.4272,
      "theoretical_loss": 4.665742707329238,
      "tokens_seen": 117899264
    },
    {
      "epoch": 0.04,
      "objective/train/docs_used": 203088,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.953324317932129,
      "objective/train/theoretical_loss": 4.665391572426282,
      "objective/train/tokens_used": 138424800,
      "theoretical_loss": 4.665391572426282,
      "tokens_seen": 117964800
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004870010030090271,
      "loss": 3.5257,
      "theoretical_loss": 4.665391572426282,
      "tokens_seen": 117964800
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048699097291875627,
      "loss": 3.6506,
      "theoretical_loss": 4.665040687130518,
      "tokens_seen": 118030336
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004869809428284855,
      "loss": 3.5996,
      "theoretical_loss": 4.664690051126065,
      "tokens_seen": 118095872
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048697091273821463,
      "loss": 3.5075,
      "theoretical_loss": 4.664339664097617,
      "tokens_seen": 118161408
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048696088264794387,
      "loss": 3.6508,
      "theoretical_loss": 4.66398952573044,
      "tokens_seen": 118226944
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000486950852557673,
      "loss": 3.7334,
      "theoretical_loss": 4.663639635710373,
      "tokens_seen": 118292480
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048694082246740223,
      "loss": 3.6102,
      "theoretical_loss": 4.663289993723826,
      "tokens_seen": 118358016
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004869307923771314,
      "loss": 3.5511,
      "theoretical_loss": 4.662940599457777,
      "tokens_seen": 118423552
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004869207622868606,
      "loss": 3.5001,
      "theoretical_loss": 4.662591452599774,
      "tokens_seen": 118489088
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004869107321965898,
      "loss": 3.485,
      "theoretical_loss": 4.662242552837929,
      "tokens_seen": 118554624
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048690070210631896,
      "loss": 3.6105,
      "theoretical_loss": 4.661893899860923,
      "tokens_seen": 118620160
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048689067201604814,
      "loss": 3.7023,
      "theoretical_loss": 4.6615454933579965,
      "tokens_seen": 118685696
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048688064192577737,
      "loss": 3.4614,
      "theoretical_loss": 4.661197333018957,
      "tokens_seen": 118751232
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004868706118355065,
      "loss": 3.5773,
      "theoretical_loss": 4.66084941853417,
      "tokens_seen": 118816768
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048686058174523573,
      "loss": 3.6587,
      "theoretical_loss": 4.6605017495945615,
      "tokens_seen": 118882304
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004868505516549649,
      "loss": 3.7256,
      "theoretical_loss": 4.660154325891618,
      "tokens_seen": 118947840
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004868405215646941,
      "loss": 3.6683,
      "theoretical_loss": 4.659807147117382,
      "tokens_seen": 119013376
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004868304914744233,
      "loss": 3.5433,
      "theoretical_loss": 4.6594602129644525,
      "tokens_seen": 119078912
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048682046138415246,
      "loss": 3.6163,
      "theoretical_loss": 4.659113523125981,
      "tokens_seen": 119144448
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048681043129388164,
      "loss": 3.6002,
      "theoretical_loss": 4.6587670772956775,
      "tokens_seen": 119209984
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004868004012036109,
      "loss": 3.744,
      "theoretical_loss": 4.658420875167799,
      "tokens_seen": 119275520
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048679037111334,
      "loss": 3.5674,
      "theoretical_loss": 4.658074916437155,
      "tokens_seen": 119341056
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048678034102306924,
      "loss": 3.705,
      "theoretical_loss": 4.657729200799105,
      "tokens_seen": 119406592
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048677031093279836,
      "loss": 3.5421,
      "theoretical_loss": 4.657383727949558,
      "tokens_seen": 119472128
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004867602808425276,
      "loss": 3.2356,
      "theoretical_loss": 4.657038497584967,
      "tokens_seen": 119537664
    },
    {
      "epoch": 0.04,
      "objective/train/docs_used": 204120,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.542322874069214,
      "objective/train/theoretical_loss": 4.656693509402331,
      "objective/train/tokens_used": 140063200,
      "theoretical_loss": 4.656693509402331,
      "tokens_seen": 119603200
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004867502507522568,
      "loss": 3.5035,
      "theoretical_loss": 4.656693509402331,
      "tokens_seen": 119603200
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048674022066198596,
      "loss": 3.6295,
      "theoretical_loss": 4.6563487630991975,
      "tokens_seen": 119668736
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048673019057171514,
      "loss": 3.6794,
      "theoretical_loss": 4.656004258373651,
      "tokens_seen": 119734272
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004867201604814443,
      "loss": 3.4914,
      "theoretical_loss": 4.655659994924323,
      "tokens_seen": 119799808
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004867101303911735,
      "loss": 3.4807,
      "theoretical_loss": 4.655315972450383,
      "tokens_seen": 119865344
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048670010030090274,
      "loss": 3.6322,
      "theoretical_loss": 4.65497219065154,
      "tokens_seen": 119930880
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048669007021063187,
      "loss": 3.5918,
      "theoretical_loss": 4.654628649228041,
      "tokens_seen": 119996416
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004866800401203611,
      "loss": 3.6249,
      "theoretical_loss": 4.654285347880672,
      "tokens_seen": 120061952
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004866700100300903,
      "loss": 3.6893,
      "theoretical_loss": 4.653942286310749,
      "tokens_seen": 120127488
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048665997993981947,
      "loss": 3.7484,
      "theoretical_loss": 4.653599464220129,
      "tokens_seen": 120193024
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048664994984954865,
      "loss": 3.6358,
      "theoretical_loss": 4.653256881311198,
      "tokens_seen": 120258560
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048663991975927783,
      "loss": 3.5567,
      "theoretical_loss": 4.6529145372868745,
      "tokens_seen": 120324096
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000486629889669007,
      "loss": 3.519,
      "theoretical_loss": 4.652572431850608,
      "tokens_seen": 120389632
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048661985957873624,
      "loss": 3.4446,
      "theoretical_loss": 4.652230564706377,
      "tokens_seen": 120455168
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048660982948846537,
      "loss": 3.4514,
      "theoretical_loss": 4.651888935558688,
      "tokens_seen": 120520704
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004865997993981946,
      "loss": 3.3068,
      "theoretical_loss": 4.651547544112575,
      "tokens_seen": 120586240
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048658976930792373,
      "loss": 3.7197,
      "theoretical_loss": 4.651206390073597,
      "tokens_seen": 120651776
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048657973921765297,
      "loss": 3.4481,
      "theoretical_loss": 4.650865473147837,
      "tokens_seen": 120717312
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048656970912738215,
      "loss": 3.3836,
      "theoretical_loss": 4.650524793041903,
      "tokens_seen": 120782848
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048655967903711133,
      "loss": 3.259,
      "theoretical_loss": 4.650184349462922,
      "tokens_seen": 120848384
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004865496489468405,
      "loss": 3.4753,
      "theoretical_loss": 4.649844142118544,
      "tokens_seen": 120913920
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048653961885656975,
      "loss": 3.6915,
      "theoretical_loss": 4.6495041707169396,
      "tokens_seen": 120979456
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004865295887662989,
      "loss": 3.562,
      "theoretical_loss": 4.649164434966794,
      "tokens_seen": 121044992
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004865195586760281,
      "loss": 3.4441,
      "theoretical_loss": 4.648824934577313,
      "tokens_seen": 121110528
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048650952858575724,
      "loss": 3.7728,
      "theoretical_loss": 4.648485669258216,
      "tokens_seen": 121176064
    },
    {
      "epoch": 0.04,
      "objective/train/docs_used": 204902,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.444392681121826,
      "objective/train/theoretical_loss": 4.648146638719739,
      "objective/train/tokens_used": 141701600,
      "theoretical_loss": 4.648146638719739,
      "tokens_seen": 121241600
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048649949849548647,
      "loss": 3.7584,
      "theoretical_loss": 4.648146638719739,
      "tokens_seen": 121241600
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048648946840521565,
      "loss": 3.3618,
      "theoretical_loss": 4.647807842672631,
      "tokens_seen": 121307136
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048647943831494483,
      "loss": 3.6952,
      "theoretical_loss": 4.647469280828153,
      "tokens_seen": 121372672
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048646940822467407,
      "loss": 3.4908,
      "theoretical_loss": 4.647130952898077,
      "tokens_seen": 121438208
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004864593781344032,
      "loss": 3.9747,
      "theoretical_loss": 4.646792858594686,
      "tokens_seen": 121503744
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048644934804413243,
      "loss": 3.3985,
      "theoretical_loss": 4.64645499763077,
      "tokens_seen": 121569280
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004864393179538616,
      "loss": 3.4271,
      "theoretical_loss": 4.646117369719629,
      "tokens_seen": 121634816
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004864292878635908,
      "loss": 3.5924,
      "theoretical_loss": 4.645779974575069,
      "tokens_seen": 121700352
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048641925777332,
      "loss": 3.4433,
      "theoretical_loss": 4.6454428119113995,
      "tokens_seen": 121765888
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048640922768304916,
      "loss": 3.6515,
      "theoretical_loss": 4.6451058814434365,
      "tokens_seen": 121831424
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048639919759277834,
      "loss": 3.8542,
      "theoretical_loss": 4.644769182886495,
      "tokens_seen": 121896960
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048638916750250757,
      "loss": 3.6482,
      "theoretical_loss": 4.644432715956399,
      "tokens_seen": 121962496
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004863791374122367,
      "loss": 3.4727,
      "theoretical_loss": 4.644096480369466,
      "tokens_seen": 122028032
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048636910732196593,
      "loss": 3.4828,
      "theoretical_loss": 4.643760475842518,
      "tokens_seen": 122093568
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004863590772316951,
      "loss": 3.3981,
      "theoretical_loss": 4.6434247020928705,
      "tokens_seen": 122159104
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004863490471414243,
      "loss": 3.4389,
      "theoretical_loss": 4.643089158838341,
      "tokens_seen": 122224640
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004863390170511535,
      "loss": 3.7861,
      "theoretical_loss": 4.642753845797243,
      "tokens_seen": 122290176
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048632898696088266,
      "loss": 3.4705,
      "theoretical_loss": 4.642418762688379,
      "tokens_seen": 122355712
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048631895687061184,
      "loss": 3.5749,
      "theoretical_loss": 4.642083909231053,
      "tokens_seen": 122421248
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004863089267803411,
      "loss": 3.6578,
      "theoretical_loss": 4.641749285145057,
      "tokens_seen": 122486784
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004862988966900702,
      "loss": 3.407,
      "theoretical_loss": 4.641414890150675,
      "tokens_seen": 122552320
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048628886659979944,
      "loss": 3.5225,
      "theoretical_loss": 4.641080723968684,
      "tokens_seen": 122617856
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048627883650952857,
      "loss": 3.4672,
      "theoretical_loss": 4.6407467863203475,
      "tokens_seen": 122683392
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004862688064192578,
      "loss": 3.336,
      "theoretical_loss": 4.640413076927418,
      "tokens_seen": 122748928
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000486258776328987,
      "loss": 3.2779,
      "theoretical_loss": 4.6400795955121374,
      "tokens_seen": 122814464
    },
    {
      "epoch": 0.04,
      "objective/train/docs_used": 206332,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 4.079704761505127,
      "objective/train/theoretical_loss": 4.639746341797229,
      "objective/train/tokens_used": 143340000,
      "theoretical_loss": 4.639746341797229,
      "tokens_seen": 122880000
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048624874623871616,
      "loss": 3.7945,
      "theoretical_loss": 4.639746341797229,
      "tokens_seen": 122880000
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048623871614844534,
      "loss": 3.7026,
      "theoretical_loss": 4.639413315505905,
      "tokens_seen": 122945536
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004862286860581745,
      "loss": 3.7991,
      "theoretical_loss": 4.639080516361861,
      "tokens_seen": 123011072
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004862186559679037,
      "loss": 3.3407,
      "theoretical_loss": 4.638747944089273,
      "tokens_seen": 123076608
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048620862587763294,
      "loss": 3.6945,
      "theoretical_loss": 4.638415598412799,
      "tokens_seen": 123142144
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048619859578736207,
      "loss": 3.3499,
      "theoretical_loss": 4.638083479057579,
      "tokens_seen": 123207680
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004861885656970913,
      "loss": 3.4411,
      "theoretical_loss": 4.637751585749234,
      "tokens_seen": 123273216
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004861785356068205,
      "loss": 3.419,
      "theoretical_loss": 4.6374199182138565,
      "tokens_seen": 123338752
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048616850551654967,
      "loss": 3.1327,
      "theoretical_loss": 4.637088476178025,
      "tokens_seen": 123404288
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048615847542627885,
      "loss": 3.3689,
      "theoretical_loss": 4.636757259368787,
      "tokens_seen": 123469824
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048614844533600803,
      "loss": 3.2499,
      "theoretical_loss": 4.636426267513668,
      "tokens_seen": 123535360
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004861384152457372,
      "loss": 3.443,
      "theoretical_loss": 4.636095500340669,
      "tokens_seen": 123600896
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048612838515546644,
      "loss": 3.4505,
      "theoretical_loss": 4.635764957578261,
      "tokens_seen": 123666432
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048611835506519557,
      "loss": 3.4719,
      "theoretical_loss": 4.635434638955388,
      "tokens_seen": 123731968
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004861083249749248,
      "loss": 3.6203,
      "theoretical_loss": 4.635104544201465,
      "tokens_seen": 123797504
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048609829488465393,
      "loss": 3.1321,
      "theoretical_loss": 4.634774673046376,
      "tokens_seen": 123863040
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048608826479438317,
      "loss": 3.3834,
      "theoretical_loss": 4.634445025220475,
      "tokens_seen": 123928576
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048607823470411235,
      "loss": 3.5649,
      "theoretical_loss": 4.634115600454582,
      "tokens_seen": 123994112
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048606820461384153,
      "loss": 3.3337,
      "theoretical_loss": 4.633786398479983,
      "tokens_seen": 124059648
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004860581745235707,
      "loss": 3.3326,
      "theoretical_loss": 4.6334574190284314,
      "tokens_seen": 124125184
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048604814443329995,
      "loss": 3.2998,
      "theoretical_loss": 4.633128661832145,
      "tokens_seen": 124190720
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004860381143430291,
      "loss": 3.581,
      "theoretical_loss": 4.632800126623803,
      "tokens_seen": 124256256
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004860280842527583,
      "loss": 3.6333,
      "theoretical_loss": 4.632471813136547,
      "tokens_seen": 124321792
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048601805416248744,
      "loss": 3.447,
      "theoretical_loss": 4.632143721103983,
      "tokens_seen": 124387328
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048600802407221667,
      "loss": 3.2293,
      "theoretical_loss": 4.631815850260173,
      "tokens_seen": 124452864
    },
    {
      "epoch": 0.04,
      "objective/train/docs_used": 207169,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.6358470916748047,
      "objective/train/theoretical_loss": 4.631488200339643,
      "objective/train/tokens_used": 144978400,
      "theoretical_loss": 4.631488200339643,
      "tokens_seen": 124518400
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048599799398194585,
      "loss": 3.4651,
      "theoretical_loss": 4.631488200339643,
      "tokens_seen": 124518400
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048598796389167503,
      "loss": 3.4181,
      "theoretical_loss": 4.63116077107737,
      "tokens_seen": 124583936
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004859779338014042,
      "loss": 3.6055,
      "theoretical_loss": 4.630833562208797,
      "tokens_seen": 124649472
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004859679037111334,
      "loss": 3.4885,
      "theoretical_loss": 4.630506573469815,
      "tokens_seen": 124715008
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004859578736208626,
      "loss": 3.5052,
      "theoretical_loss": 4.630179804596775,
      "tokens_seen": 124780544
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004859478435305918,
      "loss": 3.5164,
      "theoretical_loss": 4.629853255326481,
      "tokens_seen": 124846080
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048593781344032094,
      "loss": 3.3836,
      "theoretical_loss": 4.629526925396189,
      "tokens_seen": 124911616
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004859277833500502,
      "loss": 3.4527,
      "theoretical_loss": 4.6292008145436085,
      "tokens_seen": 124977152
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004859177532597793,
      "loss": 3.37,
      "theoretical_loss": 4.628874922506897,
      "tokens_seen": 125042688
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048590772316950854,
      "loss": 3.294,
      "theoretical_loss": 4.628549249024666,
      "tokens_seen": 125108224
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004858976930792377,
      "loss": 3.3212,
      "theoretical_loss": 4.628223793835975,
      "tokens_seen": 125173760
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004858876629889669,
      "loss": 3.3518,
      "theoretical_loss": 4.627898556680327,
      "tokens_seen": 125239296
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004858776328986961,
      "loss": 3.2403,
      "theoretical_loss": 4.627573537297678,
      "tokens_seen": 125304832
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004858676028084253,
      "loss": 3.4935,
      "theoretical_loss": 4.627248735428427,
      "tokens_seen": 125370368
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048585757271815444,
      "loss": 3.4322,
      "theoretical_loss": 4.6269241508134185,
      "tokens_seen": 125435904
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004858475426278837,
      "loss": 3.5357,
      "theoretical_loss": 4.6265997831939405,
      "tokens_seen": 125501440
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004858375125376128,
      "loss": 3.4814,
      "theoretical_loss": 4.6262756323117245,
      "tokens_seen": 125566976
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048582748244734204,
      "loss": 3.4993,
      "theoretical_loss": 4.625951697908944,
      "tokens_seen": 125632512
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004858174523570712,
      "loss": 3.4275,
      "theoretical_loss": 4.625627979728212,
      "tokens_seen": 125698048
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004858074222668004,
      "loss": 3.38,
      "theoretical_loss": 4.625304477512584,
      "tokens_seen": 125763584
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004857973921765296,
      "loss": 3.1327,
      "theoretical_loss": 4.624981191005554,
      "tokens_seen": 125829120
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048578736208625877,
      "loss": 3.4219,
      "theoretical_loss": 4.624658119951052,
      "tokens_seen": 125894656
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048577733199598795,
      "loss": 3.4453,
      "theoretical_loss": 4.624335264093447,
      "tokens_seen": 125960192
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004857673019057172,
      "loss": 3.5913,
      "theoretical_loss": 4.624012623177544,
      "tokens_seen": 126025728
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004857572718154463,
      "loss": 3.3552,
      "theoretical_loss": 4.623690196948582,
      "tokens_seen": 126091264
    },
    {
      "epoch": 0.04,
      "objective/train/docs_used": 208560,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.118776321411133,
      "objective/train/theoretical_loss": 4.623367985152234,
      "objective/train/tokens_used": 146616800,
      "theoretical_loss": 4.623367985152234,
      "tokens_seen": 126156800
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048574724172517554,
      "loss": 3.3034,
      "theoretical_loss": 4.623367985152234,
      "tokens_seen": 126156800
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048573721163490467,
      "loss": 3.4021,
      "theoretical_loss": 4.623045987534609,
      "tokens_seen": 126222336
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004857271815446339,
      "loss": 3.4968,
      "theoretical_loss": 4.622724203842246,
      "tokens_seen": 126287872
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048571715145436314,
      "loss": 3.6168,
      "theoretical_loss": 4.622402633822114,
      "tokens_seen": 126353408
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048570712136409227,
      "loss": 3.6483,
      "theoretical_loss": 4.622081277221616,
      "tokens_seen": 126418944
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004856970912738215,
      "loss": 3.1841,
      "theoretical_loss": 4.62176013378858,
      "tokens_seen": 126484480
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004856870611835507,
      "loss": 3.3364,
      "theoretical_loss": 4.621439203271267,
      "tokens_seen": 126550016
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048567703109327987,
      "loss": 3.5219,
      "theoretical_loss": 4.621118485418362,
      "tokens_seen": 126615552
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048566700100300905,
      "loss": 3.5676,
      "theoretical_loss": 4.620797979978978,
      "tokens_seen": 126681088
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048565697091273823,
      "loss": 3.0699,
      "theoretical_loss": 4.620477686702651,
      "tokens_seen": 126746624
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004856469408224674,
      "loss": 3.4134,
      "theoretical_loss": 4.620157605339347,
      "tokens_seen": 126812160
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048563691073219664,
      "loss": 3.3575,
      "theoretical_loss": 4.619837735639452,
      "tokens_seen": 126877696
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048562688064192577,
      "loss": 3.1724,
      "theoretical_loss": 4.619518077353776,
      "tokens_seen": 126943232
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000485616850551655,
      "loss": 3.5963,
      "theoretical_loss": 4.619198630233547,
      "tokens_seen": 127008768
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048560682046138413,
      "loss": 3.2372,
      "theoretical_loss": 4.6188793940304205,
      "tokens_seen": 127074304
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048559679037111337,
      "loss": 3.5238,
      "theoretical_loss": 4.618560368496466,
      "tokens_seen": 127139840
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048558676028084255,
      "loss": 3.4135,
      "theoretical_loss": 4.618241553384175,
      "tokens_seen": 127205376
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048557673019057173,
      "loss": 3.5748,
      "theoretical_loss": 4.617922948446459,
      "tokens_seen": 127270912
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004855667001003009,
      "loss": 3.2217,
      "theoretical_loss": 4.617604553436642,
      "tokens_seen": 127336448
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048555667001003015,
      "loss": 3.4823,
      "theoretical_loss": 4.617286368108466,
      "tokens_seen": 127401984
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004855466399197593,
      "loss": 3.5026,
      "theoretical_loss": 4.6169683922160925,
      "tokens_seen": 127467520
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004855366098294885,
      "loss": 3.3712,
      "theoretical_loss": 4.616650625514091,
      "tokens_seen": 127533056
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048552657973921764,
      "loss": 3.3915,
      "theoretical_loss": 4.616333067757449,
      "tokens_seen": 127598592
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048551654964894687,
      "loss": 3.5052,
      "theoretical_loss": 4.616015718701563,
      "tokens_seen": 127664128
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048550651955867605,
      "loss": 3.4679,
      "theoretical_loss": 4.615698578102245,
      "tokens_seen": 127729664
    },
    {
      "epoch": 0.04,
      "objective/train/docs_used": 209446,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.4149038791656494,
      "objective/train/theoretical_loss": 4.615381645715717,
      "objective/train/tokens_used": 148255200,
      "theoretical_loss": 4.615381645715717,
      "tokens_seen": 127795200
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048549648946840523,
      "loss": 3.3464,
      "theoretical_loss": 4.615381645715717,
      "tokens_seen": 127795200
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004854864593781344,
      "loss": 3.448,
      "theoretical_loss": 4.615064921298608,
      "tokens_seen": 127860736
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004854764292878636,
      "loss": 3.4669,
      "theoretical_loss": 4.61474840460796,
      "tokens_seen": 127926272
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004854663991975928,
      "loss": 3.2942,
      "theoretical_loss": 4.614432095401219,
      "tokens_seen": 127991808
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000485456369107322,
      "loss": 3.5544,
      "theoretical_loss": 4.614115993436242,
      "tokens_seen": 128057344
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048544633901705114,
      "loss": 3.6377,
      "theoretical_loss": 4.613800098471291,
      "tokens_seen": 128122880
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004854363089267804,
      "loss": 3.4673,
      "theoretical_loss": 4.613484410265032,
      "tokens_seen": 128188416
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004854262788365095,
      "loss": 3.3257,
      "theoretical_loss": 4.613168928576538,
      "tokens_seen": 128253952
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048541624874623874,
      "loss": 3.3088,
      "theoretical_loss": 4.612853653165283,
      "tokens_seen": 128319488
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004854062186559679,
      "loss": 3.3764,
      "theoretical_loss": 4.612538583791146,
      "tokens_seen": 128385024
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004853961885656971,
      "loss": 3.347,
      "theoretical_loss": 4.612223720214407,
      "tokens_seen": 128450560
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004853861584754263,
      "loss": 3.5082,
      "theoretical_loss": 4.611909062195749,
      "tokens_seen": 128516096
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004853761283851555,
      "loss": 3.5355,
      "theoretical_loss": 4.61159460949625,
      "tokens_seen": 128581632
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048536609829488464,
      "loss": 3.5775,
      "theoretical_loss": 4.611280361877393,
      "tokens_seen": 128647168
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004853560682046139,
      "loss": 3.4158,
      "theoretical_loss": 4.610966319101056,
      "tokens_seen": 128712704
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000485346038114343,
      "loss": 3.4332,
      "theoretical_loss": 4.610652480929515,
      "tokens_seen": 128778240
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048533600802407224,
      "loss": 3.5417,
      "theoretical_loss": 4.610338847125445,
      "tokens_seen": 128843776
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004853259779338014,
      "loss": 3.2011,
      "theoretical_loss": 4.610025417451913,
      "tokens_seen": 128909312
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004853159478435306,
      "loss": 3.3289,
      "theoretical_loss": 4.6097121916723856,
      "tokens_seen": 128974848
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004853059177532598,
      "loss": 3.4422,
      "theoretical_loss": 4.609399169550718,
      "tokens_seen": 129040384
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048529588766298897,
      "loss": 3.4443,
      "theoretical_loss": 4.609086350851165,
      "tokens_seen": 129105920
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048528585757271815,
      "loss": 3.5788,
      "theoretical_loss": 4.6087737353383655,
      "tokens_seen": 129171456
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004852758274824474,
      "loss": 3.2671,
      "theoretical_loss": 4.6084613227773605,
      "tokens_seen": 129236992
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004852657973921765,
      "loss": 3.4653,
      "theoretical_loss": 4.608149112933571,
      "tokens_seen": 129302528
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048525576730190574,
      "loss": 3.2752,
      "theoretical_loss": 4.607837105572816,
      "tokens_seen": 129368064
    },
    {
      "epoch": 0.04,
      "objective/train/docs_used": 210153,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.262286424636841,
      "objective/train/theoretical_loss": 4.607525300461299,
      "objective/train/tokens_used": 149893600,
      "theoretical_loss": 4.607525300461299,
      "tokens_seen": 129433600
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048524573721163487,
      "loss": 3.4158,
      "theoretical_loss": 4.607525300461299,
      "tokens_seen": 129433600
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004852357071213641,
      "loss": 3.84,
      "theoretical_loss": 4.607213697365613,
      "tokens_seen": 129499136
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004852256770310933,
      "loss": 3.2357,
      "theoretical_loss": 4.606902296052739,
      "tokens_seen": 129564672
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048521564694082247,
      "loss": 3.1954,
      "theoretical_loss": 4.6065910962900425,
      "tokens_seen": 129630208
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048520561685055165,
      "loss": 3.2395,
      "theoretical_loss": 4.606280097845277,
      "tokens_seen": 129695744
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004851955867602809,
      "loss": 3.6586,
      "theoretical_loss": 4.60596930048658,
      "tokens_seen": 129761280
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048518555667001,
      "loss": 3.8546,
      "theoretical_loss": 4.605658703982471,
      "tokens_seen": 129826816
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048517552657973925,
      "loss": 3.6046,
      "theoretical_loss": 4.6053483081018545,
      "tokens_seen": 129892352
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004851654964894684,
      "loss": 3.6551,
      "theoretical_loss": 4.605038112614018,
      "tokens_seen": 129957888
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004851554663991976,
      "loss": 3.7827,
      "theoretical_loss": 4.604728117288631,
      "tokens_seen": 130023424
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004851454363089268,
      "loss": 3.4682,
      "theoretical_loss": 4.604418321895739,
      "tokens_seen": 130088960
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048513540621865597,
      "loss": 3.4146,
      "theoretical_loss": 4.604108726205774,
      "tokens_seen": 130154496
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048512537612838515,
      "loss": 3.6829,
      "theoretical_loss": 4.60379932998954,
      "tokens_seen": 130220032
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048511534603811433,
      "loss": 3.2773,
      "theoretical_loss": 4.6034901330182265,
      "tokens_seen": 130285568
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004851053159478435,
      "loss": 3.441,
      "theoretical_loss": 4.603181135063394,
      "tokens_seen": 130351104
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048509528585757275,
      "loss": 3.4939,
      "theoretical_loss": 4.6028723358969845,
      "tokens_seen": 130416640
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004850852557673019,
      "loss": 3.6256,
      "theoretical_loss": 4.602563735291312,
      "tokens_seen": 130482176
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004850752256770311,
      "loss": 3.3061,
      "theoretical_loss": 4.602255333019068,
      "tokens_seen": 130547712
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048506519558676024,
      "loss": 3.4777,
      "theoretical_loss": 4.6019471288533165,
      "tokens_seen": 130613248
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004850551654964895,
      "loss": 3.5478,
      "theoretical_loss": 4.601639122567497,
      "tokens_seen": 130678784
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048504513540621866,
      "loss": 3.2939,
      "theoretical_loss": 4.601331313935418,
      "tokens_seen": 130744320
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048503510531594784,
      "loss": 3.3667,
      "theoretical_loss": 4.601023702731264,
      "tokens_seen": 130809856
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000485025075225677,
      "loss": 3.7578,
      "theoretical_loss": 4.600716288729587,
      "tokens_seen": 130875392
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048501504513540625,
      "loss": 3.3572,
      "theoretical_loss": 4.600409071705312,
      "tokens_seen": 130940928
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004850050150451354,
      "loss": 3.5029,
      "theoretical_loss": 4.60010205143373,
      "tokens_seen": 131006464
    },
    {
      "epoch": 0.04,
      "objective/train/docs_used": 211647,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.9844844341278076,
      "objective/train/theoretical_loss": 4.599795227690505,
      "objective/train/tokens_used": 151532000,
      "theoretical_loss": 4.599795227690505,
      "tokens_seen": 131072000
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004849949849548646,
      "loss": 3.4689,
      "theoretical_loss": 4.599795227690505,
      "tokens_seen": 131072000
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004849849548645938,
      "loss": 3.5353,
      "theoretical_loss": 4.5994886002516635,
      "tokens_seen": 131137536
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000484974924774323,
      "loss": 3.7795,
      "theoretical_loss": 4.599182168893604,
      "tokens_seen": 131203072
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004849648946840522,
      "loss": 3.42,
      "theoretical_loss": 4.598875933393089,
      "tokens_seen": 131268608
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048495486459378134,
      "loss": 3.4034,
      "theoretical_loss": 4.5985698935272445,
      "tokens_seen": 131334144
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004849448345035106,
      "loss": 3.5382,
      "theoretical_loss": 4.598264049073565,
      "tokens_seen": 131399680
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004849348044132397,
      "loss": 3.336,
      "theoretical_loss": 4.597958399809908,
      "tokens_seen": 131465216
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048492477432296894,
      "loss": 3.3876,
      "theoretical_loss": 4.59765294551449,
      "tokens_seen": 131530752
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004849147442326981,
      "loss": 3.418,
      "theoretical_loss": 4.597347685965897,
      "tokens_seen": 131596288
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004849047141424273,
      "loss": 3.4876,
      "theoretical_loss": 4.597042620943069,
      "tokens_seen": 131661824
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004848946840521565,
      "loss": 3.7115,
      "theoretical_loss": 4.596737750225311,
      "tokens_seen": 131727360
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004848846539618857,
      "loss": 3.6111,
      "theoretical_loss": 4.596433073592289,
      "tokens_seen": 131792896
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048487462387161484,
      "loss": 3.4538,
      "theoretical_loss": 4.596128590824026,
      "tokens_seen": 131858432
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004848645937813441,
      "loss": 3.2722,
      "theoretical_loss": 4.595824301700904,
      "tokens_seen": 131923968
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004848545636910732,
      "loss": 3.2836,
      "theoretical_loss": 4.595520206003663,
      "tokens_seen": 131989504
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048484453360080244,
      "loss": 3.2864,
      "theoretical_loss": 4.595216303513399,
      "tokens_seen": 132055040
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004848345035105316,
      "loss": 3.3418,
      "theoretical_loss": 4.594912594011566,
      "tokens_seen": 132120576
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004848244734202608,
      "loss": 3.573,
      "theoretical_loss": 4.594609077279973,
      "tokens_seen": 132186112
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048481444332999,
      "loss": 3.4152,
      "theoretical_loss": 4.594305753100782,
      "tokens_seen": 132251648
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048480441323971917,
      "loss": 3.2979,
      "theoretical_loss": 4.594002621256511,
      "tokens_seen": 132317184
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048479438314944835,
      "loss": 3.5056,
      "theoretical_loss": 4.59369968153003,
      "tokens_seen": 132382720
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004847843530591776,
      "loss": 3.6168,
      "theoretical_loss": 4.593396933704562,
      "tokens_seen": 132448256
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004847743229689067,
      "loss": 3.3142,
      "theoretical_loss": 4.593094377563681,
      "tokens_seen": 132513792
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048476429287863594,
      "loss": 3.3908,
      "theoretical_loss": 4.592792012891314,
      "tokens_seen": 132579328
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048475426278836507,
      "loss": 3.3265,
      "theoretical_loss": 4.592489839471735,
      "tokens_seen": 132644864
    },
    {
      "epoch": 0.04,
      "objective/train/docs_used": 212401,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.6346824169158936,
      "objective/train/theoretical_loss": 4.592187857089571,
      "objective/train/tokens_used": 153170400,
      "theoretical_loss": 4.592187857089571,
      "tokens_seen": 132710400
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004847442326980943,
      "loss": 3.6193,
      "theoretical_loss": 4.592187857089571,
      "tokens_seen": 132710400
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004847342026078235,
      "loss": 3.1198,
      "theoretical_loss": 4.591886065529795,
      "tokens_seen": 132775936
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048472417251755267,
      "loss": 3.5852,
      "theoretical_loss": 4.591584464577728,
      "tokens_seen": 132841472
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048471414242728185,
      "loss": 3.5391,
      "theoretical_loss": 4.591283054019041,
      "tokens_seen": 132907008
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004847041123370111,
      "loss": 3.7466,
      "theoretical_loss": 4.5909818336397485,
      "tokens_seen": 132972544
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004846940822467402,
      "loss": 3.5896,
      "theoretical_loss": 4.590680803226213,
      "tokens_seen": 133038080
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048468405215646945,
      "loss": 3.4774,
      "theoretical_loss": 4.590379962565141,
      "tokens_seen": 133103616
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004846740220661986,
      "loss": 3.5418,
      "theoretical_loss": 4.590079311443583,
      "tokens_seen": 133169152
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004846639919759278,
      "loss": 3.5195,
      "theoretical_loss": 4.589778849648934,
      "tokens_seen": 133234688
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000484653961885657,
      "loss": 3.3776,
      "theoretical_loss": 4.589478576968932,
      "tokens_seen": 133300224
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048464393179538617,
      "loss": 3.451,
      "theoretical_loss": 4.589178493191655,
      "tokens_seen": 133365760
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048463390170511535,
      "loss": 3.4017,
      "theoretical_loss": 4.588878598105527,
      "tokens_seen": 133431296
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048462387161484453,
      "loss": 3.1066,
      "theoretical_loss": 4.588578891499308,
      "tokens_seen": 133496832
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004846138415245737,
      "loss": 3.5051,
      "theoretical_loss": 4.588279373162101,
      "tokens_seen": 133562368
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048460381143430295,
      "loss": 3.3603,
      "theoretical_loss": 4.587980042883347,
      "tokens_seen": 133627904
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004845937813440321,
      "loss": 3.1542,
      "theoretical_loss": 4.587680900452824,
      "tokens_seen": 133693440
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004845837512537613,
      "loss": 3.3772,
      "theoretical_loss": 4.587381945660653,
      "tokens_seen": 133758976
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048457372116349044,
      "loss": 3.303,
      "theoretical_loss": 4.587083178297288,
      "tokens_seen": 133824512
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004845636910732197,
      "loss": 3.4483,
      "theoretical_loss": 4.5867845981535185,
      "tokens_seen": 133890048
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048455366098294886,
      "loss": 3.4031,
      "theoretical_loss": 4.586486205020474,
      "tokens_seen": 133955584
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048454363089267804,
      "loss": 3.4808,
      "theoretical_loss": 4.586187998689616,
      "tokens_seen": 134021120
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004845336008024072,
      "loss": 3.4596,
      "theoretical_loss": 4.585889978952741,
      "tokens_seen": 134086656
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048452357071213645,
      "loss": 3.2542,
      "theoretical_loss": 4.58559214560198,
      "tokens_seen": 134152192
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004845135406218656,
      "loss": 3.2377,
      "theoretical_loss": 4.585294498429796,
      "tokens_seen": 134217728
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004845035105315948,
      "loss": 3.4138,
      "theoretical_loss": 4.584997037228986,
      "tokens_seen": 134283264
    },
    {
      "epoch": 0.04,
      "objective/train/docs_used": 213682,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.553767681121826,
      "objective/train/theoretical_loss": 4.584699761792674,
      "objective/train/tokens_used": 154808800,
      "theoretical_loss": 4.584699761792674,
      "tokens_seen": 134348800
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048449348044132394,
      "loss": 3.4115,
      "theoretical_loss": 4.584699761792674,
      "tokens_seen": 134348800
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004844834503510532,
      "loss": 3.379,
      "theoretical_loss": 4.5844026719143205,
      "tokens_seen": 134414336
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048447342026078236,
      "loss": 3.2655,
      "theoretical_loss": 4.5841057673877135,
      "tokens_seen": 134479872
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048446339017051154,
      "loss": 3.6094,
      "theoretical_loss": 4.5838090480069695,
      "tokens_seen": 134545408
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004844533600802407,
      "loss": 3.4139,
      "theoretical_loss": 4.5835125135665375,
      "tokens_seen": 134610944
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004844433299899699,
      "loss": 3.241,
      "theoretical_loss": 4.583216163861191,
      "tokens_seen": 134676480
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004844332998996991,
      "loss": 3.4857,
      "theoretical_loss": 4.58291999868603,
      "tokens_seen": 134742016
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004844232698094283,
      "loss": 3.4424,
      "theoretical_loss": 4.582624017836489,
      "tokens_seen": 134807552
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048441323971915745,
      "loss": 3.5497,
      "theoretical_loss": 4.582328221108318,
      "tokens_seen": 134873088
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004844032096288867,
      "loss": 3.4341,
      "theoretical_loss": 4.5820326082976,
      "tokens_seen": 134938624
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048439317953861586,
      "loss": 3.4735,
      "theoretical_loss": 4.581737179200739,
      "tokens_seen": 135004160
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048438314944834504,
      "loss": 3.3947,
      "theoretical_loss": 4.581441933614466,
      "tokens_seen": 135069696
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004843731193580742,
      "loss": 3.6985,
      "theoretical_loss": 4.581146871335832,
      "tokens_seen": 135135232
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004843630892678034,
      "loss": 3.4447,
      "theoretical_loss": 4.580851992162214,
      "tokens_seen": 135200768
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004843530591775326,
      "loss": 3.2804,
      "theoretical_loss": 4.5805572958913086,
      "tokens_seen": 135266304
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004843430290872618,
      "loss": 3.2166,
      "theoretical_loss": 4.580262782321135,
      "tokens_seen": 135331840
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048433299899699095,
      "loss": 3.1952,
      "theoretical_loss": 4.579968451250032,
      "tokens_seen": 135397376
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004843229689067202,
      "loss": 3.6351,
      "theoretical_loss": 4.579674302476661,
      "tokens_seen": 135462912
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004843129388164493,
      "loss": 3.5732,
      "theoretical_loss": 4.579380335800001,
      "tokens_seen": 135528448
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048430290872617855,
      "loss": 3.2331,
      "theoretical_loss": 4.579086551019348,
      "tokens_seen": 135593984
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048429287863590773,
      "loss": 3.3013,
      "theoretical_loss": 4.5787929479343195,
      "tokens_seen": 135659520
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004842828485456369,
      "loss": 3.3744,
      "theoretical_loss": 4.578499526344848,
      "tokens_seen": 135725056
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004842728184553661,
      "loss": 3.4363,
      "theoretical_loss": 4.578206286051184,
      "tokens_seen": 135790592
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048426278836509527,
      "loss": 3.4997,
      "theoretical_loss": 4.5779132268538945,
      "tokens_seen": 135856128
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048425275827482445,
      "loss": 3.3039,
      "theoretical_loss": 4.577620348553859,
      "tokens_seen": 135921664
    },
    {
      "epoch": 0.04,
      "objective/train/docs_used": 214514,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.6967029571533203,
      "objective/train/theoretical_loss": 4.577327650952276,
      "objective/train/tokens_used": 156447200,
      "theoretical_loss": 4.577327650952276,
      "tokens_seen": 135987200
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004842427281845537,
      "loss": 3.7825,
      "theoretical_loss": 4.577327650952276,
      "tokens_seen": 135987200
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048423269809428287,
      "loss": 3.2603,
      "theoretical_loss": 4.5770351338506545,
      "tokens_seen": 136052736
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048422266800401205,
      "loss": 3.4404,
      "theoretical_loss": 4.57674279705082,
      "tokens_seen": 136118272
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004842126379137413,
      "loss": 3.39,
      "theoretical_loss": 4.57645064035491,
      "tokens_seen": 136183808
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004842026078234704,
      "loss": 3.1578,
      "theoretical_loss": 4.576158663565371,
      "tokens_seen": 136249344
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048419257773319965,
      "loss": 3.1521,
      "theoretical_loss": 4.575866866484967,
      "tokens_seen": 136314880
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004841825476429288,
      "loss": 3.3245,
      "theoretical_loss": 4.575575248916767,
      "tokens_seen": 136380416
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000484172517552658,
      "loss": 3.2286,
      "theoretical_loss": 4.575283810664155,
      "tokens_seen": 136445952
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004841624874623872,
      "loss": 3.1527,
      "theoretical_loss": 4.574992551530822,
      "tokens_seen": 136511488
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048415245737211637,
      "loss": 3.3615,
      "theoretical_loss": 4.574701471320768,
      "tokens_seen": 136577024
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048414242728184555,
      "loss": 3.2323,
      "theoretical_loss": 4.574410569838304,
      "tokens_seen": 136642560
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048413239719157473,
      "loss": 3.5178,
      "theoretical_loss": 4.574119846888045,
      "tokens_seen": 136708096
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004841223671013039,
      "loss": 3.4339,
      "theoretical_loss": 4.573829302274915,
      "tokens_seen": 136773632
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048411233701103315,
      "loss": 3.3785,
      "theoretical_loss": 4.573538935804146,
      "tokens_seen": 136839168
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004841023069207623,
      "loss": 3.4305,
      "theoretical_loss": 4.573248747281273,
      "tokens_seen": 136904704
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004840922768304915,
      "loss": 3.7156,
      "theoretical_loss": 4.5729587365121365,
      "tokens_seen": 136970240
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048408224674022064,
      "loss": 3.1826,
      "theoretical_loss": 4.572668903302886,
      "tokens_seen": 137035776
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004840722166499499,
      "loss": 3.291,
      "theoretical_loss": 4.572379247459969,
      "tokens_seen": 137101312
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048406218655967906,
      "loss": 3.4151,
      "theoretical_loss": 4.57208976879014,
      "tokens_seen": 137166848
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048405215646940824,
      "loss": 3.4274,
      "theoretical_loss": 4.571800467100456,
      "tokens_seen": 137232384
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004840421263791374,
      "loss": 3.5986,
      "theoretical_loss": 4.5715113421982725,
      "tokens_seen": 137297920
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048403209628886665,
      "loss": 3.6963,
      "theoretical_loss": 4.571222393891253,
      "tokens_seen": 137363456
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004840220661985958,
      "loss": 3.3796,
      "theoretical_loss": 4.570933621987356,
      "tokens_seen": 137428992
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000484012036108325,
      "loss": 3.1538,
      "theoretical_loss": 4.570645026294844,
      "tokens_seen": 137494528
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048400200601805414,
      "loss": 3.3521,
      "theoretical_loss": 4.570356606622278,
      "tokens_seen": 137560064
    },
    {
      "epoch": 0.04,
      "objective/train/docs_used": 215745,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.805065631866455,
      "objective/train/theoretical_loss": 4.570068362778516,
      "objective/train/tokens_used": 158085600,
      "theoretical_loss": 4.570068362778516,
      "tokens_seen": 137625600
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004839919759277834,
      "loss": 3.7218,
      "theoretical_loss": 4.570068362778516,
      "tokens_seen": 137625600
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048398194583751256,
      "loss": 3.2555,
      "theoretical_loss": 4.569780294572718,
      "tokens_seen": 137691136
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048397191574724174,
      "loss": 3.5768,
      "theoretical_loss": 4.569492401814339,
      "tokens_seen": 137756672
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004839618856569709,
      "loss": 3.6795,
      "theoretical_loss": 4.569204684313133,
      "tokens_seen": 137822208
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004839518555667001,
      "loss": 3.3141,
      "theoretical_loss": 4.568917141879149,
      "tokens_seen": 137887744
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004839418254764293,
      "loss": 3.3816,
      "theoretical_loss": 4.568629774322736,
      "tokens_seen": 137953280
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004839317953861585,
      "loss": 3.3705,
      "theoretical_loss": 4.568342581454532,
      "tokens_seen": 138018816
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048392176529588765,
      "loss": 3.5989,
      "theoretical_loss": 4.568055563085476,
      "tokens_seen": 138084352
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004839117352056169,
      "loss": 3.3891,
      "theoretical_loss": 4.567768719026797,
      "tokens_seen": 138149888
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048390170511534606,
      "loss": 3.2672,
      "theoretical_loss": 4.567482049090019,
      "tokens_seen": 138215424
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048389167502507524,
      "loss": 3.3265,
      "theoretical_loss": 4.567195553086961,
      "tokens_seen": 138280960
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004838816449348044,
      "loss": 3.2513,
      "theoretical_loss": 4.566909230829729,
      "tokens_seen": 138346496
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004838716148445336,
      "loss": 3.4663,
      "theoretical_loss": 4.566623082130729,
      "tokens_seen": 138412032
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004838615847542628,
      "loss": 3.5158,
      "theoretical_loss": 4.566337106802651,
      "tokens_seen": 138477568
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000483851554663992,
      "loss": 3.558,
      "theoretical_loss": 4.56605130465848,
      "tokens_seen": 138543104
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048384152457372115,
      "loss": 3.5835,
      "theoretical_loss": 4.565765675511487,
      "tokens_seen": 138608640
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004838314944834504,
      "loss": 3.4223,
      "theoretical_loss": 4.565480219175237,
      "tokens_seen": 138674176
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004838214643931795,
      "loss": 3.6202,
      "theoretical_loss": 4.56519493546358,
      "tokens_seen": 138739712
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048381143430290875,
      "loss": 3.3628,
      "theoretical_loss": 4.56490982419066,
      "tokens_seen": 138805248
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048380140421263793,
      "loss": 3.5243,
      "theoretical_loss": 4.564624885170902,
      "tokens_seen": 138870784
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004837913741223671,
      "loss": 3.2818,
      "theoretical_loss": 4.564340118219022,
      "tokens_seen": 138936320
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004837813440320963,
      "loss": 3.3265,
      "theoretical_loss": 4.56405552315002,
      "tokens_seen": 139001856
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048377131394182547,
      "loss": 3.5035,
      "theoretical_loss": 4.563771099779187,
      "tokens_seen": 139067392
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048376128385155465,
      "loss": 3.5528,
      "theoretical_loss": 4.563486847922093,
      "tokens_seen": 139132928
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004837512537612839,
      "loss": 3.5953,
      "theoretical_loss": 4.563202767394597,
      "tokens_seen": 139198464
    },
    {
      "epoch": 0.04,
      "objective/train/docs_used": 216543,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.1856539249420166,
      "objective/train/theoretical_loss": 4.562918858012843,
      "objective/train/tokens_used": 159724000,
      "theoretical_loss": 4.562918858012843,
      "tokens_seen": 139264000
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000483741223671013,
      "loss": 3.1963,
      "theoretical_loss": 4.562918858012843,
      "tokens_seen": 139264000
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048373119358074225,
      "loss": 3.7505,
      "theoretical_loss": 4.562635119593255,
      "tokens_seen": 139329536
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048372116349047143,
      "loss": 3.3943,
      "theoretical_loss": 4.562351551952542,
      "tokens_seen": 139395072
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004837111334002006,
      "loss": 3.427,
      "theoretical_loss": 4.5620681549076965,
      "tokens_seen": 139460608
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004837011033099298,
      "loss": 3.3626,
      "theoretical_loss": 4.561784928275992,
      "tokens_seen": 139526144
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000483691073219659,
      "loss": 3.2654,
      "theoretical_loss": 4.561501871874984,
      "tokens_seen": 139591680
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048368104312938816,
      "loss": 3.3665,
      "theoretical_loss": 4.561218985522507,
      "tokens_seen": 139657216
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004836710130391174,
      "loss": 3.4153,
      "theoretical_loss": 4.560936269036679,
      "tokens_seen": 139722752
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004836609829488465,
      "loss": 3.5138,
      "theoretical_loss": 4.560653722235895,
      "tokens_seen": 139788288
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048365095285857575,
      "loss": 3.5586,
      "theoretical_loss": 4.560371344938831,
      "tokens_seen": 139853824
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004836409227683049,
      "loss": 3.5356,
      "theoretical_loss": 4.560089136964439,
      "tokens_seen": 139919360
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004836308926780341,
      "loss": 3.6322,
      "theoretical_loss": 4.559807098131953,
      "tokens_seen": 139984896
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004836208625877633,
      "loss": 3.3173,
      "theoretical_loss": 4.559525228260882,
      "tokens_seen": 140050432
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004836108324974925,
      "loss": 3.5052,
      "theoretical_loss": 4.559243527171011,
      "tokens_seen": 140115968
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048360080240722166,
      "loss": 3.6296,
      "theoretical_loss": 4.558961994682403,
      "tokens_seen": 140181504
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048359077231695084,
      "loss": 3.6015,
      "theoretical_loss": 4.558680630615397,
      "tokens_seen": 140247040
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048358074222668,
      "loss": 3.2321,
      "theoretical_loss": 4.558399434790607,
      "tokens_seen": 140312576
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048357071213640926,
      "loss": 3.5938,
      "theoretical_loss": 4.558118407028921,
      "tokens_seen": 140378112
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004835606820461384,
      "loss": 3.2905,
      "theoretical_loss": 4.557837547151502,
      "tokens_seen": 140443648
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004835506519558676,
      "loss": 3.4072,
      "theoretical_loss": 4.557556854979786,
      "tokens_seen": 140509184
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004835406218655968,
      "loss": 3.4305,
      "theoretical_loss": 4.5572763303354815,
      "tokens_seen": 140574720
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000483530591775326,
      "loss": 3.4042,
      "theoretical_loss": 4.556995973040574,
      "tokens_seen": 140640256
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048352056168505516,
      "loss": 3.6395,
      "theoretical_loss": 4.556715782917314,
      "tokens_seen": 140705792
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048351053159478434,
      "loss": 3.4382,
      "theoretical_loss": 4.556435759788229,
      "tokens_seen": 140771328
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004835005015045135,
      "loss": 3.4248,
      "theoretical_loss": 4.556155903476114,
      "tokens_seen": 140836864
    },
    {
      "epoch": 0.04,
      "objective/train/docs_used": 217936,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.1173794269561768,
      "objective/train/theoretical_loss": 4.555876213804037,
      "objective/train/tokens_used": 161362400,
      "theoretical_loss": 4.555876213804037,
      "tokens_seen": 140902400
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048349047141424276,
      "loss": 3.3518,
      "theoretical_loss": 4.555876213804037,
      "tokens_seen": 140902400
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048348044132397194,
      "loss": 3.3859,
      "theoretical_loss": 4.555596690595333,
      "tokens_seen": 140967936
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004834704112337011,
      "loss": 3.2391,
      "theoretical_loss": 4.555317333673611,
      "tokens_seen": 141033472
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004834603811434303,
      "loss": 3.2149,
      "theoretical_loss": 4.555038142862742,
      "tokens_seen": 141099008
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004834503510531595,
      "loss": 3.5474,
      "theoretical_loss": 4.5547591179868725,
      "tokens_seen": 141164544
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004834403209628887,
      "loss": 3.4493,
      "theoretical_loss": 4.554480258870409,
      "tokens_seen": 141230080
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048343029087261785,
      "loss": 3.3747,
      "theoretical_loss": 4.554201565338033,
      "tokens_seen": 141295616
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004834202607823471,
      "loss": 3.6331,
      "theoretical_loss": 4.5539230372146875,
      "tokens_seen": 141361152
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048341023069207626,
      "loss": 3.3845,
      "theoretical_loss": 4.553644674325584,
      "tokens_seen": 141426688
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048340020060180544,
      "loss": 3.5964,
      "theoretical_loss": 4.553366476496198,
      "tokens_seen": 141492224
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004833901705115346,
      "loss": 3.3062,
      "theoretical_loss": 4.553088443552269,
      "tokens_seen": 141557760
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004833801404212638,
      "loss": 3.4501,
      "theoretical_loss": 4.552810575319806,
      "tokens_seen": 141623296
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000483370110330993,
      "loss": 3.2792,
      "theoretical_loss": 4.552532871625077,
      "tokens_seen": 141688832
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004833600802407222,
      "loss": 3.1857,
      "theoretical_loss": 4.5522553322946155,
      "tokens_seen": 141754368
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048335005015045135,
      "loss": 3.2261,
      "theoretical_loss": 4.551977957155217,
      "tokens_seen": 141819904
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004833400200601806,
      "loss": 3.6914,
      "theoretical_loss": 4.5517007460339425,
      "tokens_seen": 141885440
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004833299899699097,
      "loss": 3.3864,
      "theoretical_loss": 4.551423698758111,
      "tokens_seen": 141950976
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048331995987963895,
      "loss": 3.309,
      "theoretical_loss": 4.551146815155304,
      "tokens_seen": 142016512
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048330992978936813,
      "loss": 3.6274,
      "theoretical_loss": 4.550870095053366,
      "tokens_seen": 142082048
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004832998996990973,
      "loss": 3.3515,
      "theoretical_loss": 4.550593538280398,
      "tokens_seen": 142147584
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004832898696088265,
      "loss": 3.2589,
      "theoretical_loss": 4.550317144664766,
      "tokens_seen": 142213120
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048327983951855567,
      "loss": 3.4483,
      "theoretical_loss": 4.55004091403509,
      "tokens_seen": 142278656
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048326980942828485,
      "loss": 3.3192,
      "theoretical_loss": 4.5497648462202545,
      "tokens_seen": 142344192
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004832597793380141,
      "loss": 3.42,
      "theoretical_loss": 4.549488941049397,
      "tokens_seen": 142409728
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004832497492477432,
      "loss": 3.4013,
      "theoretical_loss": 4.549213198351914,
      "tokens_seen": 142475264
    },
    {
      "epoch": 0.04,
      "objective/train/docs_used": 218572,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.226104259490967,
      "objective/train/theoretical_loss": 4.548937617957463,
      "objective/train/tokens_used": 163000800,
      "theoretical_loss": 4.548937617957463,
      "tokens_seen": 142540800
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048323971915747245,
      "loss": 3.3476,
      "theoretical_loss": 4.548937617957463,
      "tokens_seen": 142540800
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048322968906720163,
      "loss": 3.1424,
      "theoretical_loss": 4.548662199695954,
      "tokens_seen": 142606336
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004832196589769308,
      "loss": 3.4049,
      "theoretical_loss": 4.548386943397556,
      "tokens_seen": 142671872
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048320962888666,
      "loss": 3.3425,
      "theoretical_loss": 4.548111848892693,
      "tokens_seen": 142737408
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004831995987963892,
      "loss": 3.6242,
      "theoretical_loss": 4.547836916012042,
      "tokens_seen": 142802944
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048318956870611836,
      "loss": 3.1838,
      "theoretical_loss": 4.547562144586539,
      "tokens_seen": 142868480
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004831795386158476,
      "loss": 3.4277,
      "theoretical_loss": 4.547287534447372,
      "tokens_seen": 142934016
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004831695085255767,
      "loss": 3.2606,
      "theoretical_loss": 4.5470130854259825,
      "tokens_seen": 142999552
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048315947843530595,
      "loss": 3.3748,
      "theoretical_loss": 4.546738797354065,
      "tokens_seen": 143065088
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004831494483450351,
      "loss": 3.4503,
      "theoretical_loss": 4.546464670063569,
      "tokens_seen": 143130624
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004831394182547643,
      "loss": 3.3705,
      "theoretical_loss": 4.546190703386695,
      "tokens_seen": 143196160
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004831293881644935,
      "loss": 3.4002,
      "theoretical_loss": 4.545916897155894,
      "tokens_seen": 143261696
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004831193580742227,
      "loss": 3.6006,
      "theoretical_loss": 4.54564325120387,
      "tokens_seen": 143327232
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048310932798395186,
      "loss": 3.614,
      "theoretical_loss": 4.545369765363578,
      "tokens_seen": 143392768
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048309929789368104,
      "loss": 3.3347,
      "theoretical_loss": 4.545096439468223,
      "tokens_seen": 143458304
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004830892678034102,
      "loss": 3.4975,
      "theoretical_loss": 4.544823273351257,
      "tokens_seen": 143523840
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048307923771313946,
      "loss": 3.5284,
      "theoretical_loss": 4.544550266846388,
      "tokens_seen": 143589376
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004830692076228686,
      "loss": 3.5193,
      "theoretical_loss": 4.544277419787566,
      "tokens_seen": 143654912
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004830591775325978,
      "loss": 3.3756,
      "theoretical_loss": 4.544004732008993,
      "tokens_seen": 143720448
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000483049147442327,
      "loss": 3.0692,
      "theoretical_loss": 4.543732203345119,
      "tokens_seen": 143785984
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004830391173520562,
      "loss": 3.3665,
      "theoretical_loss": 4.543459833630639,
      "tokens_seen": 143851520
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048302908726178536,
      "loss": 3.429,
      "theoretical_loss": 4.543187622700497,
      "tokens_seen": 143917056
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048301905717151454,
      "loss": 3.3939,
      "theoretical_loss": 4.542915570389884,
      "tokens_seen": 143982592
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004830090270812437,
      "loss": 3.4513,
      "theoretical_loss": 4.542643676534234,
      "tokens_seen": 144048128
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048299899699097296,
      "loss": 3.3302,
      "theoretical_loss": 4.542371940969231,
      "tokens_seen": 144113664
    },
    {
      "epoch": 0.04,
      "objective/train/docs_used": 219918,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.464306116104126,
      "objective/train/theoretical_loss": 4.542100363530799,
      "objective/train/tokens_used": 164639200,
      "theoretical_loss": 4.542100363530799,
      "tokens_seen": 144179200
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004829889669007021,
      "loss": 3.5182,
      "theoretical_loss": 4.542100363530799,
      "tokens_seen": 144179200
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004829789368104313,
      "loss": 3.3918,
      "theoretical_loss": 4.54182894405511,
      "tokens_seen": 144244736
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048296890672016045,
      "loss": 3.2309,
      "theoretical_loss": 4.5415576823785795,
      "tokens_seen": 144310272
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004829588766298897,
      "loss": 3.3555,
      "theoretical_loss": 4.541286578337866,
      "tokens_seen": 144375808
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048294884653961887,
      "loss": 3.485,
      "theoretical_loss": 4.541015631769872,
      "tokens_seen": 144441344
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048293881644934805,
      "loss": 3.382,
      "theoretical_loss": 4.5407448425117405,
      "tokens_seen": 144506880
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048292878635907723,
      "loss": 3.5303,
      "theoretical_loss": 4.540474210400859,
      "tokens_seen": 144572416
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048291875626880646,
      "loss": 3.3492,
      "theoretical_loss": 4.540203735274855,
      "tokens_seen": 144637952
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004829087261785356,
      "loss": 3.2747,
      "theoretical_loss": 4.5399334169716,
      "tokens_seen": 144703488
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004828986960882648,
      "loss": 3.4229,
      "theoretical_loss": 4.539663255329202,
      "tokens_seen": 144769024
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048288866599799395,
      "loss": 3.3941,
      "theoretical_loss": 4.539393250186015,
      "tokens_seen": 144834560
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004828786359077232,
      "loss": 3.5768,
      "theoretical_loss": 4.539123401380625,
      "tokens_seen": 144900096
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048286860581745237,
      "loss": 3.376,
      "theoretical_loss": 4.538853708751866,
      "tokens_seen": 144965632
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048285857572718155,
      "loss": 3.1908,
      "theoretical_loss": 4.538584172138804,
      "tokens_seen": 145031168
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048284854563691073,
      "loss": 3.4312,
      "theoretical_loss": 4.538314791380748,
      "tokens_seen": 145096704
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004828385155466399,
      "loss": 3.2382,
      "theoretical_loss": 4.538045566317242,
      "tokens_seen": 145162240
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004828284854563691,
      "loss": 3.2605,
      "theoretical_loss": 4.537776496788071,
      "tokens_seen": 145227776
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048281845536609833,
      "loss": 3.4595,
      "theoretical_loss": 4.537507582633253,
      "tokens_seen": 145293312
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048280842527582746,
      "loss": 3.4714,
      "theoretical_loss": 4.537238823693045,
      "tokens_seen": 145358848
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004827983951855567,
      "loss": 3.4331,
      "theoretical_loss": 4.536970219807939,
      "tokens_seen": 145424384
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004827883650952858,
      "loss": 3.3043,
      "theoretical_loss": 4.536701770818665,
      "tokens_seen": 145489920
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048277833500501505,
      "loss": 3.5044,
      "theoretical_loss": 4.536433476566185,
      "tokens_seen": 145555456
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048276830491474423,
      "loss": 3.4038,
      "theoretical_loss": 4.536165336891699,
      "tokens_seen": 145620992
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004827582748244734,
      "loss": 3.2575,
      "theoretical_loss": 4.535897351636638,
      "tokens_seen": 145686528
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004827482447342026,
      "loss": 3.6485,
      "theoretical_loss": 4.53562952064267,
      "tokens_seen": 145752064
    },
    {
      "epoch": 0.04,
      "objective/train/docs_used": 220709,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.8053040504455566,
      "objective/train/theoretical_loss": 4.535361843751696,
      "objective/train/tokens_used": 166277600,
      "theoretical_loss": 4.535361843751696,
      "tokens_seen": 145817600
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048273821464393183,
      "loss": 3.4441,
      "theoretical_loss": 4.535361843751696,
      "tokens_seen": 145817600
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000482728184553661,
      "loss": 3.1346,
      "theoretical_loss": 4.535094320805847,
      "tokens_seen": 145883136
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004827181544633902,
      "loss": 3.3252,
      "theoretical_loss": 4.534826951647489,
      "tokens_seen": 145948672
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004827081243731194,
      "loss": 3.4838,
      "theoretical_loss": 4.5345597361192205,
      "tokens_seen": 146014208
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048269809428284856,
      "loss": 3.6738,
      "theoretical_loss": 4.53429267406387,
      "tokens_seen": 146079744
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004826880641925778,
      "loss": 3.4136,
      "theoretical_loss": 4.5340257653244995,
      "tokens_seen": 146145280
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004826780341023069,
      "loss": 3.5642,
      "theoretical_loss": 4.5337590097444,
      "tokens_seen": 146210816
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048266800401203615,
      "loss": 3.2791,
      "theoretical_loss": 4.533492407167093,
      "tokens_seen": 146276352
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004826579739217653,
      "loss": 3.4304,
      "theoretical_loss": 4.53322595743633,
      "tokens_seen": 146341888
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004826479438314945,
      "loss": 3.4174,
      "theoretical_loss": 4.5329596603960916,
      "tokens_seen": 146407424
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004826379137412237,
      "loss": 3.3817,
      "theoretical_loss": 4.53269351589059,
      "tokens_seen": 146472960
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004826278836509529,
      "loss": 3.2871,
      "theoretical_loss": 4.532427523764261,
      "tokens_seen": 146538496
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048261785356068206,
      "loss": 3.303,
      "theoretical_loss": 4.532161683861773,
      "tokens_seen": 146604032
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048260782347041124,
      "loss": 3.5583,
      "theoretical_loss": 4.5318959960280205,
      "tokens_seen": 146669568
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004825977933801404,
      "loss": 3.0956,
      "theoretical_loss": 4.531630460108125,
      "tokens_seen": 146735104
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048258776328986966,
      "loss": 3.6116,
      "theoretical_loss": 4.531365075947434,
      "tokens_seen": 146800640
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004825777331995988,
      "loss": 3.4407,
      "theoretical_loss": 4.531099843391524,
      "tokens_seen": 146866176
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000482567703109328,
      "loss": 3.5475,
      "theoretical_loss": 4.5308347622861955,
      "tokens_seen": 146931712
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004825576730190572,
      "loss": 3.3681,
      "theoretical_loss": 4.5305698324774735,
      "tokens_seen": 146997248
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004825476429287864,
      "loss": 3.3835,
      "theoretical_loss": 4.530305053811611,
      "tokens_seen": 147062784
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048253761283851556,
      "loss": 3.2951,
      "theoretical_loss": 4.530040426135084,
      "tokens_seen": 147128320
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048252758274824474,
      "loss": 3.4417,
      "theoretical_loss": 4.529775949294593,
      "tokens_seen": 147193856
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004825175526579739,
      "loss": 3.1626,
      "theoretical_loss": 4.529511623137061,
      "tokens_seen": 147259392
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048250752256770316,
      "loss": 3.2913,
      "theoretical_loss": 4.529247447509637,
      "tokens_seen": 147324928
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004824974924774323,
      "loss": 3.4218,
      "theoretical_loss": 4.528983422259691,
      "tokens_seen": 147390464
    },
    {
      "epoch": 0.04,
      "objective/train/docs_used": 222333,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.542501926422119,
      "objective/train/theoretical_loss": 4.528719547234816,
      "objective/train/tokens_used": 167916000,
      "theoretical_loss": 4.528719547234816,
      "tokens_seen": 147456000
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004824874623871615,
      "loss": 3.6895,
      "theoretical_loss": 4.528719547234816,
      "tokens_seen": 147456000
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048247743229689065,
      "loss": 3.1272,
      "theoretical_loss": 4.528455822282828,
      "tokens_seen": 147521536
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004824674022066199,
      "loss": 3.3643,
      "theoretical_loss": 4.528192247251763,
      "tokens_seen": 147587072
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048245737211634907,
      "loss": 3.3044,
      "theoretical_loss": 4.52792882198988,
      "tokens_seen": 147652608
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048244734202607825,
      "loss": 3.2101,
      "theoretical_loss": 4.527665546345656,
      "tokens_seen": 147718144
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048243731193580743,
      "loss": 3.4948,
      "theoretical_loss": 4.5274024201677925,
      "tokens_seen": 147783680
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048242728184553666,
      "loss": 3.286,
      "theoretical_loss": 4.527139443305209,
      "tokens_seen": 147849216
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004824172517552658,
      "loss": 3.3363,
      "theoretical_loss": 4.526876615607042,
      "tokens_seen": 147914752
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000482407221664995,
      "loss": 3.5136,
      "theoretical_loss": 4.526613936922654,
      "tokens_seen": 147980288
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048239719157472415,
      "loss": 3.4325,
      "theoretical_loss": 4.526351407101618,
      "tokens_seen": 148045824
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004823871614844534,
      "loss": 3.3161,
      "theoretical_loss": 4.526089025993732,
      "tokens_seen": 148111360
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048237713139418257,
      "loss": 3.341,
      "theoretical_loss": 4.525826793449008,
      "tokens_seen": 148176896
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048236710130391175,
      "loss": 3.5225,
      "theoretical_loss": 4.525564709317678,
      "tokens_seen": 148242432
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048235707121364093,
      "loss": 3.511,
      "theoretical_loss": 4.525302773450187,
      "tokens_seen": 148307968
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004823470411233701,
      "loss": 3.2191,
      "theoretical_loss": 4.525040985697203,
      "tokens_seen": 148373504
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004823370110330993,
      "loss": 3.1906,
      "theoretical_loss": 4.524779345909604,
      "tokens_seen": 148439040
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048232698094282853,
      "loss": 3.4767,
      "theoretical_loss": 4.524517853938489,
      "tokens_seen": 148504576
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048231695085255766,
      "loss": 3.328,
      "theoretical_loss": 4.524256509635169,
      "tokens_seen": 148570112
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004823069207622869,
      "loss": 3.5906,
      "theoretical_loss": 4.523995312851174,
      "tokens_seen": 148635648
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000482296890672016,
      "loss": 3.1019,
      "theoretical_loss": 4.523734263438241,
      "tokens_seen": 148701184
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048228686058174525,
      "loss": 3.4911,
      "theoretical_loss": 4.52347336124833,
      "tokens_seen": 148766720
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048227683049147443,
      "loss": 3.5419,
      "theoretical_loss": 4.52321260613361,
      "tokens_seen": 148832256
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004822668004012036,
      "loss": 3.4556,
      "theoretical_loss": 4.522951997946466,
      "tokens_seen": 148897792
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004822567703109328,
      "loss": 3.438,
      "theoretical_loss": 4.522691536539492,
      "tokens_seen": 148963328
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048224674022066203,
      "loss": 3.5173,
      "theoretical_loss": 4.522431221765498,
      "tokens_seen": 149028864
    },
    {
      "epoch": 0.05,
      "objective/train/docs_used": 222973,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.872528314590454,
      "objective/train/theoretical_loss": 4.522171053477507,
      "objective/train/tokens_used": 169554400,
      "theoretical_loss": 4.522171053477507,
      "tokens_seen": 149094400
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048223671013039116,
      "loss": 3.1564,
      "theoretical_loss": 4.522171053477507,
      "tokens_seen": 149094400
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004822266800401204,
      "loss": 3.4409,
      "theoretical_loss": 4.5219110315287505,
      "tokens_seen": 149159936
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004822166499498495,
      "loss": 3.4656,
      "theoretical_loss": 4.521651155772675,
      "tokens_seen": 149225472
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048220661985957876,
      "loss": 3.5566,
      "theoretical_loss": 4.521391426062934,
      "tokens_seen": 149291008
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048219658976930794,
      "loss": 3.3884,
      "theoretical_loss": 4.521131842253396,
      "tokens_seen": 149356544
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004821865596790371,
      "loss": 3.6665,
      "theoretical_loss": 4.520872404198139,
      "tokens_seen": 149422080
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004821765295887663,
      "loss": 3.4762,
      "theoretical_loss": 4.520613111751445,
      "tokens_seen": 149487616
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004821664994984955,
      "loss": 3.3686,
      "theoretical_loss": 4.520353964767814,
      "tokens_seen": 149553152
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048215646940822466,
      "loss": 3.5076,
      "theoretical_loss": 4.5200949631019505,
      "tokens_seen": 149618688
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004821464393179539,
      "loss": 3.4631,
      "theoretical_loss": 4.519836106608768,
      "tokens_seen": 149684224
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000482136409227683,
      "loss": 3.2705,
      "theoretical_loss": 4.519577395143388,
      "tokens_seen": 149749760
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048212637913741226,
      "loss": 3.5931,
      "theoretical_loss": 4.519318828561142,
      "tokens_seen": 149815296
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004821163490471414,
      "loss": 3.3086,
      "theoretical_loss": 4.519060406717565,
      "tokens_seen": 149880832
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004821063189568706,
      "loss": 3.2906,
      "theoretical_loss": 4.518802129468405,
      "tokens_seen": 149946368
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004820962888665998,
      "loss": 3.3824,
      "theoretical_loss": 4.51854399666961,
      "tokens_seen": 150011904
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000482086258776329,
      "loss": 3.3789,
      "theoretical_loss": 4.518286008177341,
      "tokens_seen": 150077440
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048207622868605816,
      "loss": 3.284,
      "theoretical_loss": 4.51802816384796,
      "tokens_seen": 150142976
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004820661985957874,
      "loss": 3.2642,
      "theoretical_loss": 4.517770463538038,
      "tokens_seen": 150208512
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048205616850551653,
      "loss": 3.3741,
      "theoretical_loss": 4.517512907104347,
      "tokens_seen": 150274048
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048204613841524576,
      "loss": 3.4381,
      "theoretical_loss": 4.517255494403868,
      "tokens_seen": 150339584
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004820361083249749,
      "loss": 3.3189,
      "theoretical_loss": 4.516998225293785,
      "tokens_seen": 150405120
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004820260782347041,
      "loss": 3.3564,
      "theoretical_loss": 4.516741099631485,
      "tokens_seen": 150470656
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004820160481444333,
      "loss": 3.2927,
      "theoretical_loss": 4.51648411727456,
      "tokens_seen": 150536192
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004820060180541625,
      "loss": 3.1088,
      "theoretical_loss": 4.5162272780808035,
      "tokens_seen": 150601728
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048199598796389167,
      "loss": 3.3767,
      "theoretical_loss": 4.515970581908216,
      "tokens_seen": 150667264
    },
    {
      "epoch": 0.05,
      "objective/train/docs_used": 224256,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.578481435775757,
      "objective/train/theoretical_loss": 4.515714028614996,
      "objective/train/tokens_used": 171192800,
      "theoretical_loss": 4.515714028614996,
      "tokens_seen": 150732800
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048198595787362085,
      "loss": 3.3921,
      "theoretical_loss": 4.515714028614996,
      "tokens_seen": 150732800
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004819759277833501,
      "loss": 3.6899,
      "theoretical_loss": 4.515457618059546,
      "tokens_seen": 150798336
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048196589769307927,
      "loss": 3.365,
      "theoretical_loss": 4.515201350100471,
      "tokens_seen": 150863872
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048195586760280845,
      "loss": 3.4232,
      "theoretical_loss": 4.514945224596577,
      "tokens_seen": 150929408
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048194583751253763,
      "loss": 3.2746,
      "theoretical_loss": 4.5146892414068684,
      "tokens_seen": 150994944
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048193580742226686,
      "loss": 3.3062,
      "theoretical_loss": 4.514433400390554,
      "tokens_seen": 151060480
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000481925777331996,
      "loss": 3.5527,
      "theoretical_loss": 4.514177701407042,
      "tokens_seen": 151126016
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004819157472417252,
      "loss": 3.4158,
      "theoretical_loss": 4.51392214431594,
      "tokens_seen": 151191552
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048190571715145435,
      "loss": 3.3167,
      "theoretical_loss": 4.513666728977054,
      "tokens_seen": 151257088
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004818956870611836,
      "loss": 3.261,
      "theoretical_loss": 4.51341145525039,
      "tokens_seen": 151322624
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048188565697091277,
      "loss": 3.2799,
      "theoretical_loss": 4.513156322996155,
      "tokens_seen": 151388160
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048187562688064195,
      "loss": 3.0428,
      "theoretical_loss": 4.512901332074751,
      "tokens_seen": 151453696
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048186559679037113,
      "loss": 3.4207,
      "theoretical_loss": 4.5126464823467805,
      "tokens_seen": 151519232
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004818555667001003,
      "loss": 3.4559,
      "theoretical_loss": 4.512391773673042,
      "tokens_seen": 151584768
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004818455366098295,
      "loss": 3.4414,
      "theoretical_loss": 4.5121372059145335,
      "tokens_seen": 151650304
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048183550651955873,
      "loss": 3.1644,
      "theoretical_loss": 4.511882778932447,
      "tokens_seen": 151715840
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048182547642928786,
      "loss": 3.1517,
      "theoretical_loss": 4.511628492588174,
      "tokens_seen": 151781376
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004818154463390171,
      "loss": 3.3785,
      "theoretical_loss": 4.5113743467433,
      "tokens_seen": 151846912
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004818054162487462,
      "loss": 3.5209,
      "theoretical_loss": 4.511120341259608,
      "tokens_seen": 151912448
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048179538615847545,
      "loss": 3.5419,
      "theoretical_loss": 4.510866475999077,
      "tokens_seen": 151977984
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048178535606820463,
      "loss": 3.4455,
      "theoretical_loss": 4.510612750823878,
      "tokens_seen": 152043520
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004817753259779338,
      "loss": 3.1911,
      "theoretical_loss": 4.5103591655963795,
      "tokens_seen": 152109056
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000481765295887663,
      "loss": 3.3913,
      "theoretical_loss": 4.510105720179144,
      "tokens_seen": 152174592
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048175526579739223,
      "loss": 3.1391,
      "theoretical_loss": 4.5098524144349295,
      "tokens_seen": 152240128
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048174523570712136,
      "loss": 3.2863,
      "theoretical_loss": 4.509599248226683,
      "tokens_seen": 152305664
    },
    {
      "epoch": 0.05,
      "objective/train/docs_used": 224258,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.519541025161743,
      "objective/train/theoretical_loss": 4.509346221417552,
      "objective/train/tokens_used": 172831200,
      "theoretical_loss": 4.509346221417552,
      "tokens_seen": 152371200
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004817352056168506,
      "loss": 3.1877,
      "theoretical_loss": 4.509346221417552,
      "tokens_seen": 152371200
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004817251755265797,
      "loss": 3.5031,
      "theoretical_loss": 4.509093333870869,
      "tokens_seen": 152436736
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048171514543630896,
      "loss": 3.4969,
      "theoretical_loss": 4.508840585450166,
      "tokens_seen": 152502272
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048170511534603814,
      "loss": 3.4823,
      "theoretical_loss": 4.508587976019164,
      "tokens_seen": 152567808
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004816950852557673,
      "loss": 3.6068,
      "theoretical_loss": 4.508335505441774,
      "tokens_seen": 152633344
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004816850551654965,
      "loss": 3.2556,
      "theoretical_loss": 4.508083173582105,
      "tokens_seen": 152698880
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004816750250752257,
      "loss": 3.4433,
      "theoretical_loss": 4.507830980304451,
      "tokens_seen": 152764416
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048166499498495486,
      "loss": 3.4769,
      "theoretical_loss": 4.5075789254733,
      "tokens_seen": 152829952
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004816549648946841,
      "loss": 3.3988,
      "theoretical_loss": 4.507327008953329,
      "tokens_seen": 152895488
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004816449348044132,
      "loss": 3.4335,
      "theoretical_loss": 4.507075230609407,
      "tokens_seen": 152961024
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048163490471414246,
      "loss": 3.6205,
      "theoretical_loss": 4.506823590306591,
      "tokens_seen": 153026560
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004816248746238716,
      "loss": 3.2815,
      "theoretical_loss": 4.506572087910127,
      "tokens_seen": 153092096
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004816148445336008,
      "loss": 3.424,
      "theoretical_loss": 4.506320723285455,
      "tokens_seen": 153157632
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048160481444333,
      "loss": 3.234,
      "theoretical_loss": 4.506069496298198,
      "tokens_seen": 153223168
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004815947843530592,
      "loss": 3.5371,
      "theoretical_loss": 4.5058184068141705,
      "tokens_seen": 153288704
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048158475426278837,
      "loss": 3.4785,
      "theoretical_loss": 4.505567454699373,
      "tokens_seen": 153354240
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004815747241725176,
      "loss": 3.3644,
      "theoretical_loss": 4.505316639819997,
      "tokens_seen": 153419776
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048156469408224673,
      "loss": 3.3134,
      "theoretical_loss": 4.505065962042418,
      "tokens_seen": 153485312
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048155466399197596,
      "loss": 3.2391,
      "theoretical_loss": 4.504815421233202,
      "tokens_seen": 153550848
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004815446339017051,
      "loss": 3.4874,
      "theoretical_loss": 4.504565017259097,
      "tokens_seen": 153616384
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004815346038114343,
      "loss": 3.3429,
      "theoretical_loss": 4.504314749987044,
      "tokens_seen": 153681920
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004815245737211635,
      "loss": 3.2746,
      "theoretical_loss": 4.504064619284163,
      "tokens_seen": 153747456
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004815145436308927,
      "loss": 3.4142,
      "theoretical_loss": 4.503814625017766,
      "tokens_seen": 153812992
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048150451354062187,
      "loss": 3.1693,
      "theoretical_loss": 4.5035647670553445,
      "tokens_seen": 153878528
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048149448345035105,
      "loss": 3.3487,
      "theoretical_loss": 4.503315045264581,
      "tokens_seen": 153944064
    },
    {
      "epoch": 0.05,
      "objective/train/docs_used": 224258,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.4950404167175293,
      "objective/train/theoretical_loss": 4.503065459513339,
      "objective/train/tokens_used": 174469600,
      "theoretical_loss": 4.503065459513339,
      "tokens_seen": 154009600
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048148445336008023,
      "loss": 3.3584,
      "theoretical_loss": 4.503065459513339,
      "tokens_seen": 154009600
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048147442326980947,
      "loss": 3.3349,
      "theoretical_loss": 4.502816009669665,
      "tokens_seen": 154075136
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004814643931795386,
      "loss": 3.4341,
      "theoretical_loss": 4.502566695601795,
      "tokens_seen": 154140672
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048145436308926783,
      "loss": 3.5226,
      "theoretical_loss": 4.502317517178142,
      "tokens_seen": 154206208
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048144433299899696,
      "loss": 3.5991,
      "theoretical_loss": 4.502068474267309,
      "tokens_seen": 154271744
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004814343029087262,
      "loss": 3.3909,
      "theoretical_loss": 4.501819566738076,
      "tokens_seen": 154337280
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048142427281845537,
      "loss": 3.2495,
      "theoretical_loss": 4.501570794459411,
      "tokens_seen": 154402816
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048141424272818455,
      "loss": 3.3363,
      "theoretical_loss": 4.501322157300461,
      "tokens_seen": 154468352
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048140421263791373,
      "loss": 3.2894,
      "theoretical_loss": 4.501073655130554,
      "tokens_seen": 154533888
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048139418254764297,
      "loss": 3.352,
      "theoretical_loss": 4.500825287819205,
      "tokens_seen": 154599424
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004813841524573721,
      "loss": 3.268,
      "theoretical_loss": 4.500577055236104,
      "tokens_seen": 154664960
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048137412236710133,
      "loss": 3.3665,
      "theoretical_loss": 4.500328957251128,
      "tokens_seen": 154730496
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048136409227683046,
      "loss": 3.2677,
      "theoretical_loss": 4.500080993734329,
      "tokens_seen": 154796032
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004813540621865597,
      "loss": 3.2372,
      "theoretical_loss": 4.499833164555944,
      "tokens_seen": 154861568
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004813440320962889,
      "loss": 3.5336,
      "theoretical_loss": 4.499585469586387,
      "tokens_seen": 154927104
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048133400200601806,
      "loss": 3.2678,
      "theoretical_loss": 4.499337908696255,
      "tokens_seen": 154992640
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048132397191574724,
      "loss": 3.2836,
      "theoretical_loss": 4.499090481756321,
      "tokens_seen": 155058176
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004813139418254764,
      "loss": 3.4351,
      "theoretical_loss": 4.498843188637538,
      "tokens_seen": 155123712
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004813039117352056,
      "loss": 3.361,
      "theoretical_loss": 4.498596029211041,
      "tokens_seen": 155189248
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048129388164493483,
      "loss": 3.3694,
      "theoretical_loss": 4.498349003348137,
      "tokens_seen": 155254784
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048128385155466396,
      "loss": 3.4364,
      "theoretical_loss": 4.4981021109203185,
      "tokens_seen": 155320320
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004812738214643932,
      "loss": 3.3872,
      "theoretical_loss": 4.49785535179925,
      "tokens_seen": 155385856
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004812637913741223,
      "loss": 3.4293,
      "theoretical_loss": 4.497608725856776,
      "tokens_seen": 155451392
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048125376128385156,
      "loss": 3.2014,
      "theoretical_loss": 4.497362232964919,
      "tokens_seen": 155516928
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048124373119358074,
      "loss": 3.1822,
      "theoretical_loss": 4.497115872995876,
      "tokens_seen": 155582464
    },
    {
      "epoch": 0.05,
      "objective/train/docs_used": 225763,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.025784730911255,
      "objective/train/theoretical_loss": 4.496869645822022,
      "objective/train/tokens_used": 176108000,
      "theoretical_loss": 4.496869645822022,
      "tokens_seen": 155648000
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004812337011033099,
      "loss": 3.4056,
      "theoretical_loss": 4.496869645822022,
      "tokens_seen": 155648000
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048122367101303916,
      "loss": 3.2566,
      "theoretical_loss": 4.496623551315908,
      "tokens_seen": 155713536
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048121364092276834,
      "loss": 3.3016,
      "theoretical_loss": 4.496377589350261,
      "tokens_seen": 155779072
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004812036108324975,
      "loss": 3.3609,
      "theoretical_loss": 4.496131759797984,
      "tokens_seen": 155844608
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004811935807422267,
      "loss": 3.1468,
      "theoretical_loss": 4.495886062532153,
      "tokens_seen": 155910144
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004811835506519559,
      "loss": 3.6107,
      "theoretical_loss": 4.495640497426023,
      "tokens_seen": 155975680
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048117352056168506,
      "loss": 3.3142,
      "theoretical_loss": 4.495395064353019,
      "tokens_seen": 156041216
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004811634904714143,
      "loss": 3.3829,
      "theoretical_loss": 4.4951497631867445,
      "tokens_seen": 156106752
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004811534603811434,
      "loss": 3.4919,
      "theoretical_loss": 4.494904593800973,
      "tokens_seen": 156172288
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048114343029087266,
      "loss": 3.4889,
      "theoretical_loss": 4.4946595560696565,
      "tokens_seen": 156237824
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004811334002006018,
      "loss": 3.4904,
      "theoretical_loss": 4.494414649866915,
      "tokens_seen": 156303360
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000481123370110331,
      "loss": 3.24,
      "theoretical_loss": 4.494169875067046,
      "tokens_seen": 156368896
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004811133400200602,
      "loss": 3.2623,
      "theoretical_loss": 4.493925231544516,
      "tokens_seen": 156434432
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004811033099297894,
      "loss": 3.172,
      "theoretical_loss": 4.493680719173968,
      "tokens_seen": 156499968
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048109327983951857,
      "loss": 3.5037,
      "theoretical_loss": 4.4934363378302145,
      "tokens_seen": 156565504
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004810832497492478,
      "loss": 3.3995,
      "theoretical_loss": 4.493192087388239,
      "tokens_seen": 156631040
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048107321965897693,
      "loss": 3.2766,
      "theoretical_loss": 4.4929479677232,
      "tokens_seen": 156696576
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048106318956870616,
      "loss": 3.2331,
      "theoretical_loss": 4.4927039787104235,
      "tokens_seen": 156762112
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004810531594784353,
      "loss": 3.4322,
      "theoretical_loss": 4.4924601202254095,
      "tokens_seen": 156827648
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004810431293881645,
      "loss": 3.2955,
      "theoretical_loss": 4.492216392143826,
      "tokens_seen": 156893184
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004810330992978937,
      "loss": 3.5907,
      "theoretical_loss": 4.491972794341514,
      "tokens_seen": 156958720
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004810230692076229,
      "loss": 3.2891,
      "theoretical_loss": 4.49172932669448,
      "tokens_seen": 157024256
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048101303911735207,
      "loss": 3.4427,
      "theoretical_loss": 4.491485989078906,
      "tokens_seen": 157089792
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048100300902708125,
      "loss": 3.3958,
      "theoretical_loss": 4.491242781371138,
      "tokens_seen": 157155328
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048099297893681043,
      "loss": 3.4415,
      "theoretical_loss": 4.490999703447697,
      "tokens_seen": 157220864
    },
    {
      "epoch": 0.05,
      "objective/train/docs_used": 226317,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.2634286880493164,
      "objective/train/theoretical_loss": 4.4907567551852665,
      "objective/train/tokens_used": 177746400,
      "theoretical_loss": 4.4907567551852665,
      "tokens_seen": 157286400
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048098294884653967,
      "loss": 3.2088,
      "theoretical_loss": 4.4907567551852665,
      "tokens_seen": 157286400
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004809729187562688,
      "loss": 3.4287,
      "theoretical_loss": 4.490513936460702,
      "tokens_seen": 157351936
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048096288866599803,
      "loss": 3.0838,
      "theoretical_loss": 4.490271247151027,
      "tokens_seen": 157417472
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048095285857572716,
      "loss": 3.349,
      "theoretical_loss": 4.490028687133432,
      "tokens_seen": 157483008
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004809428284854564,
      "loss": 3.5244,
      "theoretical_loss": 4.489786256285276,
      "tokens_seen": 157548544
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048093279839518557,
      "loss": 3.3918,
      "theoretical_loss": 4.489543954484084,
      "tokens_seen": 157614080
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048092276830491475,
      "loss": 3.4131,
      "theoretical_loss": 4.489301781607551,
      "tokens_seen": 157679616
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048091273821464393,
      "loss": 3.3327,
      "theoretical_loss": 4.489059737533534,
      "tokens_seen": 157745152
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048090270812437317,
      "loss": 3.4728,
      "theoretical_loss": 4.48881782214006,
      "tokens_seen": 157810688
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004808926780341023,
      "loss": 3.2697,
      "theoretical_loss": 4.48857603530532,
      "tokens_seen": 157876224
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048088264794383153,
      "loss": 3.3249,
      "theoretical_loss": 4.488334376907673,
      "tokens_seen": 157941760
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048087261785356066,
      "loss": 3.4597,
      "theoretical_loss": 4.4880928468256425,
      "tokens_seen": 158007296
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004808625877632899,
      "loss": 3.5744,
      "theoretical_loss": 4.487851444937916,
      "tokens_seen": 158072832
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004808525576730191,
      "loss": 3.3033,
      "theoretical_loss": 4.487610171123347,
      "tokens_seen": 158138368
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048084252758274826,
      "loss": 3.3512,
      "theoretical_loss": 4.487369025260954,
      "tokens_seen": 158203904
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048083249749247744,
      "loss": 3.0467,
      "theoretical_loss": 4.48712800722992,
      "tokens_seen": 158269440
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004808224674022066,
      "loss": 3.4032,
      "theoretical_loss": 4.48688711690959,
      "tokens_seen": 158334976
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004808124373119358,
      "loss": 3.3788,
      "theoretical_loss": 4.486646354179475,
      "tokens_seen": 158400512
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048080240722166503,
      "loss": 3.4075,
      "theoretical_loss": 4.48640571891925,
      "tokens_seen": 158466048
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048079237713139416,
      "loss": 3.4903,
      "theoretical_loss": 4.48616521100875,
      "tokens_seen": 158531584
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004807823470411234,
      "loss": 3.3345,
      "theoretical_loss": 4.485924830327974,
      "tokens_seen": 158597120
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004807723169508525,
      "loss": 3.3844,
      "theoretical_loss": 4.485684576757087,
      "tokens_seen": 158662656
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048076228686058176,
      "loss": 3.4555,
      "theoretical_loss": 4.485444450176413,
      "tokens_seen": 158728192
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048075225677031094,
      "loss": 3.5851,
      "theoretical_loss": 4.485204450466437,
      "tokens_seen": 158793728
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004807422266800401,
      "loss": 3.4905,
      "theoretical_loss": 4.484964577507808,
      "tokens_seen": 158859264
    },
    {
      "epoch": 0.05,
      "objective/train/docs_used": 227589,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5188961029052734,
      "objective/train/theoretical_loss": 4.484724831181337,
      "objective/train/tokens_used": 179384800,
      "theoretical_loss": 4.484724831181337,
      "tokens_seen": 158924800
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004807321965897693,
      "loss": 2.9862,
      "theoretical_loss": 4.484724831181337,
      "tokens_seen": 158924800
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048072216649949854,
      "loss": 3.3628,
      "theoretical_loss": 4.4844852113679945,
      "tokens_seen": 158990336
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048071213640922766,
      "loss": 3.5553,
      "theoretical_loss": 4.484245717948913,
      "tokens_seen": 159055872
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004807021063189569,
      "loss": 3.0604,
      "theoretical_loss": 4.484006350805385,
      "tokens_seen": 159121408
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000480692076228686,
      "loss": 3.1705,
      "theoretical_loss": 4.483767109818862,
      "tokens_seen": 159186944
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048068204613841526,
      "loss": 3.4456,
      "theoretical_loss": 4.483527994870958,
      "tokens_seen": 159252480
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048067201604814444,
      "loss": 2.9752,
      "theoretical_loss": 4.483289005843445,
      "tokens_seen": 159318016
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004806619859578736,
      "loss": 3.2205,
      "theoretical_loss": 4.483050142618255,
      "tokens_seen": 159383552
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004806519558676028,
      "loss": 3.5844,
      "theoretical_loss": 4.482811405077482,
      "tokens_seen": 159449088
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000480641925777332,
      "loss": 3.5894,
      "theoretical_loss": 4.482572793103373,
      "tokens_seen": 159514624
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048063189568706117,
      "loss": 3.3414,
      "theoretical_loss": 4.482334306578339,
      "tokens_seen": 159580160
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004806218655967904,
      "loss": 3.3352,
      "theoretical_loss": 4.482095945384946,
      "tokens_seen": 159645696
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048061183550651953,
      "loss": 3.4557,
      "theoretical_loss": 4.481857709405919,
      "tokens_seen": 159711232
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048060180541624877,
      "loss": 3.2868,
      "theoretical_loss": 4.4816195985241425,
      "tokens_seen": 159776768
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048059177532597795,
      "loss": 3.1668,
      "theoretical_loss": 4.481381612622657,
      "tokens_seen": 159842304
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048058174523570713,
      "loss": 3.4842,
      "theoretical_loss": 4.481143751584659,
      "tokens_seen": 159907840
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004805717151454363,
      "loss": 3.4874,
      "theoretical_loss": 4.480906015293505,
      "tokens_seen": 159973376
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004805616850551655,
      "loss": 3.6502,
      "theoretical_loss": 4.480668403632706,
      "tokens_seen": 160038912
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048055165496489467,
      "loss": 3.4696,
      "theoretical_loss": 4.480430916485929,
      "tokens_seen": 160104448
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004805416248746239,
      "loss": 3.4085,
      "theoretical_loss": 4.480193553736999,
      "tokens_seen": 160169984
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048053159478435303,
      "loss": 3.296,
      "theoretical_loss": 4.479956315269897,
      "tokens_seen": 160235520
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048052156469408227,
      "loss": 3.3078,
      "theoretical_loss": 4.479719200968757,
      "tokens_seen": 160301056
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004805115346038114,
      "loss": 3.4221,
      "theoretical_loss": 4.479482210717871,
      "tokens_seen": 160366592
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048050150451354063,
      "loss": 3.3827,
      "theoretical_loss": 4.479245344401685,
      "tokens_seen": 160432128
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004804914744232698,
      "loss": 3.3653,
      "theoretical_loss": 4.479008601904798,
      "tokens_seen": 160497664
    },
    {
      "epoch": 0.05,
      "objective/train/docs_used": 228344,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.3748011589050293,
      "objective/train/theoretical_loss": 4.478771983111967,
      "objective/train/tokens_used": 181023200,
      "theoretical_loss": 4.478771983111967,
      "tokens_seen": 160563200
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000480481444332999,
      "loss": 3.298,
      "theoretical_loss": 4.478771983111967,
      "tokens_seen": 160563200
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048047141424272823,
      "loss": 3.2943,
      "theoretical_loss": 4.478535487908101,
      "tokens_seen": 160628736
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048046138415245736,
      "loss": 3.4003,
      "theoretical_loss": 4.478299116178265,
      "tokens_seen": 160694272
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004804513540621866,
      "loss": 3.6019,
      "theoretical_loss": 4.478062867807674,
      "tokens_seen": 160759808
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048044132397191577,
      "loss": 3.4234,
      "theoretical_loss": 4.4778267426817,
      "tokens_seen": 160825344
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048043129388164495,
      "loss": 3.3492,
      "theoretical_loss": 4.477590740685867,
      "tokens_seen": 160890880
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048042126379137413,
      "loss": 3.3196,
      "theoretical_loss": 4.47735486170585,
      "tokens_seen": 160956416
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048041123370110337,
      "loss": 3.2311,
      "theoretical_loss": 4.47711910562748,
      "tokens_seen": 161021952
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004804012036108325,
      "loss": 3.2435,
      "theoretical_loss": 4.4768834723367394,
      "tokens_seen": 161087488
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048039117352056173,
      "loss": 3.3293,
      "theoretical_loss": 4.4766479617197605,
      "tokens_seen": 161153024
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048038114343029086,
      "loss": 3.3614,
      "theoretical_loss": 4.476412573662829,
      "tokens_seen": 161218560
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004803711133400201,
      "loss": 3.4204,
      "theoretical_loss": 4.4761773080523835,
      "tokens_seen": 161284096
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004803610832497493,
      "loss": 3.266,
      "theoretical_loss": 4.475942164775013,
      "tokens_seen": 161349632
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048035105315947846,
      "loss": 3.3162,
      "theoretical_loss": 4.475707143717455,
      "tokens_seen": 161415168
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048034102306920764,
      "loss": 3.2002,
      "theoretical_loss": 4.475472244766601,
      "tokens_seen": 161480704
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004803309929789368,
      "loss": 3.2076,
      "theoretical_loss": 4.475237467809492,
      "tokens_seen": 161546240
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000480320962888666,
      "loss": 3.3402,
      "theoretical_loss": 4.47500281273332,
      "tokens_seen": 161611776
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048031093279839523,
      "loss": 3.405,
      "theoretical_loss": 4.474768279425424,
      "tokens_seen": 161677312
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048030090270812436,
      "loss": 3.3508,
      "theoretical_loss": 4.474533867773299,
      "tokens_seen": 161742848
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004802908726178536,
      "loss": 3.2164,
      "theoretical_loss": 4.474299577664581,
      "tokens_seen": 161808384
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004802808425275827,
      "loss": 3.2979,
      "theoretical_loss": 4.474065408987063,
      "tokens_seen": 161873920
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048027081243731196,
      "loss": 3.16,
      "theoretical_loss": 4.473831361628682,
      "tokens_seen": 161939456
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048026078234704114,
      "loss": 3.4809,
      "theoretical_loss": 4.473597435477526,
      "tokens_seen": 162004992
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004802507522567703,
      "loss": 3.3,
      "theoretical_loss": 4.473363630421831,
      "tokens_seen": 162070528
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004802407221664995,
      "loss": 3.0521,
      "theoretical_loss": 4.473129946349982,
      "tokens_seen": 162136064
    },
    {
      "epoch": 0.05,
      "objective/train/docs_used": 229659,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.4476068019866943,
      "objective/train/theoretical_loss": 4.472896383150508,
      "objective/train/tokens_used": 182661600,
      "theoretical_loss": 4.472896383150508,
      "tokens_seen": 162201600
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048023069207622874,
      "loss": 3.5461,
      "theoretical_loss": 4.472896383150508,
      "tokens_seen": 162201600
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048022066198595786,
      "loss": 3.4397,
      "theoretical_loss": 4.472662940712091,
      "tokens_seen": 162267136
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004802106318956871,
      "loss": 3.5161,
      "theoretical_loss": 4.472429618923558,
      "tokens_seen": 162332672
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048020060180541623,
      "loss": 3.2219,
      "theoretical_loss": 4.472196417673883,
      "tokens_seen": 162398208
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048019057171514546,
      "loss": 3.2822,
      "theoretical_loss": 4.471963336852187,
      "tokens_seen": 162463744
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048018054162487464,
      "loss": 3.2824,
      "theoretical_loss": 4.471730376347738,
      "tokens_seen": 162529280
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004801705115346038,
      "loss": 3.3961,
      "theoretical_loss": 4.4714975360499505,
      "tokens_seen": 162594816
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000480160481444333,
      "loss": 3.5644,
      "theoretical_loss": 4.471264815848384,
      "tokens_seen": 162660352
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004801504513540622,
      "loss": 3.3472,
      "theoretical_loss": 4.471032215632746,
      "tokens_seen": 162725888
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048014042126379137,
      "loss": 3.6874,
      "theoretical_loss": 4.470799735292889,
      "tokens_seen": 162791424
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004801303911735206,
      "loss": 3.1886,
      "theoretical_loss": 4.470567374718808,
      "tokens_seen": 162856960
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048012036108324973,
      "loss": 3.1756,
      "theoretical_loss": 4.470335133800649,
      "tokens_seen": 162922496
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048011033099297897,
      "loss": 3.3402,
      "theoretical_loss": 4.470103012428696,
      "tokens_seen": 162988032
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048010030090270815,
      "loss": 3.2792,
      "theoretical_loss": 4.469871010493383,
      "tokens_seen": 163053568
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048009027081243733,
      "loss": 3.3714,
      "theoretical_loss": 4.469639127885287,
      "tokens_seen": 163119104
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004800802407221665,
      "loss": 3.269,
      "theoretical_loss": 4.4694073644951295,
      "tokens_seen": 163184640
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004800702106318957,
      "loss": 3.2126,
      "theoretical_loss": 4.469175720213771,
      "tokens_seen": 163250176
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048006018054162487,
      "loss": 3.2064,
      "theoretical_loss": 4.468944194932225,
      "tokens_seen": 163315712
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004800501504513541,
      "loss": 3.6002,
      "theoretical_loss": 4.468712788541639,
      "tokens_seen": 163381248
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048004012036108323,
      "loss": 3.4055,
      "theoretical_loss": 4.46848150093331,
      "tokens_seen": 163446784
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048003009027081247,
      "loss": 3.4893,
      "theoretical_loss": 4.468250331998676,
      "tokens_seen": 163512320
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004800200601805416,
      "loss": 3.422,
      "theoretical_loss": 4.468019281629316,
      "tokens_seen": 163577856
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048001003009027083,
      "loss": 3.19,
      "theoretical_loss": 4.467788349716955,
      "tokens_seen": 163643392
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048,
      "loss": 3.137,
      "theoretical_loss": 4.467557536153457,
      "tokens_seen": 163708928
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004799899699097292,
      "loss": 3.2519,
      "theoretical_loss": 4.467326840830829,
      "tokens_seen": 163774464
    },
    {
      "debugging/Self-BLEU-5": 0.28281542061774223,
      "debugging/distinct-1-grams": 0.7326597736651554,
      "debugging/distinct-2-grams": 0.9067393674201281,
      "debugging/entropy-1-grams": 5.525931828352874,
      "debugging/entropy-2-grams": 6.085041606065698,
      "debugging/length": 468.25,
      "debugging/num_segments": 8,
      "debugging/score": 0.00998370262687916,
      "debugging/score_std": 0.008263218592709398,
      "epoch": 0.05,
      "objective/train/docs_used": 230304,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.3832874298095703,
      "objective/train/theoretical_loss": 4.467096263641219,
      "objective/train/tokens_used": 184300000,
      "theoretical_loss": 4.467096263641219,
      "tokens_seen": 163840000
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004799799398194584,
      "loss": 3.1978,
      "theoretical_loss": 4.467096263641219,
      "tokens_seen": 163840000
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047996990972918756,
      "loss": 3.112,
      "theoretical_loss": 4.466865804476919,
      "tokens_seen": 163905536
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047995987963891674,
      "loss": 3.2774,
      "theoretical_loss": 4.466635463230359,
      "tokens_seen": 163971072
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047994984954864597,
      "loss": 3.3353,
      "theoretical_loss": 4.466405239794113,
      "tokens_seen": 164036608
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004799398194583751,
      "loss": 3.2677,
      "theoretical_loss": 4.466175134060894,
      "tokens_seen": 164102144
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047992978936810433,
      "loss": 3.4542,
      "theoretical_loss": 4.465945145923554,
      "tokens_seen": 164167680
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004799197592778335,
      "loss": 3.4061,
      "theoretical_loss": 4.4657152752750875,
      "tokens_seen": 164233216
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004799097291875627,
      "loss": 3.3211,
      "theoretical_loss": 4.465485522008629,
      "tokens_seen": 164298752
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004798996990972919,
      "loss": 3.1246,
      "theoretical_loss": 4.465255886017452,
      "tokens_seen": 164364288
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047988966900702106,
      "loss": 3.3521,
      "theoretical_loss": 4.465026367194971,
      "tokens_seen": 164429824
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047987963891675024,
      "loss": 3.406,
      "theoretical_loss": 4.464796965434738,
      "tokens_seen": 164495360
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004798696088264795,
      "loss": 3.2096,
      "theoretical_loss": 4.464567680630443,
      "tokens_seen": 164560896
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004798595787362086,
      "loss": 3.3375,
      "theoretical_loss": 4.464338512675919,
      "tokens_seen": 164626432
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047984954864593784,
      "loss": 3.3015,
      "theoretical_loss": 4.464109461465133,
      "tokens_seen": 164691968
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047983951855566696,
      "loss": 3.1817,
      "theoretical_loss": 4.4638805268921935,
      "tokens_seen": 164757504
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004798294884653962,
      "loss": 3.3159,
      "theoretical_loss": 4.463651708851346,
      "tokens_seen": 164823040
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004798194583751254,
      "loss": 3.3364,
      "theoretical_loss": 4.463423007236974,
      "tokens_seen": 164888576
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047980942828485456,
      "loss": 3.5,
      "theoretical_loss": 4.4631944219436,
      "tokens_seen": 164954112
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047979939819458374,
      "loss": 3.5072,
      "theoretical_loss": 4.462965952865879,
      "tokens_seen": 165019648
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004797893681043129,
      "loss": 3.541,
      "theoretical_loss": 4.46273759989861,
      "tokens_seen": 165085184
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004797793380140421,
      "loss": 3.034,
      "theoretical_loss": 4.462509362936723,
      "tokens_seen": 165150720
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047976930792377134,
      "loss": 3.3458,
      "theoretical_loss": 4.46228124187529,
      "tokens_seen": 165216256
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047975927783350047,
      "loss": 3.3679,
      "theoretical_loss": 4.462053236609516,
      "tokens_seen": 165281792
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004797492477432297,
      "loss": 3.3828,
      "theoretical_loss": 4.461825347034742,
      "tokens_seen": 165347328
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004797392176529589,
      "loss": 3.3819,
      "theoretical_loss": 4.461597573046449,
      "tokens_seen": 165412864
    },
    {
      "epoch": 0.05,
      "objective/train/docs_used": 231649,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.920825958251953,
      "objective/train/theoretical_loss": 4.461369914540247,
      "objective/train/tokens_used": 185938400,
      "theoretical_loss": 4.461369914540247,
      "tokens_seen": 165478400
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047972918756268807,
      "loss": 3.1735,
      "theoretical_loss": 4.461369914540247,
      "tokens_seen": 165478400
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004797191574724173,
      "loss": 3.4608,
      "theoretical_loss": 4.4611423714118885,
      "tokens_seen": 165543936
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047970912738214643,
      "loss": 3.1453,
      "theoretical_loss": 4.460914943557256,
      "tokens_seen": 165609472
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047969909729187566,
      "loss": 3.3385,
      "theoretical_loss": 4.460687630872371,
      "tokens_seen": 165675008
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047968906720160484,
      "loss": 3.3347,
      "theoretical_loss": 4.46046043325339,
      "tokens_seen": 165740544
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000479679037111334,
      "loss": 3.1779,
      "theoretical_loss": 4.460233350596599,
      "tokens_seen": 165806080
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004796690070210632,
      "loss": 3.3418,
      "theoretical_loss": 4.460006382798425,
      "tokens_seen": 165871616
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004796589769307924,
      "loss": 3.1079,
      "theoretical_loss": 4.459779529755423,
      "tokens_seen": 165937152
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047964894684052157,
      "loss": 3.378,
      "theoretical_loss": 4.459552791364288,
      "tokens_seen": 166002688
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004796389167502508,
      "loss": 3.5243,
      "theoretical_loss": 4.459326167521844,
      "tokens_seen": 166068224
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047962888665997993,
      "loss": 3.4306,
      "theoretical_loss": 4.4590996581250515,
      "tokens_seen": 166133760
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047961885656970917,
      "loss": 3.3535,
      "theoretical_loss": 4.458873263071002,
      "tokens_seen": 166199296
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047960882647943835,
      "loss": 3.311,
      "theoretical_loss": 4.458646982256921,
      "tokens_seen": 166264832
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047959879638916753,
      "loss": 3.0939,
      "theoretical_loss": 4.458420815580169,
      "tokens_seen": 166330368
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004795887662988967,
      "loss": 3.1869,
      "theoretical_loss": 4.458194762938234,
      "tokens_seen": 166395904
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004795787362086259,
      "loss": 3.3403,
      "theoretical_loss": 4.457968824228743,
      "tokens_seen": 166461440
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047956870611835507,
      "loss": 3.3175,
      "theoretical_loss": 4.457742999349449,
      "tokens_seen": 166526976
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004795586760280843,
      "loss": 3.4267,
      "theoretical_loss": 4.4575172881982414,
      "tokens_seen": 166592512
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047954864593781343,
      "loss": 3.3691,
      "theoretical_loss": 4.457291690673139,
      "tokens_seen": 166658048
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047953861584754267,
      "loss": 3.4988,
      "theoretical_loss": 4.457066206672291,
      "tokens_seen": 166723584
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004795285857572718,
      "loss": 3.1917,
      "theoretical_loss": 4.456840836093983,
      "tokens_seen": 166789120
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047951855566700103,
      "loss": 3.1793,
      "theoretical_loss": 4.456615578836625,
      "tokens_seen": 166854656
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004795085255767302,
      "loss": 3.4376,
      "theoretical_loss": 4.456390434798762,
      "tokens_seen": 166920192
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004794984954864594,
      "loss": 3.3839,
      "theoretical_loss": 4.45616540387907,
      "tokens_seen": 166985728
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004794884653961886,
      "loss": 3.4359,
      "theoretical_loss": 4.4559404859763525,
      "tokens_seen": 167051264
    },
    {
      "epoch": 0.05,
      "objective/train/docs_used": 232239,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.8191466331481934,
      "objective/train/theoretical_loss": 4.455715680989545,
      "objective/train/tokens_used": 187576800,
      "theoretical_loss": 4.455715680989545,
      "tokens_seen": 167116800
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047947843530591776,
      "loss": 3.3247,
      "theoretical_loss": 4.455715680989545,
      "tokens_seen": 167116800
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047946840521564694,
      "loss": 3.442,
      "theoretical_loss": 4.455490988817713,
      "tokens_seen": 167182336
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047945837512537617,
      "loss": 3.4273,
      "theoretical_loss": 4.4552664093600525,
      "tokens_seen": 167247872
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004794483450351053,
      "loss": 3.3412,
      "theoretical_loss": 4.455041942515887,
      "tokens_seen": 167313408
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047943831494483453,
      "loss": 3.2788,
      "theoretical_loss": 4.454817588184669,
      "tokens_seen": 167378944
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004794282848545637,
      "loss": 3.3126,
      "theoretical_loss": 4.454593346265984,
      "tokens_seen": 167444480
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004794182547642929,
      "loss": 3.5354,
      "theoretical_loss": 4.454369216659542,
      "tokens_seen": 167510016
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004794082246740221,
      "loss": 3.2904,
      "theoretical_loss": 4.454145199265183,
      "tokens_seen": 167575552
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047939819458375126,
      "loss": 3.3353,
      "theoretical_loss": 4.453921293982877,
      "tokens_seen": 167641088
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047938816449348044,
      "loss": 3.3379,
      "theoretical_loss": 4.453697500712722,
      "tokens_seen": 167706624
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004793781344032097,
      "loss": 3.1888,
      "theoretical_loss": 4.453473819354942,
      "tokens_seen": 167772160
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004793681043129388,
      "loss": 3.3098,
      "theoretical_loss": 4.453250249809889,
      "tokens_seen": 167837696
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047935807422266804,
      "loss": 3.3648,
      "theoretical_loss": 4.453026791978045,
      "tokens_seen": 167903232
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047934804413239716,
      "loss": 3.4082,
      "theoretical_loss": 4.4528034457600185,
      "tokens_seen": 167968768
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004793380140421264,
      "loss": 3.472,
      "theoretical_loss": 4.452580211056542,
      "tokens_seen": 168034304
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004793279839518556,
      "loss": 3.1374,
      "theoretical_loss": 4.452357087768481,
      "tokens_seen": 168099840
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047931795386158476,
      "loss": 3.3576,
      "theoretical_loss": 4.45213407579682,
      "tokens_seen": 168165376
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047930792377131394,
      "loss": 3.4063,
      "theoretical_loss": 4.451911175042679,
      "tokens_seen": 168230912
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004792978936810431,
      "loss": 3.5888,
      "theoretical_loss": 4.451688385407296,
      "tokens_seen": 168296448
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004792878635907723,
      "loss": 3.4353,
      "theoretical_loss": 4.451465706792041,
      "tokens_seen": 168361984
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047927783350050154,
      "loss": 3.6028,
      "theoretical_loss": 4.4512431390984055,
      "tokens_seen": 168427520
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047926780341023067,
      "loss": 3.5768,
      "theoretical_loss": 4.451020682228011,
      "tokens_seen": 168493056
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004792577733199599,
      "loss": 3.1725,
      "theoretical_loss": 4.450798336082601,
      "tokens_seen": 168558592
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004792477432296891,
      "loss": 3.0739,
      "theoretical_loss": 4.450576100564046,
      "tokens_seen": 168624128
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047923771313941827,
      "loss": 3.16,
      "theoretical_loss": 4.450353975574341,
      "tokens_seen": 168689664
    },
    {
      "epoch": 0.05,
      "objective/train/docs_used": 233635,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.160566568374634,
      "objective/train/theoretical_loss": 4.450131961015606,
      "objective/train/tokens_used": 189215200,
      "theoretical_loss": 4.450131961015606,
      "tokens_seen": 168755200
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047922768304914745,
      "loss": 3.2785,
      "theoretical_loss": 4.450131961015606,
      "tokens_seen": 168755200
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047921765295887663,
      "loss": 3.401,
      "theoretical_loss": 4.449910056790086,
      "tokens_seen": 168820736
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004792076228686058,
      "loss": 3.122,
      "theoretical_loss": 4.44968826280015,
      "tokens_seen": 168886272
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047919759277833504,
      "loss": 3.4615,
      "theoretical_loss": 4.4494665789482895,
      "tokens_seen": 168951808
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047918756268806417,
      "loss": 3.2752,
      "theoretical_loss": 4.449245005137125,
      "tokens_seen": 169017344
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004791775325977934,
      "loss": 3.2905,
      "theoretical_loss": 4.449023541269395,
      "tokens_seen": 169082880
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047916750250752253,
      "loss": 3.3756,
      "theoretical_loss": 4.448802187247966,
      "tokens_seen": 169148416
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047915747241725177,
      "loss": 3.5968,
      "theoretical_loss": 4.448580942975825,
      "tokens_seen": 169213952
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047914744232698095,
      "loss": 3.2809,
      "theoretical_loss": 4.448359808356084,
      "tokens_seen": 169279488
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047913741223671013,
      "loss": 3.5359,
      "theoretical_loss": 4.448138783291979,
      "tokens_seen": 169345024
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004791273821464393,
      "loss": 3.4119,
      "theoretical_loss": 4.447917867686863,
      "tokens_seen": 169410560
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047911735205616855,
      "loss": 3.4354,
      "theoretical_loss": 4.44769706144422,
      "tokens_seen": 169476096
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004791073219658977,
      "loss": 3.4738,
      "theoretical_loss": 4.44747636446765,
      "tokens_seen": 169541632
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004790972918756269,
      "loss": 3.6068,
      "theoretical_loss": 4.447255776660878,
      "tokens_seen": 169607168
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047908726178535604,
      "loss": 3.2765,
      "theoretical_loss": 4.44703529792775,
      "tokens_seen": 169672704
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047907723169508527,
      "loss": 3.2577,
      "theoretical_loss": 4.446814928172234,
      "tokens_seen": 169738240
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047906720160481445,
      "loss": 3.3136,
      "theoretical_loss": 4.446594667298421,
      "tokens_seen": 169803776
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047905717151454363,
      "loss": 3.4342,
      "theoretical_loss": 4.446374515210521,
      "tokens_seen": 169869312
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004790471414242728,
      "loss": 3.305,
      "theoretical_loss": 4.446154471812866,
      "tokens_seen": 169934848
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000479037111334002,
      "loss": 3.4233,
      "theoretical_loss": 4.445934537009911,
      "tokens_seen": 170000384
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004790270812437312,
      "loss": 3.5624,
      "theoretical_loss": 4.445714710706228,
      "tokens_seen": 170065920
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004790170511534604,
      "loss": 3.4564,
      "theoretical_loss": 4.445494992806513,
      "tokens_seen": 170131456
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047900702106318954,
      "loss": 3.4515,
      "theoretical_loss": 4.44527538321558,
      "tokens_seen": 170196992
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004789969909729188,
      "loss": 3.3197,
      "theoretical_loss": 4.445055881838365,
      "tokens_seen": 170262528
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004789869608826479,
      "loss": 3.4949,
      "theoretical_loss": 4.444836488579924,
      "tokens_seen": 170328064
    },
    {
      "epoch": 0.05,
      "objective/train/docs_used": 234408,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.2099924087524414,
      "objective/train/theoretical_loss": 4.44461720334543,
      "objective/train/tokens_used": 190853600,
      "theoretical_loss": 4.44461720334543,
      "tokens_seen": 170393600
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047897693079237714,
      "loss": 3.2463,
      "theoretical_loss": 4.44461720334543,
      "tokens_seen": 170393600
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047896690070210637,
      "loss": 3.4087,
      "theoretical_loss": 4.444398026040179,
      "tokens_seen": 170459136
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004789568706118355,
      "loss": 3.4015,
      "theoretical_loss": 4.444178956569585,
      "tokens_seen": 170524672
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047894684052156473,
      "loss": 3.2236,
      "theoretical_loss": 4.443959994839181,
      "tokens_seen": 170590208
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004789368104312939,
      "loss": 3.3624,
      "theoretical_loss": 4.44374114075462,
      "tokens_seen": 170655744
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004789267803410231,
      "loss": 3.4993,
      "theoretical_loss": 4.443522394221671,
      "tokens_seen": 170721280
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004789167502507523,
      "loss": 3.3937,
      "theoretical_loss": 4.443303755146225,
      "tokens_seen": 170786816
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047890672016048146,
      "loss": 3.2099,
      "theoretical_loss": 4.443085223434291,
      "tokens_seen": 170852352
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047889669007021064,
      "loss": 3.2229,
      "theoretical_loss": 4.442866798991993,
      "tokens_seen": 170917888
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004788866599799399,
      "loss": 3.38,
      "theoretical_loss": 4.442648481725577,
      "tokens_seen": 170983424
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000478876629889669,
      "loss": 3.5082,
      "theoretical_loss": 4.442430271541404,
      "tokens_seen": 171048960
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047886659979939824,
      "loss": 3.2701,
      "theoretical_loss": 4.442212168345956,
      "tokens_seen": 171114496
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047885656970912736,
      "loss": 3.405,
      "theoretical_loss": 4.4419941720458285,
      "tokens_seen": 171180032
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004788465396188566,
      "loss": 3.0497,
      "theoretical_loss": 4.441776282547736,
      "tokens_seen": 171245568
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004788365095285858,
      "loss": 3.2281,
      "theoretical_loss": 4.441558499758511,
      "tokens_seen": 171311104
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047882647943831496,
      "loss": 3.3353,
      "theoretical_loss": 4.441340823585101,
      "tokens_seen": 171376640
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047881644934804414,
      "loss": 3.2502,
      "theoretical_loss": 4.441123253934572,
      "tokens_seen": 171442176
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004788064192577733,
      "loss": 3.36,
      "theoretical_loss": 4.440905790714105,
      "tokens_seen": 171507712
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004787963891675025,
      "loss": 3.2246,
      "theoretical_loss": 4.440688433830999,
      "tokens_seen": 171573248
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047878635907723174,
      "loss": 3.1875,
      "theoretical_loss": 4.440471183192667,
      "tokens_seen": 171638784
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047877632898696087,
      "loss": 3.2297,
      "theoretical_loss": 4.440254038706639,
      "tokens_seen": 171704320
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004787662988966901,
      "loss": 3.4318,
      "theoretical_loss": 4.440037000280561,
      "tokens_seen": 171769856
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004787562688064193,
      "loss": 3.298,
      "theoretical_loss": 4.439820067822195,
      "tokens_seen": 171835392
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047874623871614847,
      "loss": 3.5067,
      "theoretical_loss": 4.439603241239416,
      "tokens_seen": 171900928
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047873620862587765,
      "loss": 3.1999,
      "theoretical_loss": 4.439386520440218,
      "tokens_seen": 171966464
    },
    {
      "epoch": 0.05,
      "objective/train/docs_used": 235647,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.488489866256714,
      "objective/train/theoretical_loss": 4.439169905332706,
      "objective/train/tokens_used": 192492000,
      "theoretical_loss": 4.439169905332706,
      "tokens_seen": 172032000
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047872617853560683,
      "loss": 3.6249,
      "theoretical_loss": 4.439169905332706,
      "tokens_seen": 172032000
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000478716148445336,
      "loss": 3.4185,
      "theoretical_loss": 4.438953395825102,
      "tokens_seen": 172097536
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047870611835506524,
      "loss": 3.4638,
      "theoretical_loss": 4.438736991825744,
      "tokens_seen": 172163072
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047869608826479437,
      "loss": 2.9901,
      "theoretical_loss": 4.438520693243079,
      "tokens_seen": 172228608
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004786860581745236,
      "loss": 3.3508,
      "theoretical_loss": 4.4383044999856756,
      "tokens_seen": 172294144
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047867602808425273,
      "loss": 3.006,
      "theoretical_loss": 4.438088411962211,
      "tokens_seen": 172359680
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047866599799398197,
      "loss": 3.2261,
      "theoretical_loss": 4.437872429081477,
      "tokens_seen": 172425216
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047865596790371115,
      "loss": 3.3128,
      "theoretical_loss": 4.437656551252381,
      "tokens_seen": 172490752
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047864593781344033,
      "loss": 3.1811,
      "theoretical_loss": 4.4374407783839445,
      "tokens_seen": 172556288
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004786359077231695,
      "loss": 3.6537,
      "theoretical_loss": 4.437225110385297,
      "tokens_seen": 172621824
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047862587763289875,
      "loss": 3.3197,
      "theoretical_loss": 4.4370095471656885,
      "tokens_seen": 172687360
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004786158475426279,
      "loss": 3.2659,
      "theoretical_loss": 4.436794088634477,
      "tokens_seen": 172752896
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004786058174523571,
      "loss": 3.432,
      "theoretical_loss": 4.4365787347011345,
      "tokens_seen": 172818432
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047859578736208624,
      "loss": 3.4144,
      "theoretical_loss": 4.436363485275246,
      "tokens_seen": 172883968
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047858575727181547,
      "loss": 3.4017,
      "theoretical_loss": 4.436148340266508,
      "tokens_seen": 172949504
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047857572718154465,
      "loss": 3.7135,
      "theoretical_loss": 4.435933299584729,
      "tokens_seen": 173015040
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047856569709127383,
      "loss": 3.2931,
      "theoretical_loss": 4.4357183631398325,
      "tokens_seen": 173080576
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000478555667001003,
      "loss": 3.3342,
      "theoretical_loss": 4.435503530841849,
      "tokens_seen": 173146112
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004785456369107322,
      "loss": 3.1036,
      "theoretical_loss": 4.435288802600926,
      "tokens_seen": 173211648
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004785356068204614,
      "loss": 3.4167,
      "theoretical_loss": 4.4350741783273175,
      "tokens_seen": 173277184
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004785255767301906,
      "loss": 3.3526,
      "theoretical_loss": 4.434859657931392,
      "tokens_seen": 173342720
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047851554663991974,
      "loss": 3.3756,
      "theoretical_loss": 4.434645241323629,
      "tokens_seen": 173408256
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000478505516549649,
      "loss": 3.3329,
      "theoretical_loss": 4.434430928414617,
      "tokens_seen": 173473792
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004784954864593781,
      "loss": 3.2552,
      "theoretical_loss": 4.434216719115057,
      "tokens_seen": 173539328
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047848545636910734,
      "loss": 3.1126,
      "theoretical_loss": 4.43400261333576,
      "tokens_seen": 173604864
    },
    {
      "epoch": 0.05,
      "objective/train/docs_used": 236140,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.18463397026062,
      "objective/train/theoretical_loss": 4.433788610987646,
      "objective/train/tokens_used": 194130400,
      "theoretical_loss": 4.433788610987646,
      "tokens_seen": 173670400
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004784754262788365,
      "loss": 3.3947,
      "theoretical_loss": 4.433788610987646,
      "tokens_seen": 173670400
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004784653961885657,
      "loss": 3.4461,
      "theoretical_loss": 4.433574711981749,
      "tokens_seen": 173735936
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004784553660982949,
      "loss": 3.2924,
      "theoretical_loss": 4.433360916229209,
      "tokens_seen": 173801472
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004784453360080241,
      "loss": 3.3511,
      "theoretical_loss": 4.433147223641278,
      "tokens_seen": 173867008
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047843530591775324,
      "loss": 3.1406,
      "theoretical_loss": 4.432933634129318,
      "tokens_seen": 173932544
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004784252758274825,
      "loss": 3.3653,
      "theoretical_loss": 4.4327201476047975,
      "tokens_seen": 173998080
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004784152457372116,
      "loss": 3.1941,
      "theoretical_loss": 4.432506763979299,
      "tokens_seen": 174063616
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047840521564694084,
      "loss": 3.2088,
      "theoretical_loss": 4.432293483164512,
      "tokens_seen": 174129152
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047839518555667,
      "loss": 3.4825,
      "theoretical_loss": 4.432080305072233,
      "tokens_seen": 174194688
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004783851554663992,
      "loss": 3.1969,
      "theoretical_loss": 4.43186722961437,
      "tokens_seen": 174260224
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004783751253761284,
      "loss": 3.3585,
      "theoretical_loss": 4.431654256702938,
      "tokens_seen": 174325760
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047836509528585756,
      "loss": 3.2937,
      "theoretical_loss": 4.431441386250063,
      "tokens_seen": 174391296
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047835506519558675,
      "loss": 3.3733,
      "theoretical_loss": 4.4312286181679745,
      "tokens_seen": 174456832
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000478345035105316,
      "loss": 3.1708,
      "theoretical_loss": 4.431015952369016,
      "tokens_seen": 174522368
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004783350050150451,
      "loss": 3.4721,
      "theoretical_loss": 4.430803388765636,
      "tokens_seen": 174587904
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047832497492477434,
      "loss": 3.3855,
      "theoretical_loss": 4.430590927270388,
      "tokens_seen": 174653440
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047831494483450347,
      "loss": 3.2205,
      "theoretical_loss": 4.430378567795938,
      "tokens_seen": 174718976
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004783049147442327,
      "loss": 3.3129,
      "theoretical_loss": 4.430166310255057,
      "tokens_seen": 174784512
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004782948846539619,
      "loss": 3.6459,
      "theoretical_loss": 4.429954154560624,
      "tokens_seen": 174850048
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047828485456369107,
      "loss": 3.3131,
      "theoretical_loss": 4.429742100625624,
      "tokens_seen": 174915584
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047827482447342025,
      "loss": 3.2811,
      "theoretical_loss": 4.429530148363151,
      "tokens_seen": 174981120
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004782647943831495,
      "loss": 3.2996,
      "theoretical_loss": 4.429318297686402,
      "tokens_seen": 175046656
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004782547642928786,
      "loss": 3.3326,
      "theoretical_loss": 4.429106548508685,
      "tokens_seen": 175112192
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047824473420260785,
      "loss": 3.1878,
      "theoretical_loss": 4.428894900743411,
      "tokens_seen": 175177728
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000478234704112337,
      "loss": 3.214,
      "theoretical_loss": 4.428683354304098,
      "tokens_seen": 175243264
    },
    {
      "epoch": 0.05,
      "objective/train/docs_used": 237488,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5905725955963135,
      "objective/train/theoretical_loss": 4.428471909104372,
      "objective/train/tokens_used": 195768800,
      "theoretical_loss": 4.428471909104372,
      "tokens_seen": 175308800
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004782246740220662,
      "loss": 3.1411,
      "theoretical_loss": 4.428471909104372,
      "tokens_seen": 175308800
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047821464393179544,
      "loss": 3.3937,
      "theoretical_loss": 4.428260565057964,
      "tokens_seen": 175374336
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047820461384152457,
      "loss": 3.3474,
      "theoretical_loss": 4.428049322078708,
      "tokens_seen": 175439872
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004781945837512538,
      "loss": 3.3299,
      "theoretical_loss": 4.427838180080547,
      "tokens_seen": 175505408
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047818455366098293,
      "loss": 3.3159,
      "theoretical_loss": 4.4276271389775275,
      "tokens_seen": 175570944
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047817452357071217,
      "loss": 3.3342,
      "theoretical_loss": 4.427416198683803,
      "tokens_seen": 175636480
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047816449348044135,
      "loss": 3.2288,
      "theoretical_loss": 4.427205359113629,
      "tokens_seen": 175702016
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047815446339017053,
      "loss": 3.5414,
      "theoretical_loss": 4.42699462018137,
      "tokens_seen": 175767552
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004781444332998997,
      "loss": 3.389,
      "theoretical_loss": 4.42678398180149,
      "tokens_seen": 175833088
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047813440320962895,
      "loss": 3.2056,
      "theoretical_loss": 4.426573443888563,
      "tokens_seen": 175898624
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004781243731193581,
      "loss": 3.2972,
      "theoretical_loss": 4.426363006357263,
      "tokens_seen": 175964160
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004781143430290873,
      "loss": 3.4419,
      "theoretical_loss": 4.426152669122374,
      "tokens_seen": 176029696
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047810431293881644,
      "loss": 3.285,
      "theoretical_loss": 4.425942432098774,
      "tokens_seen": 176095232
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047809428284854567,
      "loss": 3.3969,
      "theoretical_loss": 4.425732295201455,
      "tokens_seen": 176160768
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047808425275827485,
      "loss": 3.2307,
      "theoretical_loss": 4.425522258345508,
      "tokens_seen": 176226304
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047807422266800403,
      "loss": 3.4295,
      "theoretical_loss": 4.425312321446127,
      "tokens_seen": 176291840
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004780641925777332,
      "loss": 3.1458,
      "theoretical_loss": 4.425102484418613,
      "tokens_seen": 176357376
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004780541624874624,
      "loss": 3.3972,
      "theoretical_loss": 4.424892747178365,
      "tokens_seen": 176422912
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004780441323971916,
      "loss": 3.527,
      "theoretical_loss": 4.42468310964089,
      "tokens_seen": 176488448
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004780341023069208,
      "loss": 3.4912,
      "theoretical_loss": 4.424473571721794,
      "tokens_seen": 176553984
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047802407221664994,
      "loss": 3.2588,
      "theoretical_loss": 4.42426413333679,
      "tokens_seen": 176619520
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004780140421263792,
      "loss": 3.398,
      "theoretical_loss": 4.424054794401689,
      "tokens_seen": 176685056
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004780040120361083,
      "loss": 3.3066,
      "theoretical_loss": 4.423845554832406,
      "tokens_seen": 176750592
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047799398194583754,
      "loss": 3.4778,
      "theoretical_loss": 4.42363641454496,
      "tokens_seen": 176816128
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004779839518555667,
      "loss": 3.4736,
      "theoretical_loss": 4.423427373455471,
      "tokens_seen": 176881664
    },
    {
      "epoch": 0.05,
      "objective/train/docs_used": 238146,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.402047634124756,
      "objective/train/theoretical_loss": 4.42321843148016,
      "objective/train/tokens_used": 197407200,
      "theoretical_loss": 4.42321843148016,
      "tokens_seen": 176947200
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004779739217652959,
      "loss": 3.4011,
      "theoretical_loss": 4.42321843148016,
      "tokens_seen": 176947200
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004779638916750251,
      "loss": 3.2247,
      "theoretical_loss": 4.423009588535351,
      "tokens_seen": 177012736
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004779538615847543,
      "loss": 3.5019,
      "theoretical_loss": 4.422800844537466,
      "tokens_seen": 177078272
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047794383149448344,
      "loss": 3.3271,
      "theoretical_loss": 4.422592199403036,
      "tokens_seen": 177143808
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004779338014042127,
      "loss": 3.1612,
      "theoretical_loss": 4.422383653048685,
      "tokens_seen": 177209344
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004779237713139418,
      "loss": 3.5149,
      "theoretical_loss": 4.422175205391145,
      "tokens_seen": 177274880
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047791374122367104,
      "loss": 3.4788,
      "theoretical_loss": 4.421966856347243,
      "tokens_seen": 177340416
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004779037111334002,
      "loss": 3.3556,
      "theoretical_loss": 4.421758605833912,
      "tokens_seen": 177405952
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004778936810431294,
      "loss": 3.2858,
      "theoretical_loss": 4.421550453768181,
      "tokens_seen": 177471488
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004778836509528586,
      "loss": 3.2064,
      "theoretical_loss": 4.421342400067183,
      "tokens_seen": 177537024
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047787362086258776,
      "loss": 3.2573,
      "theoretical_loss": 4.42113444464815,
      "tokens_seen": 177602560
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047786359077231695,
      "loss": 3.1305,
      "theoretical_loss": 4.420926587428411,
      "tokens_seen": 177668096
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004778535606820462,
      "loss": 3.3526,
      "theoretical_loss": 4.420718828325403,
      "tokens_seen": 177733632
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004778435305917753,
      "loss": 3.4669,
      "theoretical_loss": 4.420511167256656,
      "tokens_seen": 177799168
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047783350050150454,
      "loss": 3.2721,
      "theoretical_loss": 4.4203036041398,
      "tokens_seen": 177864704
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047782347041123367,
      "loss": 3.4746,
      "theoretical_loss": 4.420096138892568,
      "tokens_seen": 177930240
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004778134403209629,
      "loss": 3.2158,
      "theoretical_loss": 4.419888771432789,
      "tokens_seen": 177995776
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004778034102306921,
      "loss": 3.5373,
      "theoretical_loss": 4.419681501678395,
      "tokens_seen": 178061312
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047779338014042127,
      "loss": 3.2966,
      "theoretical_loss": 4.419474329547413,
      "tokens_seen": 178126848
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047778335005015045,
      "loss": 3.4379,
      "theoretical_loss": 4.419267254957971,
      "tokens_seen": 178192384
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004777733199598797,
      "loss": 3.5645,
      "theoretical_loss": 4.419060277828295,
      "tokens_seen": 178257920
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004777632898696088,
      "loss": 3.3582,
      "theoretical_loss": 4.41885339807671,
      "tokens_seen": 178323456
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047775325977933805,
      "loss": 3.2314,
      "theoretical_loss": 4.4186466156216415,
      "tokens_seen": 178388992
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004777432296890672,
      "loss": 3.3675,
      "theoretical_loss": 4.418439930381609,
      "tokens_seen": 178454528
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004777331995987964,
      "loss": 3.1722,
      "theoretical_loss": 4.418233342275233,
      "tokens_seen": 178520064
    },
    {
      "epoch": 0.05,
      "objective/train/docs_used": 239453,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.536768674850464,
      "objective/train/theoretical_loss": 4.418026851221231,
      "objective/train/tokens_used": 199045600,
      "theoretical_loss": 4.418026851221231,
      "tokens_seen": 178585600
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004777231695085256,
      "loss": 3.5599,
      "theoretical_loss": 4.418026851221231,
      "tokens_seen": 178585600
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047771313941825477,
      "loss": 3.3849,
      "theoretical_loss": 4.4178204571384185,
      "tokens_seen": 178651136
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047770310932798395,
      "loss": 3.2237,
      "theoretical_loss": 4.41761415994571,
      "tokens_seen": 178716672
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047769307923771313,
      "loss": 3.3179,
      "theoretical_loss": 4.417407959562116,
      "tokens_seen": 178782208
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004776830491474423,
      "loss": 3.0157,
      "theoretical_loss": 4.417201855906742,
      "tokens_seen": 178847744
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047767301905717155,
      "loss": 3.4205,
      "theoretical_loss": 4.416995848898797,
      "tokens_seen": 178913280
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004776629889669007,
      "loss": 3.5319,
      "theoretical_loss": 4.4167899384575815,
      "tokens_seen": 178978816
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004776529588766299,
      "loss": 3.3096,
      "theoretical_loss": 4.416584124502495,
      "tokens_seen": 179044352
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047764292878635904,
      "loss": 3.5886,
      "theoretical_loss": 4.416378406953033,
      "tokens_seen": 179109888
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004776328986960883,
      "loss": 3.4965,
      "theoretical_loss": 4.41617278572879,
      "tokens_seen": 179175424
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047762286860581746,
      "loss": 3.2969,
      "theoretical_loss": 4.4159672607494524,
      "tokens_seen": 179240960
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047761283851554664,
      "loss": 3.3703,
      "theoretical_loss": 4.415761831934808,
      "tokens_seen": 179306496
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004776028084252758,
      "loss": 3.363,
      "theoretical_loss": 4.415556499204737,
      "tokens_seen": 179372032
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047759277833500505,
      "loss": 3.4131,
      "theoretical_loss": 4.415351262479216,
      "tokens_seen": 179437568
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004775827482447342,
      "loss": 3.2837,
      "theoretical_loss": 4.415146121678321,
      "tokens_seen": 179503104
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004775727181544634,
      "loss": 3.3687,
      "theoretical_loss": 4.414941076722219,
      "tokens_seen": 179568640
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047756268806419254,
      "loss": 3.2755,
      "theoretical_loss": 4.4147361275311745,
      "tokens_seen": 179634176
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004775526579739218,
      "loss": 3.4791,
      "theoretical_loss": 4.414531274025548,
      "tokens_seen": 179699712
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047754262788365096,
      "loss": 3.3726,
      "theoretical_loss": 4.414326516125795,
      "tokens_seen": 179765248
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047753259779338014,
      "loss": 3.4113,
      "theoretical_loss": 4.414121853752466,
      "tokens_seen": 179830784
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004775225677031093,
      "loss": 3.3512,
      "theoretical_loss": 4.413917286826205,
      "tokens_seen": 179896320
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004775125376128385,
      "loss": 3.4069,
      "theoretical_loss": 4.413712815267752,
      "tokens_seen": 179961856
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004775025075225677,
      "loss": 3.1901,
      "theoretical_loss": 4.413508438997944,
      "tokens_seen": 180027392
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004774924774322969,
      "loss": 3.5482,
      "theoretical_loss": 4.4133041579377075,
      "tokens_seen": 180092928
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047748244734202605,
      "loss": 3.3269,
      "theoretical_loss": 4.413099972008068,
      "tokens_seen": 180158464
    },
    {
      "epoch": 0.05,
      "objective/train/docs_used": 240101,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.189091444015503,
      "objective/train/theoretical_loss": 4.412895881130142,
      "objective/train/tokens_used": 200684000,
      "theoretical_loss": 4.412895881130142,
      "tokens_seen": 180224000
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004774724172517553,
      "loss": 3.2624,
      "theoretical_loss": 4.412895881130142,
      "tokens_seen": 180224000
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004774623871614845,
      "loss": 3.4578,
      "theoretical_loss": 4.412691885225141,
      "tokens_seen": 180289536
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047745235707121364,
      "loss": 3.3567,
      "theoretical_loss": 4.412487984214373,
      "tokens_seen": 180355072
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004774423269809429,
      "loss": 3.2258,
      "theoretical_loss": 4.412284178019235,
      "tokens_seen": 180420608
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000477432296890672,
      "loss": 3.2944,
      "theoretical_loss": 4.412080466561221,
      "tokens_seen": 180486144
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047742226680040124,
      "loss": 3.3614,
      "theoretical_loss": 4.411876849761917,
      "tokens_seen": 180551680
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004774122367101304,
      "loss": 3.1865,
      "theoretical_loss": 4.411673327543005,
      "tokens_seen": 180617216
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004774022066198596,
      "loss": 3.517,
      "theoretical_loss": 4.4114698998262565,
      "tokens_seen": 180682752
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004773921765295888,
      "loss": 3.5934,
      "theoretical_loss": 4.411266566533539,
      "tokens_seen": 180748288
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047738214643931797,
      "loss": 3.0994,
      "theoretical_loss": 4.41106332758681,
      "tokens_seen": 180813824
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047737211634904715,
      "loss": 3.6293,
      "theoretical_loss": 4.41086018290812,
      "tokens_seen": 180879360
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004773620862587764,
      "loss": 3.3159,
      "theoretical_loss": 4.410657132419617,
      "tokens_seen": 180944896
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004773520561685055,
      "loss": 3.2151,
      "theoretical_loss": 4.410454176043537,
      "tokens_seen": 181010432
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047734202607823474,
      "loss": 3.3866,
      "theoretical_loss": 4.410251313702208,
      "tokens_seen": 181075968
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047733199598796387,
      "loss": 3.3264,
      "theoretical_loss": 4.410048545318052,
      "tokens_seen": 181141504
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004773219658976931,
      "loss": 3.3692,
      "theoretical_loss": 4.409845870813582,
      "tokens_seen": 181207040
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004773119358074223,
      "loss": 3.4274,
      "theoretical_loss": 4.409643290111404,
      "tokens_seen": 181272576
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047730190571715147,
      "loss": 3.2483,
      "theoretical_loss": 4.409440803134215,
      "tokens_seen": 181338112
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047729187562688065,
      "loss": 3.2365,
      "theoretical_loss": 4.409238409804804,
      "tokens_seen": 181403648
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004772818455366099,
      "loss": 3.2074,
      "theoretical_loss": 4.409036110046051,
      "tokens_seen": 181469184
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000477271815446339,
      "loss": 3.0851,
      "theoretical_loss": 4.408833903780926,
      "tokens_seen": 181534720
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047726178535606825,
      "loss": 3.2768,
      "theoretical_loss": 4.408631790932494,
      "tokens_seen": 181600256
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004772517552657974,
      "loss": 3.7316,
      "theoretical_loss": 4.408429771423909,
      "tokens_seen": 181665792
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004772417251755266,
      "loss": 3.1774,
      "theoretical_loss": 4.408227845178414,
      "tokens_seen": 181731328
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004772316950852558,
      "loss": 3.3076,
      "theoretical_loss": 4.408026012119344,
      "tokens_seen": 181796864
    },
    {
      "epoch": 0.06,
      "objective/train/docs_used": 241412,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.7472660541534424,
      "objective/train/theoretical_loss": 4.407824272170128,
      "objective/train/tokens_used": 202322400,
      "theoretical_loss": 4.407824272170128,
      "tokens_seen": 181862400
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047722166499498497,
      "loss": 3.2786,
      "theoretical_loss": 4.407824272170128,
      "tokens_seen": 181862400
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047721163490471415,
      "loss": 3.4129,
      "theoretical_loss": 4.407622625254279,
      "tokens_seen": 181927936
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047720160481444333,
      "loss": 2.838,
      "theoretical_loss": 4.407421071295406,
      "tokens_seen": 181993472
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004771915747241725,
      "loss": 3.2404,
      "theoretical_loss": 4.407219610217206,
      "tokens_seen": 182059008
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047718154463390175,
      "loss": 3.4927,
      "theoretical_loss": 4.407018241943467,
      "tokens_seen": 182124544
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004771715145436309,
      "loss": 3.1684,
      "theoretical_loss": 4.406816966398064,
      "tokens_seen": 182190080
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004771614844533601,
      "loss": 3.3542,
      "theoretical_loss": 4.406615783504965,
      "tokens_seen": 182255616
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047715145436308924,
      "loss": 3.2972,
      "theoretical_loss": 4.4064146931882275,
      "tokens_seen": 182321152
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004771414242728185,
      "loss": 3.0567,
      "theoretical_loss": 4.406213695371996,
      "tokens_seen": 182386688
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047713139418254766,
      "loss": 3.2913,
      "theoretical_loss": 4.406012789980506,
      "tokens_seen": 182452224
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047712136409227684,
      "loss": 3.425,
      "theoretical_loss": 4.405811976938084,
      "tokens_seen": 182517760
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000477111334002006,
      "loss": 3.3947,
      "theoretical_loss": 4.405611256169143,
      "tokens_seen": 182583296
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047710130391173525,
      "loss": 3.2788,
      "theoretical_loss": 4.405410627598185,
      "tokens_seen": 182648832
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004770912738214644,
      "loss": 3.2396,
      "theoretical_loss": 4.405210091149802,
      "tokens_seen": 182714368
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004770812437311936,
      "loss": 3.6049,
      "theoretical_loss": 4.405009646748674,
      "tokens_seen": 182779904
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047707121364092274,
      "loss": 3.2774,
      "theoretical_loss": 4.404809294319572,
      "tokens_seen": 182845440
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000477061183550652,
      "loss": 3.1352,
      "theoretical_loss": 4.40460903378735,
      "tokens_seen": 182910976
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047705115346038116,
      "loss": 3.5256,
      "theoretical_loss": 4.404408865076955,
      "tokens_seen": 182976512
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047704112337011034,
      "loss": 3.3179,
      "theoretical_loss": 4.404208788113422,
      "tokens_seen": 183042048
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004770310932798395,
      "loss": 3.368,
      "theoretical_loss": 4.404008802821871,
      "tokens_seen": 183107584
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004770210631895687,
      "loss": 3.4253,
      "theoretical_loss": 4.4038089091275125,
      "tokens_seen": 183173120
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004770110330992979,
      "loss": 3.3234,
      "theoretical_loss": 4.403609106955645,
      "tokens_seen": 183238656
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004770010030090271,
      "loss": 3.3547,
      "theoretical_loss": 4.403409396231651,
      "tokens_seen": 183304192
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047699097291875625,
      "loss": 3.3308,
      "theoretical_loss": 4.403209776881004,
      "tokens_seen": 183369728
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004769809428284855,
      "loss": 3.3826,
      "theoretical_loss": 4.403010248829265,
      "tokens_seen": 183435264
    },
    {
      "epoch": 0.06,
      "objective/train/docs_used": 241901,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.666914463043213,
      "objective/train/theoretical_loss": 4.4028108120020795,
      "objective/train/tokens_used": 203960800,
      "theoretical_loss": 4.4028108120020795,
      "tokens_seen": 183500800
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047697091273821466,
      "loss": 3.347,
      "theoretical_loss": 4.4028108120020795,
      "tokens_seen": 183500800
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047696088264794384,
      "loss": 3.3715,
      "theoretical_loss": 4.402611466325182,
      "tokens_seen": 183566336
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000476950852557673,
      "loss": 3.5634,
      "theoretical_loss": 4.4024122117243945,
      "tokens_seen": 183631872
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004769408224674022,
      "loss": 3.0663,
      "theoretical_loss": 4.402213048125624,
      "tokens_seen": 183697408
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004769307923771314,
      "loss": 3.059,
      "theoretical_loss": 4.4020139754548655,
      "tokens_seen": 183762944
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004769207622868606,
      "loss": 3.1995,
      "theoretical_loss": 4.401814993638199,
      "tokens_seen": 183828480
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047691073219658975,
      "loss": 3.1642,
      "theoretical_loss": 4.4016161026017935,
      "tokens_seen": 183894016
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000476900702106319,
      "loss": 2.9642,
      "theoretical_loss": 4.401417302271902,
      "tokens_seen": 183959552
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004768906720160481,
      "loss": 3.405,
      "theoretical_loss": 4.401218592574865,
      "tokens_seen": 184025088
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047688064192577735,
      "loss": 3.362,
      "theoretical_loss": 4.401019973437108,
      "tokens_seen": 184090624
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047687061183550653,
      "loss": 3.416,
      "theoretical_loss": 4.400821444785143,
      "tokens_seen": 184156160
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004768605817452357,
      "loss": 3.1617,
      "theoretical_loss": 4.400623006545567,
      "tokens_seen": 184221696
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004768505516549649,
      "loss": 3.4902,
      "theoretical_loss": 4.400424658645065,
      "tokens_seen": 184287232
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047684052156469407,
      "loss": 3.2861,
      "theoretical_loss": 4.400226401010404,
      "tokens_seen": 184352768
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047683049147442325,
      "loss": 3.3124,
      "theoretical_loss": 4.40002823356844,
      "tokens_seen": 184418304
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004768204613841525,
      "loss": 2.9837,
      "theoretical_loss": 4.39983015624611,
      "tokens_seen": 184483840
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004768104312938816,
      "loss": 3.2424,
      "theoretical_loss": 4.39963216897044,
      "tokens_seen": 184549376
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047680040120361085,
      "loss": 3.2719,
      "theoretical_loss": 4.3994342716685395,
      "tokens_seen": 184614912
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047679037111334003,
      "loss": 3.1485,
      "theoretical_loss": 4.399236464267602,
      "tokens_seen": 184680448
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004767803410230692,
      "loss": 3.3108,
      "theoretical_loss": 4.399038746694908,
      "tokens_seen": 184745984
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004767703109327984,
      "loss": 3.135,
      "theoretical_loss": 4.398841118877819,
      "tokens_seen": 184811520
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004767602808425276,
      "loss": 3.4858,
      "theoretical_loss": 4.398643580743785,
      "tokens_seen": 184877056
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047675025075225676,
      "loss": 3.272,
      "theoretical_loss": 4.398446132220338,
      "tokens_seen": 184942592
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000476740220661986,
      "loss": 3.4381,
      "theoretical_loss": 4.3982487732350934,
      "tokens_seen": 185008128
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004767301905717151,
      "loss": 3.3967,
      "theoretical_loss": 4.398051503715753,
      "tokens_seen": 185073664
    },
    {
      "epoch": 0.06,
      "objective/train/docs_used": 243537,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.3711562156677246,
      "objective/train/theoretical_loss": 4.397854323590102,
      "objective/train/tokens_used": 205599200,
      "theoretical_loss": 4.397854323590102,
      "tokens_seen": 185139200
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047672016048144435,
      "loss": 3.0928,
      "theoretical_loss": 4.397854323590102,
      "tokens_seen": 185139200
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047671013039117353,
      "loss": 3.2136,
      "theoretical_loss": 4.397657232786008,
      "tokens_seen": 185204736
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004767001003009027,
      "loss": 3.3584,
      "theoretical_loss": 4.397460231231424,
      "tokens_seen": 185270272
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047669007021063195,
      "loss": 3.4504,
      "theoretical_loss": 4.397263318854384,
      "tokens_seen": 185335808
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004766800401203611,
      "loss": 3.3679,
      "theoretical_loss": 4.39706649558301,
      "tokens_seen": 185401344
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004766700100300903,
      "loss": 3.254,
      "theoretical_loss": 4.396869761345503,
      "tokens_seen": 185466880
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047665997993981944,
      "loss": 3.0167,
      "theoretical_loss": 4.396673116070147,
      "tokens_seen": 185532416
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004766499498495487,
      "loss": 3.397,
      "theoretical_loss": 4.396476559685315,
      "tokens_seen": 185597952
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047663991975927786,
      "loss": 3.1028,
      "theoretical_loss": 4.396280092119455,
      "tokens_seen": 185663488
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047662988966900704,
      "loss": 3.3503,
      "theoretical_loss": 4.3960837133011035,
      "tokens_seen": 185729024
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004766198595787362,
      "loss": 3.2027,
      "theoretical_loss": 4.395887423158877,
      "tokens_seen": 185794560
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047660982948846545,
      "loss": 3.36,
      "theoretical_loss": 4.395691221621476,
      "tokens_seen": 185860096
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004765997993981946,
      "loss": 3.4191,
      "theoretical_loss": 4.395495108617682,
      "tokens_seen": 185925632
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004765897693079238,
      "loss": 3.4474,
      "theoretical_loss": 4.39529908407636,
      "tokens_seen": 185991168
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047657973921765294,
      "loss": 3.2397,
      "theoretical_loss": 4.3951031479264575,
      "tokens_seen": 186056704
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004765697091273822,
      "loss": 3.4729,
      "theoretical_loss": 4.394907300097002,
      "tokens_seen": 186122240
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047655967903711136,
      "loss": 3.2495,
      "theoretical_loss": 4.394711540517106,
      "tokens_seen": 186187776
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047654964894684054,
      "loss": 3.4181,
      "theoretical_loss": 4.39451586911596,
      "tokens_seen": 186253312
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004765396188565697,
      "loss": 3.5468,
      "theoretical_loss": 4.39432028582284,
      "tokens_seen": 186318848
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004765295887662989,
      "loss": 3.111,
      "theoretical_loss": 4.394124790567101,
      "tokens_seen": 186384384
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004765195586760281,
      "loss": 3.3198,
      "theoretical_loss": 4.3939293832781825,
      "tokens_seen": 186449920
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004765095285857573,
      "loss": 3.2602,
      "theoretical_loss": 4.393734063885599,
      "tokens_seen": 186515456
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047649949849548645,
      "loss": 3.0038,
      "theoretical_loss": 4.3935388323189555,
      "tokens_seen": 186580992
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004764894684052157,
      "loss": 3.371,
      "theoretical_loss": 4.39334368850793,
      "tokens_seen": 186646528
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047647943831494486,
      "loss": 3.4606,
      "theoretical_loss": 4.3931486323822835,
      "tokens_seen": 186712064
    },
    {
      "epoch": 0.06,
      "objective/train/docs_used": 246703,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.3764562606811523,
      "objective/train/theoretical_loss": 4.392953663871862,
      "objective/train/tokens_used": 207237600,
      "theoretical_loss": 4.392953663871862,
      "tokens_seen": 186777600
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047646940822467404,
      "loss": 3.2311,
      "theoretical_loss": 4.392953663871862,
      "tokens_seen": 186777600
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004764593781344032,
      "loss": 3.2135,
      "theoretical_loss": 4.392758782906586,
      "tokens_seen": 186843136
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004764493480441324,
      "loss": 3.5582,
      "theoretical_loss": 4.392563989416462,
      "tokens_seen": 186908672
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004764393179538616,
      "loss": 3.4679,
      "theoretical_loss": 4.392369283331574,
      "tokens_seen": 186974208
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004764292878635908,
      "loss": 3.3004,
      "theoretical_loss": 4.392174664582085,
      "tokens_seen": 187039744
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047641925777331995,
      "loss": 3.3981,
      "theoretical_loss": 4.391980133098244,
      "tokens_seen": 187105280
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004764092276830492,
      "loss": 3.248,
      "theoretical_loss": 4.391785688810373,
      "tokens_seen": 187170816
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004763991975927783,
      "loss": 3.6385,
      "theoretical_loss": 4.391591331648879,
      "tokens_seen": 187236352
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047638916750250755,
      "loss": 3.2769,
      "theoretical_loss": 4.391397061544247,
      "tokens_seen": 187301888
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047637913741223673,
      "loss": 3.1636,
      "theoretical_loss": 4.391202878427042,
      "tokens_seen": 187367424
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004763691073219659,
      "loss": 3.3382,
      "theoretical_loss": 4.3910087822279085,
      "tokens_seen": 187432960
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004763590772316951,
      "loss": 3.3282,
      "theoretical_loss": 4.390814772877571,
      "tokens_seen": 187498496
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047634904714142427,
      "loss": 3.4497,
      "theoretical_loss": 4.390620850306832,
      "tokens_seen": 187564032
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047633901705115345,
      "loss": 3.5448,
      "theoretical_loss": 4.390427014446575,
      "tokens_seen": 187629568
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004763289869608827,
      "loss": 3.5607,
      "theoretical_loss": 4.390233265227764,
      "tokens_seen": 187695104
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004763189568706118,
      "loss": 3.2674,
      "theoretical_loss": 4.390039602581437,
      "tokens_seen": 187760640
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047630892678034105,
      "loss": 3.3578,
      "theoretical_loss": 4.389846026438715,
      "tokens_seen": 187826176
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047629889669007023,
      "loss": 3.3827,
      "theoretical_loss": 4.3896525367307975,
      "tokens_seen": 187891712
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004762888665997994,
      "loss": 3.2597,
      "theoretical_loss": 4.389459133388962,
      "tokens_seen": 187957248
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004762788365095286,
      "loss": 3.1751,
      "theoretical_loss": 4.3892658163445635,
      "tokens_seen": 188022784
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004762688064192578,
      "loss": 3.3537,
      "theoretical_loss": 4.389072585529037,
      "tokens_seen": 188088320
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047625877632898696,
      "loss": 2.8638,
      "theoretical_loss": 4.388879440873897,
      "tokens_seen": 188153856
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004762487462387162,
      "loss": 3.4062,
      "theoretical_loss": 4.388686382310732,
      "tokens_seen": 188219392
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004762387161484453,
      "loss": 3.1585,
      "theoretical_loss": 4.388493409771213,
      "tokens_seen": 188284928
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047622868605817455,
      "loss": 3.134,
      "theoretical_loss": 4.388300523187087,
      "tokens_seen": 188350464
    },
    {
      "epoch": 0.06,
      "objective/train/docs_used": 251659,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.317779541015625,
      "objective/train/theoretical_loss": 4.3881077224901786,
      "objective/train/tokens_used": 208876000,
      "theoretical_loss": 4.3881077224901786,
      "tokens_seen": 188416000
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004762186559679037,
      "loss": 3.4901,
      "theoretical_loss": 4.3881077224901786,
      "tokens_seen": 188416000
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004762086258776329,
      "loss": 3.4647,
      "theoretical_loss": 4.38791500761239,
      "tokens_seen": 188481536
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004761985957873621,
      "loss": 3.1316,
      "theoretical_loss": 4.387722378485703,
      "tokens_seen": 188547072
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004761885656970913,
      "loss": 3.1812,
      "theoretical_loss": 4.3875298350421765,
      "tokens_seen": 188612608
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047617853560682046,
      "loss": 3.2664,
      "theoretical_loss": 4.387337377213943,
      "tokens_seen": 188678144
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047616850551654964,
      "loss": 3.4667,
      "theoretical_loss": 4.387145004933218,
      "tokens_seen": 188743680
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004761584754262788,
      "loss": 3.1033,
      "theoretical_loss": 4.38695271813229,
      "tokens_seen": 188809216
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047614844533600806,
      "loss": 3.425,
      "theoretical_loss": 4.386760516743526,
      "tokens_seen": 188874752
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004761384152457372,
      "loss": 3.301,
      "theoretical_loss": 4.38656840069937,
      "tokens_seen": 188940288
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004761283851554664,
      "loss": 3.3707,
      "theoretical_loss": 4.386376369932344,
      "tokens_seen": 189005824
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004761183550651956,
      "loss": 3.3374,
      "theoretical_loss": 4.386184424375044,
      "tokens_seen": 189071360
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004761083249749248,
      "loss": 3.279,
      "theoretical_loss": 4.385992563960145,
      "tokens_seen": 189136896
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047609829488465396,
      "loss": 3.101,
      "theoretical_loss": 4.385800788620397,
      "tokens_seen": 189202432
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047608826479438314,
      "loss": 3.205,
      "theoretical_loss": 4.385609098288628,
      "tokens_seen": 189267968
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004760782347041123,
      "loss": 3.2755,
      "theoretical_loss": 4.385417492897741,
      "tokens_seen": 189333504
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047606820461384156,
      "loss": 3.3262,
      "theoretical_loss": 4.385225972380715,
      "tokens_seen": 189399040
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004760581745235707,
      "loss": 3.2513,
      "theoretical_loss": 4.385034536670606,
      "tokens_seen": 189464576
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004760481444332999,
      "loss": 3.422,
      "theoretical_loss": 4.384843185700544,
      "tokens_seen": 189530112
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047603811434302905,
      "loss": 3.6571,
      "theoretical_loss": 4.384651919403739,
      "tokens_seen": 189595648
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004760280842527583,
      "loss": 3.1847,
      "theoretical_loss": 4.384460737713471,
      "tokens_seen": 189661184
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047601805416248746,
      "loss": 3.3644,
      "theoretical_loss": 4.384269640563101,
      "tokens_seen": 189726720
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047600802407221665,
      "loss": 3.4833,
      "theoretical_loss": 4.384078627886062,
      "tokens_seen": 189792256
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004759979939819458,
      "loss": 3.2601,
      "theoretical_loss": 4.383887699615863,
      "tokens_seen": 189857792
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047598796389167506,
      "loss": 3.3371,
      "theoretical_loss": 4.38369685568609,
      "tokens_seen": 189923328
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004759779338014042,
      "loss": 3.2443,
      "theoretical_loss": 4.383506096030401,
      "tokens_seen": 189988864
    },
    {
      "epoch": 0.06,
      "objective/train/docs_used": 256551,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.438608169555664,
      "objective/train/theoretical_loss": 4.383315420582533,
      "objective/train/tokens_used": 210514400,
      "theoretical_loss": 4.383315420582533,
      "tokens_seen": 190054400
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004759679037111334,
      "loss": 3.2904,
      "theoretical_loss": 4.383315420582533,
      "tokens_seen": 190054400
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004759578736208626,
      "loss": 3.0384,
      "theoretical_loss": 4.383124829276294,
      "tokens_seen": 190119936
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004759478435305918,
      "loss": 3.1615,
      "theoretical_loss": 4.38293432204557,
      "tokens_seen": 190185472
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000475937813440321,
      "loss": 3.4141,
      "theoretical_loss": 4.382743898824321,
      "tokens_seen": 190251008
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047592778335005015,
      "loss": 3.2814,
      "theoretical_loss": 4.3825535595465785,
      "tokens_seen": 190316544
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004759177532597794,
      "loss": 3.3771,
      "theoretical_loss": 4.382363304146453,
      "tokens_seen": 190382080
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004759077231695085,
      "loss": 3.4016,
      "theoretical_loss": 4.382173132558126,
      "tokens_seen": 190447616
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047589769307923775,
      "loss": 3.306,
      "theoretical_loss": 4.381983044715856,
      "tokens_seen": 190513152
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047588766298896693,
      "loss": 3.2248,
      "theoretical_loss": 4.381793040553973,
      "tokens_seen": 190578688
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004758776328986961,
      "loss": 3.2466,
      "theoretical_loss": 4.381603120006883,
      "tokens_seen": 190644224
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004758676028084253,
      "loss": 3.325,
      "theoretical_loss": 4.381413283009065,
      "tokens_seen": 190709760
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047585757271815447,
      "loss": 3.1018,
      "theoretical_loss": 4.381223529495073,
      "tokens_seen": 190775296
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047584754262788365,
      "loss": 3.2762,
      "theoretical_loss": 4.381033859399532,
      "tokens_seen": 190840832
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004758375125376129,
      "loss": 3.5844,
      "theoretical_loss": 4.380844272657145,
      "tokens_seen": 190906368
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000475827482447342,
      "loss": 3.3999,
      "theoretical_loss": 4.380654769202683,
      "tokens_seen": 190971904
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047581745235707125,
      "loss": 3.3852,
      "theoretical_loss": 4.380465348970995,
      "tokens_seen": 191037440
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047580742226680043,
      "loss": 3.2935,
      "theoretical_loss": 4.380276011897003,
      "tokens_seen": 191102976
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004757973921765296,
      "loss": 3.2371,
      "theoretical_loss": 4.380086757915698,
      "tokens_seen": 191168512
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004757873620862588,
      "loss": 3.5679,
      "theoretical_loss": 4.379897586962148,
      "tokens_seen": 191234048
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000475777331995988,
      "loss": 3.338,
      "theoretical_loss": 4.379708498971494,
      "tokens_seen": 191299584
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047576730190571716,
      "loss": 3.1515,
      "theoretical_loss": 4.379519493878948,
      "tokens_seen": 191365120
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004757572718154464,
      "loss": 3.1375,
      "theoretical_loss": 4.379330571619795,
      "tokens_seen": 191430656
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004757472417251755,
      "loss": 3.2006,
      "theoretical_loss": 4.379141732129394,
      "tokens_seen": 191496192
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047573721163490475,
      "loss": 3.5164,
      "theoretical_loss": 4.378952975343175,
      "tokens_seen": 191561728
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004757271815446339,
      "loss": 3.4549,
      "theoretical_loss": 4.378764301196642,
      "tokens_seen": 191627264
    },
    {
      "epoch": 0.06,
      "objective/train/docs_used": 261606,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.4967339038848877,
      "objective/train/theoretical_loss": 4.37857570962537,
      "objective/train/tokens_used": 212152800,
      "theoretical_loss": 4.37857570962537,
      "tokens_seen": 191692800
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004757171514543631,
      "loss": 3.2403,
      "theoretical_loss": 4.37857570962537,
      "tokens_seen": 191692800
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004757071213640923,
      "loss": 3.305,
      "theoretical_loss": 4.378387200565006,
      "tokens_seen": 191758336
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004756970912738215,
      "loss": 3.175,
      "theoretical_loss": 4.378198773951272,
      "tokens_seen": 191823872
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047568706118355066,
      "loss": 3.3906,
      "theoretical_loss": 4.378010429719957,
      "tokens_seen": 191889408
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047567703109327984,
      "loss": 3.4164,
      "theoretical_loss": 4.377822167806928,
      "tokens_seen": 191954944
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000475667001003009,
      "loss": 3.2804,
      "theoretical_loss": 4.377633988148117,
      "tokens_seen": 192020480
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047565697091273826,
      "loss": 3.2627,
      "theoretical_loss": 4.377445890679534,
      "tokens_seen": 192086016
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004756469408224674,
      "loss": 3.1732,
      "theoretical_loss": 4.377257875337257,
      "tokens_seen": 192151552
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004756369107321966,
      "loss": 3.4658,
      "theoretical_loss": 4.377069942057436,
      "tokens_seen": 192217088
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004756268806419258,
      "loss": 3.3786,
      "theoretical_loss": 4.376882090776293,
      "tokens_seen": 192282624
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000475616850551655,
      "loss": 3.1876,
      "theoretical_loss": 4.376694321430121,
      "tokens_seen": 192348160
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047560682046138416,
      "loss": 3.2698,
      "theoretical_loss": 4.376506633955286,
      "tokens_seen": 192413696
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047559679037111334,
      "loss": 3.3082,
      "theoretical_loss": 4.376319028288219,
      "tokens_seen": 192479232
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004755867602808425,
      "loss": 3.4333,
      "theoretical_loss": 4.37613150436543,
      "tokens_seen": 192544768
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047557673019057176,
      "loss": 3.4606,
      "theoretical_loss": 4.375944062123496,
      "tokens_seen": 192610304
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004755667001003009,
      "loss": 3.0892,
      "theoretical_loss": 4.375756701499063,
      "tokens_seen": 192675840
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004755566700100301,
      "loss": 3.2806,
      "theoretical_loss": 4.3755694224288515,
      "tokens_seen": 192741376
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047554663991975925,
      "loss": 3.4885,
      "theoretical_loss": 4.375382224849648,
      "tokens_seen": 192806912
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004755366098294885,
      "loss": 3.217,
      "theoretical_loss": 4.375195108698316,
      "tokens_seen": 192872448
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047552657973921766,
      "loss": 3.277,
      "theoretical_loss": 4.375008073911781,
      "tokens_seen": 192937984
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047551654964894685,
      "loss": 3.6578,
      "theoretical_loss": 4.374821120427047,
      "tokens_seen": 193003520
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047550651955867603,
      "loss": 3.4625,
      "theoretical_loss": 4.374634248181182,
      "tokens_seen": 193069056
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047549648946840526,
      "loss": 3.3939,
      "theoretical_loss": 4.3744474571113265,
      "tokens_seen": 193134592
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004754864593781344,
      "loss": 3.4205,
      "theoretical_loss": 4.374260747154692,
      "tokens_seen": 193200128
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004754764292878636,
      "loss": 3.0036,
      "theoretical_loss": 4.374074118248559,
      "tokens_seen": 193265664
    },
    {
      "epoch": 0.06,
      "objective/train/docs_used": 266689,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.22265625,
      "objective/train/theoretical_loss": 4.373887570330275,
      "objective/train/tokens_used": 213791200,
      "theoretical_loss": 4.373887570330275,
      "tokens_seen": 193331200
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047546639919759275,
      "loss": 3.3538,
      "theoretical_loss": 4.373887570330275,
      "tokens_seen": 193331200
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000475456369107322,
      "loss": 3.0157,
      "theoretical_loss": 4.373701103337263,
      "tokens_seen": 193396736
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047544633901705117,
      "loss": 3.3198,
      "theoretical_loss": 4.373514717207009,
      "tokens_seen": 193462272
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047543630892678035,
      "loss": 3.303,
      "theoretical_loss": 4.373328411877073,
      "tokens_seen": 193527808
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047542627883650953,
      "loss": 3.4534,
      "theoretical_loss": 4.373142187285083,
      "tokens_seen": 193593344
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004754162487462387,
      "loss": 3.442,
      "theoretical_loss": 4.372956043368736,
      "tokens_seen": 193658880
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004754062186559679,
      "loss": 3.0814,
      "theoretical_loss": 4.372769980065797,
      "tokens_seen": 193724416
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047539618856569713,
      "loss": 3.2632,
      "theoretical_loss": 4.372583997314104,
      "tokens_seen": 193789952
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047538615847542625,
      "loss": 3.0204,
      "theoretical_loss": 4.372398095051559,
      "tokens_seen": 193855488
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004753761283851555,
      "loss": 3.4108,
      "theoretical_loss": 4.372212273216136,
      "tokens_seen": 193921024
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004753660982948846,
      "loss": 3.3715,
      "theoretical_loss": 4.372026531745877,
      "tokens_seen": 193986560
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047535606820461385,
      "loss": 3.1585,
      "theoretical_loss": 4.371840870578891,
      "tokens_seen": 194052096
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047534603811434303,
      "loss": 3.3463,
      "theoretical_loss": 4.37165528965336,
      "tokens_seen": 194117632
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004753360080240722,
      "loss": 3.1645,
      "theoretical_loss": 4.371469788907529,
      "tokens_seen": 194183168
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004753259779338014,
      "loss": 3.542,
      "theoretical_loss": 4.371284368279714,
      "tokens_seen": 194248704
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047531594784353063,
      "loss": 3.4428,
      "theoretical_loss": 4.3710990277083,
      "tokens_seen": 194314240
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047530591775325976,
      "loss": 3.2136,
      "theoretical_loss": 4.3709137671317375,
      "tokens_seen": 194379776
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000475295887662989,
      "loss": 3.5123,
      "theoretical_loss": 4.37072858648855,
      "tokens_seen": 194445312
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004752858575727181,
      "loss": 3.4088,
      "theoretical_loss": 4.370543485717322,
      "tokens_seen": 194510848
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047527582748244736,
      "loss": 3.2998,
      "theoretical_loss": 4.370358464756713,
      "tokens_seen": 194576384
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047526579739217654,
      "loss": 3.0609,
      "theoretical_loss": 4.370173523545443,
      "tokens_seen": 194641920
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004752557673019057,
      "loss": 3.1257,
      "theoretical_loss": 4.3699886620223065,
      "tokens_seen": 194707456
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004752457372116349,
      "loss": 3.3654,
      "theoretical_loss": 4.369803880126162,
      "tokens_seen": 194772992
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004752357071213641,
      "loss": 3.362,
      "theoretical_loss": 4.3696191777959354,
      "tokens_seen": 194838528
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047522567703109326,
      "loss": 3.3996,
      "theoretical_loss": 4.369434554970621,
      "tokens_seen": 194904064
    },
    {
      "epoch": 0.06,
      "objective/train/docs_used": 269630,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.62288236618042,
      "objective/train/theoretical_loss": 4.369250011589279,
      "objective/train/tokens_used": 215429600,
      "theoretical_loss": 4.369250011589279,
      "tokens_seen": 194969600
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004752156469408225,
      "loss": 3.4504,
      "theoretical_loss": 4.369250011589279,
      "tokens_seen": 194969600
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004752056168505517,
      "loss": 3.3655,
      "theoretical_loss": 4.369065547591038,
      "tokens_seen": 195035136
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047519558676028086,
      "loss": 3.2661,
      "theoretical_loss": 4.368881162915095,
      "tokens_seen": 195100672
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047518555667001004,
      "loss": 3.3123,
      "theoretical_loss": 4.36869685750071,
      "tokens_seen": 195166208
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004751755265797392,
      "loss": 3.2071,
      "theoretical_loss": 4.3685126312872145,
      "tokens_seen": 195231744
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047516549648946846,
      "loss": 3.5141,
      "theoretical_loss": 4.368328484214002,
      "tokens_seen": 195297280
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004751554663991976,
      "loss": 3.4057,
      "theoretical_loss": 4.368144416220538,
      "tokens_seen": 195362816
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004751454363089268,
      "loss": 3.34,
      "theoretical_loss": 4.3679604272463495,
      "tokens_seen": 195428352
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000475135406218656,
      "loss": 3.2522,
      "theoretical_loss": 4.367776517231033,
      "tokens_seen": 195493888
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004751253761283852,
      "loss": 3.1821,
      "theoretical_loss": 4.367592686114252,
      "tokens_seen": 195559424
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047511534603811436,
      "loss": 3.3659,
      "theoretical_loss": 4.367408933835733,
      "tokens_seen": 195624960
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047510531594784354,
      "loss": 3.3274,
      "theoretical_loss": 4.367225260335272,
      "tokens_seen": 195690496
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004750952858575727,
      "loss": 3.1391,
      "theoretical_loss": 4.36704166555273,
      "tokens_seen": 195756032
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047508525576730196,
      "loss": 3.2901,
      "theoretical_loss": 4.366858149428032,
      "tokens_seen": 195821568
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004750752256770311,
      "loss": 3.0,
      "theoretical_loss": 4.366674711901173,
      "tokens_seen": 195887104
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004750651955867603,
      "loss": 3.3856,
      "theoretical_loss": 4.366491352912211,
      "tokens_seen": 195952640
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047505516549648945,
      "loss": 3.369,
      "theoretical_loss": 4.366308072401271,
      "tokens_seen": 196018176
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004750451354062187,
      "loss": 3.1375,
      "theoretical_loss": 4.366124870308541,
      "tokens_seen": 196083712
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047503510531594787,
      "loss": 3.3023,
      "theoretical_loss": 4.365941746574278,
      "tokens_seen": 196149248
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047502507522567705,
      "loss": 3.4409,
      "theoretical_loss": 4.3657587011388035,
      "tokens_seen": 196214784
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047501504513540623,
      "loss": 3.4023,
      "theoretical_loss": 4.365575733942503,
      "tokens_seen": 196280320
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047500501504513546,
      "loss": 3.1508,
      "theoretical_loss": 4.365392844925829,
      "tokens_seen": 196345856
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004749949849548646,
      "loss": 3.2641,
      "theoretical_loss": 4.365210034029298,
      "tokens_seen": 196411392
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004749849548645938,
      "loss": 3.2689,
      "theoretical_loss": 4.365027301193491,
      "tokens_seen": 196476928
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047497492477432295,
      "loss": 3.3907,
      "theoretical_loss": 4.364844646359056,
      "tokens_seen": 196542464
    },
    {
      "epoch": 0.06,
      "objective/train/docs_used": 270311,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.453747272491455,
      "objective/train/theoretical_loss": 4.364662069466704,
      "objective/train/tokens_used": 217068000,
      "theoretical_loss": 4.364662069466704,
      "tokens_seen": 196608000
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004749648946840522,
      "loss": 3.4165,
      "theoretical_loss": 4.364662069466704,
      "tokens_seen": 196608000
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047495486459378137,
      "loss": 3.2901,
      "theoretical_loss": 4.364479570457213,
      "tokens_seen": 196673536
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047494483450351055,
      "loss": 3.3529,
      "theoretical_loss": 4.364297149271423,
      "tokens_seen": 196739072
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047493480441323973,
      "loss": 3.4174,
      "theoretical_loss": 4.3641148058502415,
      "tokens_seen": 196804608
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004749247743229689,
      "loss": 3.203,
      "theoretical_loss": 4.363932540134638,
      "tokens_seen": 196870144
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004749147442326981,
      "loss": 3.1933,
      "theoretical_loss": 4.363750352065647,
      "tokens_seen": 196935680
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047490471414242733,
      "loss": 3.2939,
      "theoretical_loss": 4.363568241584368,
      "tokens_seen": 197001216
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047489468405215646,
      "loss": 3.1258,
      "theoretical_loss": 4.363386208631966,
      "tokens_seen": 197066752
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004748846539618857,
      "loss": 3.2392,
      "theoretical_loss": 4.363204253149667,
      "tokens_seen": 197132288
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004748746238716148,
      "loss": 3.283,
      "theoretical_loss": 4.3630223750787644,
      "tokens_seen": 197197824
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047486459378134405,
      "loss": 3.3828,
      "theoretical_loss": 4.362840574360612,
      "tokens_seen": 197263360
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047485456369107323,
      "loss": 3.0395,
      "theoretical_loss": 4.362658850936631,
      "tokens_seen": 197328896
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004748445336008024,
      "loss": 3.2909,
      "theoretical_loss": 4.362477204748305,
      "tokens_seen": 197394432
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004748345035105316,
      "loss": 3.1958,
      "theoretical_loss": 4.362295635737179,
      "tokens_seen": 197459968
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047482447342026083,
      "loss": 2.9771,
      "theoretical_loss": 4.362114143844867,
      "tokens_seen": 197525504
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047481444332998996,
      "loss": 3.289,
      "theoretical_loss": 4.3619327290130405,
      "tokens_seen": 197591040
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004748044132397192,
      "loss": 3.2772,
      "theoretical_loss": 4.3617513911834385,
      "tokens_seen": 197656576
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004747943831494483,
      "loss": 3.149,
      "theoretical_loss": 4.361570130297863,
      "tokens_seen": 197722112
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047478435305917756,
      "loss": 3.3416,
      "theoretical_loss": 4.3613889462981765,
      "tokens_seen": 197787648
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047477432296890674,
      "loss": 3.4184,
      "theoretical_loss": 4.361207839126308,
      "tokens_seen": 197853184
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004747642928786359,
      "loss": 3.3547,
      "theoretical_loss": 4.361026808724247,
      "tokens_seen": 197918720
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004747542627883651,
      "loss": 3.1468,
      "theoretical_loss": 4.360845855034049,
      "tokens_seen": 197984256
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004747442326980943,
      "loss": 3.2702,
      "theoretical_loss": 4.360664977997828,
      "tokens_seen": 198049792
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047473420260782346,
      "loss": 2.9289,
      "theoretical_loss": 4.360484177557766,
      "tokens_seen": 198115328
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004747241725175527,
      "loss": 3.3127,
      "theoretical_loss": 4.360303453656103,
      "tokens_seen": 198180864
    },
    {
      "epoch": 0.06,
      "objective/train/docs_used": 271760,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.187415838241577,
      "objective/train/theoretical_loss": 4.360122806235145,
      "objective/train/tokens_used": 218706400,
      "theoretical_loss": 4.360122806235145,
      "tokens_seen": 198246400
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004747141424272818,
      "loss": 3.2371,
      "theoretical_loss": 4.360122806235145,
      "tokens_seen": 198246400
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047470411233701106,
      "loss": 3.1704,
      "theoretical_loss": 4.359942235237257,
      "tokens_seen": 198311936
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004746940822467402,
      "loss": 3.375,
      "theoretical_loss": 4.359761740604871,
      "tokens_seen": 198377472
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004746840521564694,
      "loss": 3.0178,
      "theoretical_loss": 4.359581322280479,
      "tokens_seen": 198443008
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004746740220661986,
      "loss": 2.9285,
      "theoretical_loss": 4.359400980206634,
      "tokens_seen": 198508544
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004746639919759278,
      "loss": 3.3454,
      "theoretical_loss": 4.359220714325954,
      "tokens_seen": 198574080
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047465396188565696,
      "loss": 3.3081,
      "theoretical_loss": 4.359040524581116,
      "tokens_seen": 198639616
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004746439317953862,
      "loss": 3.4269,
      "theoretical_loss": 4.358860410914861,
      "tokens_seen": 198705152
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004746339017051153,
      "loss": 3.2979,
      "theoretical_loss": 4.358680373269993,
      "tokens_seen": 198770688
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047462387161484456,
      "loss": 3.2567,
      "theoretical_loss": 4.358500411589375,
      "tokens_seen": 198836224
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004746138415245737,
      "loss": 3.373,
      "theoretical_loss": 4.358320525815934,
      "tokens_seen": 198901760
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004746038114343029,
      "loss": 3.4155,
      "theoretical_loss": 4.358140715892658,
      "tokens_seen": 198967296
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004745937813440321,
      "loss": 3.5005,
      "theoretical_loss": 4.357960981762595,
      "tokens_seen": 199032832
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004745837512537613,
      "loss": 3.145,
      "theoretical_loss": 4.357781323368857,
      "tokens_seen": 199098368
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047457372116349047,
      "loss": 3.1356,
      "theoretical_loss": 4.357601740654617,
      "tokens_seen": 199163904
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047456369107321965,
      "loss": 3.3813,
      "theoretical_loss": 4.357422233563106,
      "tokens_seen": 199229440
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047455366098294883,
      "loss": 3.2667,
      "theoretical_loss": 4.357242802037623,
      "tokens_seen": 199294976
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047454363089267807,
      "loss": 3.5508,
      "theoretical_loss": 4.35706344602152,
      "tokens_seen": 199360512
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004745336008024072,
      "loss": 3.3062,
      "theoretical_loss": 4.356884165458217,
      "tokens_seen": 199426048
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047452357071213643,
      "loss": 3.3419,
      "theoretical_loss": 4.356704960291191,
      "tokens_seen": 199491584
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047451354062186555,
      "loss": 3.1005,
      "theoretical_loss": 4.35652583046398,
      "tokens_seen": 199557120
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004745035105315948,
      "loss": 3.1586,
      "theoretical_loss": 4.356346775920185,
      "tokens_seen": 199622656
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047449348044132397,
      "loss": 2.9514,
      "theoretical_loss": 4.356167796603467,
      "tokens_seen": 199688192
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047448345035105315,
      "loss": 3.5216,
      "theoretical_loss": 4.355988892457546,
      "tokens_seen": 199753728
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004744734202607824,
      "loss": 3.4248,
      "theoretical_loss": 4.355810063426204,
      "tokens_seen": 199819264
    },
    {
      "epoch": 0.06,
      "objective/train/docs_used": 272554,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.5305120944976807,
      "objective/train/theoretical_loss": 4.355631309453283,
      "objective/train/tokens_used": 220344800,
      "theoretical_loss": 4.355631309453283,
      "tokens_seen": 199884800
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047446339017051157,
      "loss": 3.2918,
      "theoretical_loss": 4.355631309453283,
      "tokens_seen": 199884800
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047445336008024075,
      "loss": 3.3689,
      "theoretical_loss": 4.355452630482685,
      "tokens_seen": 199950336
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047444332998996993,
      "loss": 3.1337,
      "theoretical_loss": 4.355274026458375,
      "tokens_seen": 200015872
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004744332998996991,
      "loss": 3.1743,
      "theoretical_loss": 4.355095497324373,
      "tokens_seen": 200081408
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004744232698094283,
      "loss": 3.3082,
      "theoretical_loss": 4.354917043024765,
      "tokens_seen": 200146944
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047441323971915753,
      "loss": 3.4852,
      "theoretical_loss": 4.354738663503692,
      "tokens_seen": 200212480
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047440320962888666,
      "loss": 3.2587,
      "theoretical_loss": 4.354560358705358,
      "tokens_seen": 200278016
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004743931795386159,
      "loss": 2.9646,
      "theoretical_loss": 4.354382128574027,
      "tokens_seen": 200343552
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000474383149448345,
      "loss": 3.1041,
      "theoretical_loss": 4.35420397305402,
      "tokens_seen": 200409088
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047437311935807425,
      "loss": 3.532,
      "theoretical_loss": 4.35402589208972,
      "tokens_seen": 200474624
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047436308926780343,
      "loss": 3.3644,
      "theoretical_loss": 4.353847885625571,
      "tokens_seen": 200540160
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004743530591775326,
      "loss": 3.3038,
      "theoretical_loss": 4.353669953606072,
      "tokens_seen": 200605696
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004743430290872618,
      "loss": 3.1885,
      "theoretical_loss": 4.353492095975787,
      "tokens_seen": 200671232
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047433299899699103,
      "loss": 3.1056,
      "theoretical_loss": 4.353314312679333,
      "tokens_seen": 200736768
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047432296890672016,
      "loss": 3.1679,
      "theoretical_loss": 4.353136603661392,
      "tokens_seen": 200802304
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004743129388164494,
      "loss": 3.3277,
      "theoretical_loss": 4.352958968866704,
      "tokens_seen": 200867840
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004743029087261785,
      "loss": 3.3515,
      "theoretical_loss": 4.352781408240065,
      "tokens_seen": 200933376
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047429287863590776,
      "loss": 3.4612,
      "theoretical_loss": 4.352603921726334,
      "tokens_seen": 200998912
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047428284854563694,
      "loss": 3.2698,
      "theoretical_loss": 4.352426509270425,
      "tokens_seen": 201064448
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004742728184553661,
      "loss": 3.2832,
      "theoretical_loss": 4.352249170817315,
      "tokens_seen": 201129984
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004742627883650953,
      "loss": 3.2757,
      "theoretical_loss": 4.352071906312037,
      "tokens_seen": 201195520
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004742527582748245,
      "loss": 3.2153,
      "theoretical_loss": 4.351894715699684,
      "tokens_seen": 201261056
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047424272818455366,
      "loss": 3.5588,
      "theoretical_loss": 4.351717598925406,
      "tokens_seen": 201326592
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004742326980942829,
      "loss": 3.1392,
      "theoretical_loss": 4.351540555934414,
      "tokens_seen": 201392128
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000474222668004012,
      "loss": 3.2376,
      "theoretical_loss": 4.351363586671976,
      "tokens_seen": 201457664
    },
    {
      "epoch": 0.06,
      "objective/train/docs_used": 274046,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8291563987731934,
      "objective/train/theoretical_loss": 4.351186691083417,
      "objective/train/tokens_used": 221983200,
      "theoretical_loss": 4.351186691083417,
      "tokens_seen": 201523200
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047421263791374126,
      "loss": 3.179,
      "theoretical_loss": 4.351186691083417,
      "tokens_seen": 201523200
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004742026078234704,
      "loss": 3.047,
      "theoretical_loss": 4.351009869114124,
      "tokens_seen": 201588736
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004741925777331996,
      "loss": 3.2204,
      "theoretical_loss": 4.350833120709539,
      "tokens_seen": 201654272
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004741825476429288,
      "loss": 3.2638,
      "theoretical_loss": 4.350656445815164,
      "tokens_seen": 201719808
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000474172517552658,
      "loss": 3.3704,
      "theoretical_loss": 4.350479844376557,
      "tokens_seen": 201785344
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047416248746238716,
      "loss": 3.2403,
      "theoretical_loss": 4.350303316339337,
      "tokens_seen": 201850880
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004741524573721164,
      "loss": 3.1801,
      "theoretical_loss": 4.350126861649178,
      "tokens_seen": 201916416
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004741424272818455,
      "loss": 3.0457,
      "theoretical_loss": 4.349950480251813,
      "tokens_seen": 201981952
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047413239719157476,
      "loss": 3.3629,
      "theoretical_loss": 4.349774172093033,
      "tokens_seen": 202047488
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004741223671013039,
      "loss": 3.2814,
      "theoretical_loss": 4.349597937118687,
      "tokens_seen": 202113024
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004741123370110331,
      "loss": 3.6036,
      "theoretical_loss": 4.3494217752746795,
      "tokens_seen": 202178560
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004741023069207623,
      "loss": 3.2269,
      "theoretical_loss": 4.349245686506976,
      "tokens_seen": 202244096
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004740922768304915,
      "loss": 3.4228,
      "theoretical_loss": 4.349069670761597,
      "tokens_seen": 202309632
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047408224674022067,
      "loss": 3.4884,
      "theoretical_loss": 4.348893727984619,
      "tokens_seen": 202375168
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047407221664994985,
      "loss": 3.2129,
      "theoretical_loss": 4.348717858122178,
      "tokens_seen": 202440704
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047406218655967903,
      "loss": 3.1145,
      "theoretical_loss": 4.348542061120469,
      "tokens_seen": 202506240
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047405215646940827,
      "loss": 3.4848,
      "theoretical_loss": 4.348366336925739,
      "tokens_seen": 202571776
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004740421263791374,
      "loss": 3.3138,
      "theoretical_loss": 4.3481906854842975,
      "tokens_seen": 202637312
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047403209628886663,
      "loss": 3.3594,
      "theoretical_loss": 4.348015106742507,
      "tokens_seen": 202702848
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047402206619859575,
      "loss": 3.4342,
      "theoretical_loss": 4.347839600646786,
      "tokens_seen": 202768384
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000474012036108325,
      "loss": 3.2728,
      "theoretical_loss": 4.347664167143615,
      "tokens_seen": 202833920
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047400200601805417,
      "loss": 3.491,
      "theoretical_loss": 4.347488806179528,
      "tokens_seen": 202899456
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047399197592778335,
      "loss": 3.2498,
      "theoretical_loss": 4.347313517701114,
      "tokens_seen": 202964992
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047398194583751253,
      "loss": 3.2725,
      "theoretical_loss": 4.347138301655021,
      "tokens_seen": 203030528
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047397191574724177,
      "loss": 3.2918,
      "theoretical_loss": 4.346963157987954,
      "tokens_seen": 203096064
    },
    {
      "epoch": 0.06,
      "objective/train/docs_used": 274684,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.094592332839966,
      "objective/train/theoretical_loss": 4.346788086646671,
      "objective/train/tokens_used": 223621600,
      "theoretical_loss": 4.346788086646671,
      "tokens_seen": 203161600
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004739618856569709,
      "loss": 3.2821,
      "theoretical_loss": 4.346788086646671,
      "tokens_seen": 203161600
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047395185556670013,
      "loss": 3.1994,
      "theoretical_loss": 4.346613087577991,
      "tokens_seen": 203227136
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047394182547642926,
      "loss": 3.276,
      "theoretical_loss": 4.346438160728785,
      "tokens_seen": 203292672
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004739317953861585,
      "loss": 3.4814,
      "theoretical_loss": 4.346263306045983,
      "tokens_seen": 203358208
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004739217652958877,
      "loss": 3.3933,
      "theoretical_loss": 4.346088523476569,
      "tokens_seen": 203423744
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047391173520561686,
      "loss": 3.1577,
      "theoretical_loss": 4.345913812967584,
      "tokens_seen": 203489280
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047390170511534604,
      "loss": 3.2551,
      "theoretical_loss": 4.345739174466127,
      "tokens_seen": 203554816
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004738916750250752,
      "loss": 3.3351,
      "theoretical_loss": 4.345564607919348,
      "tokens_seen": 203620352
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004738816449348044,
      "loss": 3.2087,
      "theoretical_loss": 4.3453901132744575,
      "tokens_seen": 203685888
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047387161484453363,
      "loss": 3.5707,
      "theoretical_loss": 4.345215690478719,
      "tokens_seen": 203751424
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047386158475426276,
      "loss": 3.2015,
      "theoretical_loss": 4.345041339479453,
      "tokens_seen": 203816960
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000473851554663992,
      "loss": 3.2998,
      "theoretical_loss": 4.3448670602240345,
      "tokens_seen": 203882496
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004738415245737211,
      "loss": 3.2642,
      "theoretical_loss": 4.344692852659895,
      "tokens_seen": 203948032
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047383149448345036,
      "loss": 3.307,
      "theoretical_loss": 4.34451871673452,
      "tokens_seen": 204013568
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047382146439317954,
      "loss": 3.3531,
      "theoretical_loss": 4.344344652395451,
      "tokens_seen": 204079104
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004738114343029087,
      "loss": 3.3066,
      "theoretical_loss": 4.3441706595902865,
      "tokens_seen": 204144640
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004738014042126379,
      "loss": 3.0495,
      "theoretical_loss": 4.343996738266677,
      "tokens_seen": 204210176
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047379137412236714,
      "loss": 3.3121,
      "theoretical_loss": 4.343822888372331,
      "tokens_seen": 204275712
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047378134403209626,
      "loss": 3.32,
      "theoretical_loss": 4.343649109855009,
      "tokens_seen": 204341248
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004737713139418255,
      "loss": 3.2897,
      "theoretical_loss": 4.343475402662529,
      "tokens_seen": 204406784
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004737612838515546,
      "loss": 3.325,
      "theoretical_loss": 4.343301766742763,
      "tokens_seen": 204472320
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047375125376128386,
      "loss": 3.2351,
      "theoretical_loss": 4.343128202043638,
      "tokens_seen": 204537856
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047374122367101304,
      "loss": 3.4129,
      "theoretical_loss": 4.342954708513136,
      "tokens_seen": 204603392
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004737311935807422,
      "loss": 3.2237,
      "theoretical_loss": 4.342781286099291,
      "tokens_seen": 204668928
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047372116349047146,
      "loss": 3.224,
      "theoretical_loss": 4.3426079347501965,
      "tokens_seen": 204734464
    },
    {
      "epoch": 0.06,
      "objective/train/docs_used": 275206,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.4942283630371094,
      "objective/train/theoretical_loss": 4.342434654413995,
      "objective/train/tokens_used": 225260000,
      "theoretical_loss": 4.342434654413995,
      "tokens_seen": 204800000
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004737111334002006,
      "loss": 3.4119,
      "theoretical_loss": 4.342434654413995,
      "tokens_seen": 204800000
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004737011033099298,
      "loss": 3.3759,
      "theoretical_loss": 4.342261445038888,
      "tokens_seen": 204865536
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000473691073219659,
      "loss": 3.1406,
      "theoretical_loss": 4.342088306573128,
      "tokens_seen": 204931072
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004736810431293882,
      "loss": 3.2741,
      "theoretical_loss": 4.341915238965026,
      "tokens_seen": 204996608
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047367101303911736,
      "loss": 3.2403,
      "theoretical_loss": 4.34174224216294,
      "tokens_seen": 205062144
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004736609829488466,
      "loss": 3.1585,
      "theoretical_loss": 4.34156931611529,
      "tokens_seen": 205127680
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004736509528585757,
      "loss": 3.2647,
      "theoretical_loss": 4.341396460770547,
      "tokens_seen": 205193216
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047364092276830496,
      "loss": 3.3709,
      "theoretical_loss": 4.341223676077232,
      "tokens_seen": 205258752
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004736308926780341,
      "loss": 3.2875,
      "theoretical_loss": 4.341050961983926,
      "tokens_seen": 205324288
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004736208625877633,
      "loss": 3.2638,
      "theoretical_loss": 4.340878318439261,
      "tokens_seen": 205389824
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004736108324974925,
      "loss": 3.563,
      "theoretical_loss": 4.340705745391922,
      "tokens_seen": 205455360
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004736008024072217,
      "loss": 3.3661,
      "theoretical_loss": 4.3405332427906504,
      "tokens_seen": 205520896
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047359077231695087,
      "loss": 3.5278,
      "theoretical_loss": 4.340360810584238,
      "tokens_seen": 205586432
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047358074222668005,
      "loss": 3.3486,
      "theoretical_loss": 4.340188448721532,
      "tokens_seen": 205651968
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047357071213640923,
      "loss": 3.3382,
      "theoretical_loss": 4.3400161571514335,
      "tokens_seen": 205717504
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047356068204613847,
      "loss": 3.156,
      "theoretical_loss": 4.339843935822895,
      "tokens_seen": 205783040
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004735506519558676,
      "loss": 3.4644,
      "theoretical_loss": 4.339671784684923,
      "tokens_seen": 205848576
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047354062186559683,
      "loss": 3.0953,
      "theoretical_loss": 4.339499703686579,
      "tokens_seen": 205914112
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047353059177532595,
      "loss": 3.1301,
      "theoretical_loss": 4.339327692776977,
      "tokens_seen": 205979648
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004735205616850552,
      "loss": 3.3351,
      "theoretical_loss": 4.339155751905282,
      "tokens_seen": 206045184
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047351053159478437,
      "loss": 3.3348,
      "theoretical_loss": 4.338983881020713,
      "tokens_seen": 206110720
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047350050150451355,
      "loss": 3.3371,
      "theoretical_loss": 4.338812080072545,
      "tokens_seen": 206176256
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047349047141424273,
      "loss": 3.2751,
      "theoretical_loss": 4.338640349010101,
      "tokens_seen": 206241792
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047348044132397197,
      "loss": 3.1762,
      "theoretical_loss": 4.3384686877827585,
      "tokens_seen": 206307328
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004734704112337011,
      "loss": 3.3436,
      "theoretical_loss": 4.338297096339951,
      "tokens_seen": 206372864
    },
    {
      "epoch": 0.06,
      "objective/train/docs_used": 276466,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.012336015701294,
      "objective/train/theoretical_loss": 4.33812557463116,
      "objective/train/tokens_used": 226898400,
      "theoretical_loss": 4.33812557463116,
      "tokens_seen": 206438400
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047346038114343033,
      "loss": 3.0366,
      "theoretical_loss": 4.33812557463116,
      "tokens_seen": 206438400
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047345035105315946,
      "loss": 3.2347,
      "theoretical_loss": 4.3379541226059235,
      "tokens_seen": 206503936
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004734403209628887,
      "loss": 3.1413,
      "theoretical_loss": 4.337782740213827,
      "tokens_seen": 206569472
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004734302908726179,
      "loss": 3.0938,
      "theoretical_loss": 4.337611427404514,
      "tokens_seen": 206635008
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047342026078234706,
      "loss": 3.0309,
      "theoretical_loss": 4.337440184127679,
      "tokens_seen": 206700544
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047341023069207624,
      "loss": 3.3016,
      "theoretical_loss": 4.337269010333065,
      "tokens_seen": 206766080
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004734002006018054,
      "loss": 3.315,
      "theoretical_loss": 4.337097905970471,
      "tokens_seen": 206831616
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004733901705115346,
      "loss": 3.5151,
      "theoretical_loss": 4.336926870989748,
      "tokens_seen": 206897152
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047338014042126383,
      "loss": 3.2054,
      "theoretical_loss": 4.336755905340797,
      "tokens_seen": 206962688
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047337011033099296,
      "loss": 3.1878,
      "theoretical_loss": 4.336585008973573,
      "tokens_seen": 207028224
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004733600802407222,
      "loss": 3.2425,
      "theoretical_loss": 4.336414181838082,
      "tokens_seen": 207093760
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004733500501504513,
      "loss": 3.5244,
      "theoretical_loss": 4.336243423884382,
      "tokens_seen": 207159296
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047334002006018056,
      "loss": 3.153,
      "theoretical_loss": 4.336072735062583,
      "tokens_seen": 207224832
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047332998996990974,
      "loss": 3.2415,
      "theoretical_loss": 4.335902115322847,
      "tokens_seen": 207290368
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004733199598796389,
      "loss": 3.0556,
      "theoretical_loss": 4.335731564615387,
      "tokens_seen": 207355904
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004733099297893681,
      "loss": 3.1923,
      "theoretical_loss": 4.335561082890468,
      "tokens_seen": 207421440
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047329989969909734,
      "loss": 3.2044,
      "theoretical_loss": 4.335390670098407,
      "tokens_seen": 207486976
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047328986960882646,
      "loss": 3.5337,
      "theoretical_loss": 4.335220326189571,
      "tokens_seen": 207552512
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004732798395185557,
      "loss": 3.2499,
      "theoretical_loss": 4.335050051114379,
      "tokens_seen": 207618048
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004732698094282848,
      "loss": 3.1384,
      "theoretical_loss": 4.334879844823304,
      "tokens_seen": 207683584
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047325977933801406,
      "loss": 3.1074,
      "theoretical_loss": 4.334709707266865,
      "tokens_seen": 207749120
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047324974924774324,
      "loss": 3.1987,
      "theoretical_loss": 4.334539638395636,
      "tokens_seen": 207814656
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004732397191574724,
      "loss": 3.1185,
      "theoretical_loss": 4.334369638160242,
      "tokens_seen": 207880192
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004732296890672016,
      "loss": 3.2328,
      "theoretical_loss": 4.334199706511358,
      "tokens_seen": 207945728
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004732196589769308,
      "loss": 3.0938,
      "theoretical_loss": 4.334029843399709,
      "tokens_seen": 208011264
    },
    {
      "epoch": 0.06,
      "objective/train/docs_used": 277178,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.992138385772705,
      "objective/train/theoretical_loss": 4.333860048776074,
      "objective/train/tokens_used": 228536800,
      "theoretical_loss": 4.333860048776074,
      "tokens_seen": 208076800
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047320962888665997,
      "loss": 3.1004,
      "theoretical_loss": 4.333860048776074,
      "tokens_seen": 208076800
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004731995987963892,
      "loss": 3.2221,
      "theoretical_loss": 4.33369032259128,
      "tokens_seen": 208142336
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047318956870611833,
      "loss": 3.2591,
      "theoretical_loss": 4.333520664796206,
      "tokens_seen": 208207872
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047317953861584757,
      "loss": 3.4427,
      "theoretical_loss": 4.33335107534178,
      "tokens_seen": 208273408
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047316950852557675,
      "loss": 3.2497,
      "theoretical_loss": 4.333181554178985,
      "tokens_seen": 208338944
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047315947843530593,
      "loss": 3.293,
      "theoretical_loss": 4.3330121012588485,
      "tokens_seen": 208404480
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004731494483450351,
      "loss": 3.5665,
      "theoretical_loss": 4.332842716532454,
      "tokens_seen": 208470016
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004731394182547643,
      "loss": 3.0097,
      "theoretical_loss": 4.332673399950932,
      "tokens_seen": 208535552
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047312938816449347,
      "loss": 3.1251,
      "theoretical_loss": 4.332504151465464,
      "tokens_seen": 208601088
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004731193580742227,
      "loss": 3.0934,
      "theoretical_loss": 4.332334971027284,
      "tokens_seen": 208666624
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047310932798395183,
      "loss": 3.4214,
      "theoretical_loss": 4.332165858587672,
      "tokens_seen": 208732160
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047309929789368107,
      "loss": 3.0211,
      "theoretical_loss": 4.331996814097963,
      "tokens_seen": 208797696
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004730892678034102,
      "loss": 3.2003,
      "theoretical_loss": 4.331827837509538,
      "tokens_seen": 208863232
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047307923771313943,
      "loss": 3.3196,
      "theoretical_loss": 4.331658928773831,
      "tokens_seen": 208928768
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004730692076228686,
      "loss": 3.2202,
      "theoretical_loss": 4.331490087842324,
      "tokens_seen": 208994304
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004730591775325978,
      "loss": 3.1351,
      "theoretical_loss": 4.33132131466655,
      "tokens_seen": 209059840
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000473049147442327,
      "loss": 3.212,
      "theoretical_loss": 4.3311526091980905,
      "tokens_seen": 209125376
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047303911735205615,
      "loss": 3.2426,
      "theoretical_loss": 4.330983971388578,
      "tokens_seen": 209190912
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047302908726178534,
      "loss": 3.2021,
      "theoretical_loss": 4.330815401189695,
      "tokens_seen": 209256448
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047301905717151457,
      "loss": 3.2868,
      "theoretical_loss": 4.330646898553173,
      "tokens_seen": 209321984
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004730090270812437,
      "loss": 3.2098,
      "theoretical_loss": 4.330478463430792,
      "tokens_seen": 209387520
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047299899699097293,
      "loss": 3.0737,
      "theoretical_loss": 4.330310095774383,
      "tokens_seen": 209453056
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004729889669007021,
      "loss": 3.3302,
      "theoretical_loss": 4.330141795535828,
      "tokens_seen": 209518592
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004729789368104313,
      "loss": 3.1422,
      "theoretical_loss": 4.329973562667053,
      "tokens_seen": 209584128
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047296890672016053,
      "loss": 3.3357,
      "theoretical_loss": 4.3298053971200385,
      "tokens_seen": 209649664
    },
    {
      "epoch": 0.06,
      "objective/train/docs_used": 278592,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8049657344818115,
      "objective/train/theoretical_loss": 4.329637298846812,
      "objective/train/tokens_used": 230175200,
      "theoretical_loss": 4.329637298846812,
      "tokens_seen": 209715200
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047295887662988966,
      "loss": 2.8862,
      "theoretical_loss": 4.329637298846812,
      "tokens_seen": 209715200
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004729488465396189,
      "loss": 3.3877,
      "theoretical_loss": 4.329469267799451,
      "tokens_seen": 209780736
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004729388164493481,
      "loss": 3.2828,
      "theoretical_loss": 4.32930130393008,
      "tokens_seen": 209846272
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047292878635907726,
      "loss": 3.2585,
      "theoretical_loss": 4.329133407190876,
      "tokens_seen": 209911808
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047291875626880644,
      "loss": 3.21,
      "theoretical_loss": 4.3289655775340625,
      "tokens_seen": 209977344
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004729087261785356,
      "loss": 3.2672,
      "theoretical_loss": 4.328797814911912,
      "tokens_seen": 210042880
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004728986960882648,
      "loss": 3.3022,
      "theoretical_loss": 4.328630119276747,
      "tokens_seen": 210108416
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047288866599799403,
      "loss": 3.0099,
      "theoretical_loss": 4.328462490580938,
      "tokens_seen": 210173952
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047287863590772316,
      "loss": 3.2968,
      "theoretical_loss": 4.328294928776903,
      "tokens_seen": 210239488
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004728686058174524,
      "loss": 3.1646,
      "theoretical_loss": 4.328127433817112,
      "tokens_seen": 210305024
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004728585757271815,
      "loss": 3.2345,
      "theoretical_loss": 4.327960005654081,
      "tokens_seen": 210370560
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047284854563691076,
      "loss": 3.3589,
      "theoretical_loss": 4.327792644240374,
      "tokens_seen": 210436096
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047283851554663994,
      "loss": 3.0922,
      "theoretical_loss": 4.327625349528605,
      "tokens_seen": 210501632
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004728284854563691,
      "loss": 3.3172,
      "theoretical_loss": 4.327458121471436,
      "tokens_seen": 210567168
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004728184553660983,
      "loss": 3.3431,
      "theoretical_loss": 4.3272909600215765,
      "tokens_seen": 210632704
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047280842527582754,
      "loss": 3.1134,
      "theoretical_loss": 4.327123865131786,
      "tokens_seen": 210698240
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047279839518555666,
      "loss": 3.1697,
      "theoretical_loss": 4.326956836754871,
      "tokens_seen": 210763776
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004727883650952859,
      "loss": 3.4017,
      "theoretical_loss": 4.326789874843685,
      "tokens_seen": 210829312
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000472778335005015,
      "loss": 3.1166,
      "theoretical_loss": 4.326622979351132,
      "tokens_seen": 210894848
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047276830491474426,
      "loss": 3.0739,
      "theoretical_loss": 4.326456150230163,
      "tokens_seen": 210960384
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047275827482447344,
      "loss": 3.2434,
      "theoretical_loss": 4.326289387433776,
      "tokens_seen": 211025920
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004727482447342026,
      "loss": 3.1457,
      "theoretical_loss": 4.326122690915017,
      "tokens_seen": 211091456
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004727382146439318,
      "loss": 3.5019,
      "theoretical_loss": 4.325956060626982,
      "tokens_seen": 211156992
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000472728184553661,
      "loss": 3.1354,
      "theoretical_loss": 4.325789496522812,
      "tokens_seen": 211222528
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047271815446339017,
      "loss": 3.1215,
      "theoretical_loss": 4.325622998555697,
      "tokens_seen": 211288064
    },
    {
      "epoch": 0.06,
      "objective/train/docs_used": 279319,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.034355640411377,
      "objective/train/theoretical_loss": 4.3254565666788745,
      "objective/train/tokens_used": 231813600,
      "theoretical_loss": 4.3254565666788745,
      "tokens_seen": 211353600
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004727081243731194,
      "loss": 3.3646,
      "theoretical_loss": 4.3254565666788745,
      "tokens_seen": 211353600
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047269809428284853,
      "loss": 3.1249,
      "theoretical_loss": 4.325290200845629,
      "tokens_seen": 211419136
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047268806419257777,
      "loss": 2.903,
      "theoretical_loss": 4.3251239010092934,
      "tokens_seen": 211484672
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047267803410230695,
      "loss": 2.9262,
      "theoretical_loss": 4.324957667123249,
      "tokens_seen": 211550208
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047266800401203613,
      "loss": 3.339,
      "theoretical_loss": 4.32479149914092,
      "tokens_seen": 211615744
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004726579739217653,
      "loss": 3.2618,
      "theoretical_loss": 4.324625397015783,
      "tokens_seen": 211681280
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004726479438314945,
      "loss": 3.1819,
      "theoretical_loss": 4.3244593607013595,
      "tokens_seen": 211746816
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047263791374122367,
      "loss": 2.9808,
      "theoretical_loss": 4.324293390151218,
      "tokens_seen": 211812352
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004726278836509529,
      "loss": 3.3896,
      "theoretical_loss": 4.324127485318975,
      "tokens_seen": 211877888
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047261785356068203,
      "loss": 3.5236,
      "theoretical_loss": 4.323961646158294,
      "tokens_seen": 211943424
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047260782347041127,
      "loss": 3.0748,
      "theoretical_loss": 4.323795872622884,
      "tokens_seen": 212008960
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004725977933801404,
      "loss": 3.2701,
      "theoretical_loss": 4.323630164666502,
      "tokens_seen": 212074496
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047258776328986963,
      "loss": 3.0506,
      "theoretical_loss": 4.323464522242954,
      "tokens_seen": 212140032
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004725777331995988,
      "loss": 3.3146,
      "theoretical_loss": 4.323298945306089,
      "tokens_seen": 212205568
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000472567703109328,
      "loss": 3.3863,
      "theoretical_loss": 4.3231334338098035,
      "tokens_seen": 212271104
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004725576730190572,
      "loss": 3.5083,
      "theoretical_loss": 4.322967987708043,
      "tokens_seen": 212336640
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047254764292878636,
      "loss": 3.0861,
      "theoretical_loss": 4.322802606954798,
      "tokens_seen": 212402176
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047253761283851554,
      "loss": 3.2103,
      "theoretical_loss": 4.322637291504106,
      "tokens_seen": 212467712
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047252758274824477,
      "loss": 3.2339,
      "theoretical_loss": 4.32247204131005,
      "tokens_seen": 212533248
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004725175526579739,
      "loss": 3.3825,
      "theoretical_loss": 4.322306856326761,
      "tokens_seen": 212598784
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047250752256770313,
      "loss": 3.2205,
      "theoretical_loss": 4.322141736508415,
      "tokens_seen": 212664320
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004724974924774323,
      "loss": 3.2196,
      "theoretical_loss": 4.321976681809236,
      "tokens_seen": 212729856
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004724874623871615,
      "loss": 3.1189,
      "theoretical_loss": 4.321811692183491,
      "tokens_seen": 212795392
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004724774322968907,
      "loss": 3.2039,
      "theoretical_loss": 4.321646767585497,
      "tokens_seen": 212860928
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047246740220661986,
      "loss": 3.1653,
      "theoretical_loss": 4.3214819079696145,
      "tokens_seen": 212926464
    },
    {
      "epoch": 0.06,
      "objective/train/docs_used": 280599,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.3949620723724365,
      "objective/train/theoretical_loss": 4.321317113290252,
      "objective/train/tokens_used": 233452000,
      "theoretical_loss": 4.321317113290252,
      "tokens_seen": 212992000
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047245737211634904,
      "loss": 3.3408,
      "theoretical_loss": 4.321317113290252,
      "tokens_seen": 212992000
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004724473420260783,
      "loss": 3.1235,
      "theoretical_loss": 4.321152383501863,
      "tokens_seen": 213057536
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004724373119358074,
      "loss": 3.2593,
      "theoretical_loss": 4.320987718558945,
      "tokens_seen": 213123072
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047242728184553664,
      "loss": 3.3497,
      "theoretical_loss": 4.320823118416046,
      "tokens_seen": 213188608
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047241725175526576,
      "loss": 3.5142,
      "theoretical_loss": 4.320658583027755,
      "tokens_seen": 213254144
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000472407221664995,
      "loss": 3.4399,
      "theoretical_loss": 4.32049411234871,
      "tokens_seen": 213319680
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004723971915747242,
      "loss": 3.2726,
      "theoretical_loss": 4.3203297063335935,
      "tokens_seen": 213385216
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047238716148445336,
      "loss": 3.2783,
      "theoretical_loss": 4.320165364937134,
      "tokens_seen": 213450752
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047237713139418254,
      "loss": 3.2021,
      "theoretical_loss": 4.320001088114105,
      "tokens_seen": 213516288
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004723671013039117,
      "loss": 3.2311,
      "theoretical_loss": 4.319836875819325,
      "tokens_seen": 213581824
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004723570712136409,
      "loss": 2.9671,
      "theoretical_loss": 4.31967272800766,
      "tokens_seen": 213647360
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047234704112337014,
      "loss": 3.1442,
      "theoretical_loss": 4.319508644634021,
      "tokens_seen": 213712896
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047233701103309927,
      "loss": 2.8623,
      "theoretical_loss": 4.319344625653361,
      "tokens_seen": 213778432
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004723269809428285,
      "loss": 3.1813,
      "theoretical_loss": 4.319180671020684,
      "tokens_seen": 213843968
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004723169508525577,
      "loss": 3.1773,
      "theoretical_loss": 4.319016780691033,
      "tokens_seen": 213909504
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047230692076228686,
      "loss": 3.2263,
      "theoretical_loss": 4.318852954619501,
      "tokens_seen": 213975040
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047229689067201605,
      "loss": 3.2179,
      "theoretical_loss": 4.318689192761225,
      "tokens_seen": 214040576
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004722868605817452,
      "loss": 3.3982,
      "theoretical_loss": 4.318525495071385,
      "tokens_seen": 214106112
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004722768304914744,
      "loss": 3.2654,
      "theoretical_loss": 4.318361861505207,
      "tokens_seen": 214171648
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047226680040120364,
      "loss": 3.0422,
      "theoretical_loss": 4.318198292017964,
      "tokens_seen": 214237184
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047225677031093277,
      "loss": 3.2805,
      "theoretical_loss": 4.318034786564971,
      "tokens_seen": 214302720
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000472246740220662,
      "loss": 3.115,
      "theoretical_loss": 4.31787134510159,
      "tokens_seen": 214368256
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047223671013039113,
      "loss": 3.4185,
      "theoretical_loss": 4.3177079675832255,
      "tokens_seen": 214433792
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047222668004012037,
      "loss": 3.2431,
      "theoretical_loss": 4.317544653965329,
      "tokens_seen": 214499328
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004722166499498496,
      "loss": 3.2717,
      "theoretical_loss": 4.3173814042033944,
      "tokens_seen": 214564864
    },
    {
      "epoch": 0.07,
      "objective/train/docs_used": 281081,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.1554346084594727,
      "objective/train/theoretical_loss": 4.317218218252963,
      "objective/train/tokens_used": 235090400,
      "theoretical_loss": 4.317218218252963,
      "tokens_seen": 214630400
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047220661985957873,
      "loss": 3.2487,
      "theoretical_loss": 4.317218218252963,
      "tokens_seen": 214630400
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047219658976930797,
      "loss": 3.2627,
      "theoretical_loss": 4.317055096069618,
      "tokens_seen": 214695936
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047218655967903715,
      "loss": 3.132,
      "theoretical_loss": 4.316892037608987,
      "tokens_seen": 214761472
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047217652958876633,
      "loss": 3.2713,
      "theoretical_loss": 4.316729042826745,
      "tokens_seen": 214827008
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004721664994984955,
      "loss": 3.2406,
      "theoretical_loss": 4.316566111678609,
      "tokens_seen": 214892544
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004721564694082247,
      "loss": 3.3419,
      "theoretical_loss": 4.316403244120339,
      "tokens_seen": 214958080
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047214643931795387,
      "loss": 3.0689,
      "theoretical_loss": 4.3162404401077445,
      "tokens_seen": 215023616
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004721364092276831,
      "loss": 3.3983,
      "theoretical_loss": 4.316077699596671,
      "tokens_seen": 215089152
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047212637913741223,
      "loss": 3.0837,
      "theoretical_loss": 4.315915022543016,
      "tokens_seen": 215154688
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047211634904714147,
      "loss": 3.1345,
      "theoretical_loss": 4.315752408902716,
      "tokens_seen": 215220224
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004721063189568706,
      "loss": 3.2947,
      "theoretical_loss": 4.315589858631755,
      "tokens_seen": 215285760
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047209628886659983,
      "loss": 3.2647,
      "theoretical_loss": 4.315427371686157,
      "tokens_seen": 215351296
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000472086258776329,
      "loss": 3.1737,
      "theoretical_loss": 4.315264948021994,
      "tokens_seen": 215416832
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004720762286860582,
      "loss": 3.1847,
      "theoretical_loss": 4.315102587595379,
      "tokens_seen": 215482368
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004720661985957874,
      "loss": 3.0922,
      "theoretical_loss": 4.31494029036247,
      "tokens_seen": 215547904
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047205616850551656,
      "loss": 3.287,
      "theoretical_loss": 4.314778056279468,
      "tokens_seen": 215613440
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047204613841524574,
      "loss": 2.8694,
      "theoretical_loss": 4.314615885302619,
      "tokens_seen": 215678976
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047203610832497497,
      "loss": 3.373,
      "theoretical_loss": 4.314453777388209,
      "tokens_seen": 215744512
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004720260782347041,
      "loss": 3.2566,
      "theoretical_loss": 4.314291732492573,
      "tokens_seen": 215810048
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047201604814443333,
      "loss": 2.9877,
      "theoretical_loss": 4.314129750572087,
      "tokens_seen": 215875584
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004720060180541625,
      "loss": 3.3091,
      "theoretical_loss": 4.3139678315831675,
      "tokens_seen": 215941120
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004719959879638917,
      "loss": 3.0783,
      "theoretical_loss": 4.313805975482278,
      "tokens_seen": 216006656
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004719859578736209,
      "loss": 3.1844,
      "theoretical_loss": 4.313644182225926,
      "tokens_seen": 216072192
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047197592778335006,
      "loss": 3.2515,
      "theoretical_loss": 4.313482451770659,
      "tokens_seen": 216137728
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047196589769307924,
      "loss": 2.9306,
      "theoretical_loss": 4.313320784073069,
      "tokens_seen": 216203264
    },
    {
      "epoch": 0.07,
      "objective/train/docs_used": 282379,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.2120323181152344,
      "objective/train/theoretical_loss": 4.3131591790897925,
      "objective/train/tokens_used": 236728800,
      "theoretical_loss": 4.3131591790897925,
      "tokens_seen": 216268800
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004719558676028085,
      "loss": 3.1633,
      "theoretical_loss": 4.3131591790897925,
      "tokens_seen": 216268800
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004719458375125376,
      "loss": 3.3385,
      "theoretical_loss": 4.3129976367775065,
      "tokens_seen": 216334336
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047193580742226684,
      "loss": 3.1109,
      "theoretical_loss": 4.312836157092934,
      "tokens_seen": 216399872
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047192577733199596,
      "loss": 3.3772,
      "theoretical_loss": 4.312674739992839,
      "tokens_seen": 216465408
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004719157472417252,
      "loss": 3.3361,
      "theoretical_loss": 4.31251338543403,
      "tokens_seen": 216530944
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004719057171514544,
      "loss": 3.3251,
      "theoretical_loss": 4.312352093373354,
      "tokens_seen": 216596480
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047189568706118356,
      "loss": 3.0503,
      "theoretical_loss": 4.312190863767708,
      "tokens_seen": 216662016
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047188565697091274,
      "loss": 3.0291,
      "theoretical_loss": 4.312029696574027,
      "tokens_seen": 216727552
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004718756268806419,
      "loss": 3.0042,
      "theoretical_loss": 4.311868591749287,
      "tokens_seen": 216793088
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004718655967903711,
      "loss": 3.0656,
      "theoretical_loss": 4.311707549250514,
      "tokens_seen": 216858624
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047185556670010034,
      "loss": 3.1287,
      "theoretical_loss": 4.311546569034767,
      "tokens_seen": 216924160
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047184553660982947,
      "loss": 3.1405,
      "theoretical_loss": 4.311385651059155,
      "tokens_seen": 216989696
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004718355065195587,
      "loss": 3.425,
      "theoretical_loss": 4.311224795280825,
      "tokens_seen": 217055232
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004718254764292879,
      "loss": 3.282,
      "theoretical_loss": 4.3110640016569715,
      "tokens_seen": 217120768
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047181544633901706,
      "loss": 3.4067,
      "theoretical_loss": 4.310903270144825,
      "tokens_seen": 217186304
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047180541624874625,
      "loss": 3.1557,
      "theoretical_loss": 4.310742600701664,
      "tokens_seen": 217251840
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004717953861584754,
      "loss": 3.1474,
      "theoretical_loss": 4.310581993284805,
      "tokens_seen": 217317376
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004717853560682046,
      "loss": 3.2756,
      "theoretical_loss": 4.310421447851609,
      "tokens_seen": 217382912
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047177532597793384,
      "loss": 3.4118,
      "theoretical_loss": 4.310260964359479,
      "tokens_seen": 217448448
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047176529588766297,
      "loss": 3.2428,
      "theoretical_loss": 4.310100542765858,
      "tokens_seen": 217513984
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004717552657973922,
      "loss": 3.2141,
      "theoretical_loss": 4.309940183028236,
      "tokens_seen": 217579520
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047174523570712133,
      "loss": 3.3118,
      "theoretical_loss": 4.309779885104139,
      "tokens_seen": 217645056
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047173520561685057,
      "loss": 3.276,
      "theoretical_loss": 4.309619648951139,
      "tokens_seen": 217710592
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047172517552657975,
      "loss": 2.8709,
      "theoretical_loss": 4.3094594745268475,
      "tokens_seen": 217776128
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047171514543630893,
      "loss": 3.3411,
      "theoretical_loss": 4.30929936178892,
      "tokens_seen": 217841664
    },
    {
      "epoch": 0.07,
      "objective/train/docs_used": 283045,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.1448984146118164,
      "objective/train/theoretical_loss": 4.309139310695053,
      "objective/train/tokens_used": 238367200,
      "theoretical_loss": 4.309139310695053,
      "tokens_seen": 217907200
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004717051153460381,
      "loss": 3.247,
      "theoretical_loss": 4.309139310695053,
      "tokens_seen": 217907200
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047169508525576735,
      "loss": 2.9108,
      "theoretical_loss": 4.308979321202983,
      "tokens_seen": 217972736
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004716850551654965,
      "loss": 3.2046,
      "theoretical_loss": 4.308819393270491,
      "tokens_seen": 218038272
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004716750250752257,
      "loss": 3.076,
      "theoretical_loss": 4.308659526855396,
      "tokens_seen": 218103808
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047166499498495484,
      "loss": 3.245,
      "theoretical_loss": 4.308499721915563,
      "tokens_seen": 218169344
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047165496489468407,
      "loss": 3.2956,
      "theoretical_loss": 4.308339978408897,
      "tokens_seen": 218234880
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047164493480441325,
      "loss": 3.0581,
      "theoretical_loss": 4.308180296293341,
      "tokens_seen": 218300416
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047163490471414243,
      "loss": 3.1606,
      "theoretical_loss": 4.308020675526883,
      "tokens_seen": 218365952
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004716248746238716,
      "loss": 3.2829,
      "theoretical_loss": 4.307861116067554,
      "tokens_seen": 218431488
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004716148445336008,
      "loss": 3.3331,
      "theoretical_loss": 4.30770161787342,
      "tokens_seen": 218497024
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047160481444333,
      "loss": 3.2521,
      "theoretical_loss": 4.307542180902594,
      "tokens_seen": 218562560
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004715947843530592,
      "loss": 3.0335,
      "theoretical_loss": 4.307382805113228,
      "tokens_seen": 218628096
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047158475426278834,
      "loss": 2.9783,
      "theoretical_loss": 4.307223490463516,
      "tokens_seen": 218693632
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004715747241725176,
      "loss": 3.1517,
      "theoretical_loss": 4.307064236911692,
      "tokens_seen": 218759168
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004715646940822467,
      "loss": 3.002,
      "theoretical_loss": 4.30690504441603,
      "tokens_seen": 218824704
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047155466399197594,
      "loss": 3.13,
      "theoretical_loss": 4.306745912934849,
      "tokens_seen": 218890240
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004715446339017051,
      "loss": 3.2757,
      "theoretical_loss": 4.306586842426504,
      "tokens_seen": 218955776
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004715346038114343,
      "loss": 3.1729,
      "theoretical_loss": 4.306427832849394,
      "tokens_seen": 219021312
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004715245737211635,
      "loss": 2.8273,
      "theoretical_loss": 4.306268884161959,
      "tokens_seen": 219086848
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004715145436308927,
      "loss": 3.0639,
      "theoretical_loss": 4.306109996322679,
      "tokens_seen": 219152384
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047150451354062184,
      "loss": 3.2826,
      "theoretical_loss": 4.305951169290073,
      "tokens_seen": 219217920
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004714944834503511,
      "loss": 3.2331,
      "theoretical_loss": 4.305792403022703,
      "tokens_seen": 219283456
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004714844533600802,
      "loss": 3.4157,
      "theoretical_loss": 4.305633697479171,
      "tokens_seen": 219348992
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047147442326980944,
      "loss": 3.2171,
      "theoretical_loss": 4.305475052618119,
      "tokens_seen": 219414528
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004714643931795387,
      "loss": 3.271,
      "theoretical_loss": 4.30531646839823,
      "tokens_seen": 219480064
    },
    {
      "epoch": 0.07,
      "objective/train/docs_used": 284124,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.518115997314453,
      "objective/train/theoretical_loss": 4.305157944778228,
      "objective/train/tokens_used": 240005600,
      "theoretical_loss": 4.305157944778228,
      "tokens_seen": 219545600
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004714543630892678,
      "loss": 3.2974,
      "theoretical_loss": 4.305157944778228,
      "tokens_seen": 219545600
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047144433299899704,
      "loss": 3.1396,
      "theoretical_loss": 4.304999481716876,
      "tokens_seen": 219611136
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047143430290872616,
      "loss": 3.3291,
      "theoretical_loss": 4.304841079172979,
      "tokens_seen": 219676672
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004714242728184554,
      "loss": 3.3637,
      "theoretical_loss": 4.30468273710538,
      "tokens_seen": 219742208
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004714142427281846,
      "loss": 3.22,
      "theoretical_loss": 4.304524455472965,
      "tokens_seen": 219807744
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047140421263791376,
      "loss": 3.2553,
      "theoretical_loss": 4.304366234234659,
      "tokens_seen": 219873280
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047139418254764294,
      "loss": 3.2947,
      "theoretical_loss": 4.304208073349426,
      "tokens_seen": 219938816
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004713841524573721,
      "loss": 3.2111,
      "theoretical_loss": 4.304049972776271,
      "tokens_seen": 220004352
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004713741223671013,
      "loss": 3.2684,
      "theoretical_loss": 4.30389193247424,
      "tokens_seen": 220069888
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047136409227683054,
      "loss": 3.1529,
      "theoretical_loss": 4.303733952402419,
      "tokens_seen": 220135424
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047135406218655967,
      "loss": 3.1474,
      "theoretical_loss": 4.303576032519931,
      "tokens_seen": 220200960
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004713440320962889,
      "loss": 3.0742,
      "theoretical_loss": 4.303418172785943,
      "tokens_seen": 220266496
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004713340020060181,
      "loss": 3.1944,
      "theoretical_loss": 4.303260373159659,
      "tokens_seen": 220332032
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047132397191574726,
      "loss": 3.0874,
      "theoretical_loss": 4.303102633600322,
      "tokens_seen": 220397568
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047131394182547645,
      "loss": 3.0249,
      "theoretical_loss": 4.30294495406722,
      "tokens_seen": 220463104
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047130391173520563,
      "loss": 3.434,
      "theoretical_loss": 4.3027873345196745,
      "tokens_seen": 220528640
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004712938816449348,
      "loss": 3.4907,
      "theoretical_loss": 4.302629774917049,
      "tokens_seen": 220594176
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047128385155466404,
      "loss": 3.1007,
      "theoretical_loss": 4.302472275218748,
      "tokens_seen": 220659712
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047127382146439317,
      "loss": 3.3572,
      "theoretical_loss": 4.302314835384214,
      "tokens_seen": 220725248
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004712637913741224,
      "loss": 3.2629,
      "theoretical_loss": 4.30215745537293,
      "tokens_seen": 220790784
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047125376128385153,
      "loss": 3.0362,
      "theoretical_loss": 4.302000135144416,
      "tokens_seen": 220856320
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047124373119358077,
      "loss": 3.1044,
      "theoretical_loss": 4.301842874658235,
      "tokens_seen": 220921856
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047123370110330995,
      "loss": 3.2226,
      "theoretical_loss": 4.301685673873987,
      "tokens_seen": 220987392
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047122367101303913,
      "loss": 3.2153,
      "theoretical_loss": 4.301528532751312,
      "tokens_seen": 221052928
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004712136409227683,
      "loss": 3.5554,
      "theoretical_loss": 4.301371451249888,
      "tokens_seen": 221118464
    },
    {
      "epoch": 0.07,
      "objective/train/docs_used": 284674,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.90860652923584,
      "objective/train/theoretical_loss": 4.301214429329433,
      "objective/train/tokens_used": 241644000,
      "theoretical_loss": 4.301214429329433,
      "tokens_seen": 221184000
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047120361083249755,
      "loss": 3.2309,
      "theoretical_loss": 4.301214429329433,
      "tokens_seen": 221184000
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004711935807422267,
      "loss": 3.2685,
      "theoretical_loss": 4.301057466949707,
      "tokens_seen": 221249536
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004711835506519559,
      "loss": 3.1091,
      "theoretical_loss": 4.300900564070504,
      "tokens_seen": 221315072
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047117352056168504,
      "loss": 3.0911,
      "theoretical_loss": 4.30074372065166,
      "tokens_seen": 221380608
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047116349047141427,
      "loss": 3.3026,
      "theoretical_loss": 4.300586936653049,
      "tokens_seen": 221446144
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047115346038114345,
      "loss": 3.2759,
      "theoretical_loss": 4.300430212034587,
      "tokens_seen": 221511680
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047114343029087263,
      "loss": 3.1257,
      "theoretical_loss": 4.300273546756223,
      "tokens_seen": 221577216
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004711334002006018,
      "loss": 3.1675,
      "theoretical_loss": 4.300116940777951,
      "tokens_seen": 221642752
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000471123370110331,
      "loss": 2.9268,
      "theoretical_loss": 4.299960394059799,
      "tokens_seen": 221708288
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004711133400200602,
      "loss": 3.1183,
      "theoretical_loss": 4.299803906561835,
      "tokens_seen": 221773824
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004711033099297894,
      "loss": 3.3301,
      "theoretical_loss": 4.29964747824417,
      "tokens_seen": 221839360
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047109327983951854,
      "loss": 3.0193,
      "theoretical_loss": 4.299491109066947,
      "tokens_seen": 221904896
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004710832497492478,
      "loss": 3.4411,
      "theoretical_loss": 4.299334798990351,
      "tokens_seen": 221970432
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004710732196589769,
      "loss": 3.2296,
      "theoretical_loss": 4.2991785479746065,
      "tokens_seen": 222035968
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047106318956870614,
      "loss": 3.0854,
      "theoretical_loss": 4.299022355979974,
      "tokens_seen": 222101504
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004710531594784353,
      "loss": 3.2255,
      "theoretical_loss": 4.298866222966755,
      "tokens_seen": 222167040
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004710431293881645,
      "loss": 3.2013,
      "theoretical_loss": 4.298710148895286,
      "tokens_seen": 222232576
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004710330992978937,
      "loss": 3.2448,
      "theoretical_loss": 4.298554133725946,
      "tokens_seen": 222298112
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004710230692076229,
      "loss": 3.1119,
      "theoretical_loss": 4.298398177419149,
      "tokens_seen": 222363648
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047101303911735204,
      "loss": 3.1882,
      "theoretical_loss": 4.298242279935349,
      "tokens_seen": 222429184
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004710030090270813,
      "loss": 3.0531,
      "theoretical_loss": 4.2980864412350375,
      "tokens_seen": 222494720
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004709929789368104,
      "loss": 3.0501,
      "theoretical_loss": 4.297930661278745,
      "tokens_seen": 222560256
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047098294884653964,
      "loss": 3.1275,
      "theoretical_loss": 4.297774940027038,
      "tokens_seen": 222625792
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004709729187562688,
      "loss": 3.3449,
      "theoretical_loss": 4.297619277440523,
      "tokens_seen": 222691328
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000470962888665998,
      "loss": 3.2312,
      "theoretical_loss": 4.297463673479846,
      "tokens_seen": 222756864
    },
    {
      "epoch": 0.07,
      "objective/train/docs_used": 286126,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6270840167999268,
      "objective/train/theoretical_loss": 4.297308128105687,
      "objective/train/tokens_used": 243282400,
      "theoretical_loss": 4.297308128105687,
      "tokens_seen": 222822400
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004709528585757272,
      "loss": 3.1513,
      "theoretical_loss": 4.297308128105687,
      "tokens_seen": 222822400
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047094282848545636,
      "loss": 3.1221,
      "theoretical_loss": 4.297152641278767,
      "tokens_seen": 222887936
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047093279839518555,
      "loss": 3.2143,
      "theoretical_loss": 4.296997212959842,
      "tokens_seen": 222953472
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004709227683049148,
      "loss": 3.0636,
      "theoretical_loss": 4.296841843109711,
      "tokens_seen": 223019008
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004709127382146439,
      "loss": 3.2735,
      "theoretical_loss": 4.296686531689204,
      "tokens_seen": 223084544
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047090270812437314,
      "loss": 2.9135,
      "theoretical_loss": 4.296531278659193,
      "tokens_seen": 223150080
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047089267803410227,
      "loss": 3.1243,
      "theoretical_loss": 4.296376083980589,
      "tokens_seen": 223215616
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004708826479438315,
      "loss": 3.0309,
      "theoretical_loss": 4.296220947614337,
      "tokens_seen": 223281152
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004708726178535607,
      "loss": 3.154,
      "theoretical_loss": 4.296065869521421,
      "tokens_seen": 223346688
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047086258776328987,
      "loss": 2.8156,
      "theoretical_loss": 4.295910849662862,
      "tokens_seen": 223412224
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047085255767301905,
      "loss": 3.0104,
      "theoretical_loss": 4.2957558879997215,
      "tokens_seen": 223477760
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004708425275827483,
      "loss": 3.1348,
      "theoretical_loss": 4.295600984493093,
      "tokens_seen": 223543296
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004708324974924774,
      "loss": 3.047,
      "theoretical_loss": 4.295446139104112,
      "tokens_seen": 223608832
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047082246740220665,
      "loss": 3.3838,
      "theoretical_loss": 4.295291351793951,
      "tokens_seen": 223674368
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004708124373119358,
      "loss": 3.0171,
      "theoretical_loss": 4.295136622523817,
      "tokens_seen": 223739904
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000470802407221665,
      "loss": 3.3862,
      "theoretical_loss": 4.294981951254956,
      "tokens_seen": 223805440
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004707923771313942,
      "loss": 3.1766,
      "theoretical_loss": 4.294827337948651,
      "tokens_seen": 223870976
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047078234704112337,
      "loss": 3.0689,
      "theoretical_loss": 4.294672782566224,
      "tokens_seen": 223936512
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047077231695085255,
      "loss": 3.1193,
      "theoretical_loss": 4.29451828506903,
      "tokens_seen": 224002048
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047076228686058173,
      "loss": 3.4053,
      "theoretical_loss": 4.294363845418465,
      "tokens_seen": 224067584
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004707522567703109,
      "loss": 3.1876,
      "theoretical_loss": 4.29420946357596,
      "tokens_seen": 224133120
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047074222668004015,
      "loss": 3.2133,
      "theoretical_loss": 4.294055139502985,
      "tokens_seen": 224198656
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004707321965897693,
      "loss": 3.2064,
      "theoretical_loss": 4.293900873161043,
      "tokens_seen": 224264192
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004707221664994985,
      "loss": 3.295,
      "theoretical_loss": 4.293746664511678,
      "tokens_seen": 224329728
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047071213640922775,
      "loss": 2.9466,
      "theoretical_loss": 4.293592513516469,
      "tokens_seen": 224395264
    },
    {
      "epoch": 0.07,
      "objective/train/docs_used": 286836,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.5235793590545654,
      "objective/train/theoretical_loss": 4.293438420137031,
      "objective/train/tokens_used": 244920800,
      "theoretical_loss": 4.293438420137031,
      "tokens_seen": 224460800
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004707021063189569,
      "loss": 3.2967,
      "theoretical_loss": 4.293438420137031,
      "tokens_seen": 224460800
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004706920762286861,
      "loss": 3.1972,
      "theoretical_loss": 4.293284384335017,
      "tokens_seen": 224526336
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047068204613841524,
      "loss": 3.2584,
      "theoretical_loss": 4.293130406072118,
      "tokens_seen": 224591872
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047067201604814447,
      "loss": 3.1142,
      "theoretical_loss": 4.292976485310057,
      "tokens_seen": 224657408
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047066198595787365,
      "loss": 3.1867,
      "theoretical_loss": 4.2928226220106005,
      "tokens_seen": 224722944
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047065195586760283,
      "loss": 3.1763,
      "theoretical_loss": 4.292668816135545,
      "tokens_seen": 224788480
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000470641925777332,
      "loss": 3.5007,
      "theoretical_loss": 4.292515067646727,
      "tokens_seen": 224854016
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004706318956870612,
      "loss": 2.8707,
      "theoretical_loss": 4.29236137650602,
      "tokens_seen": 224919552
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004706218655967904,
      "loss": 3.2017,
      "theoretical_loss": 4.2922077426753305,
      "tokens_seen": 224985088
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004706118355065196,
      "loss": 3.0439,
      "theoretical_loss": 4.292054166116605,
      "tokens_seen": 225050624
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047060180541624874,
      "loss": 3.2209,
      "theoretical_loss": 4.291900646791825,
      "tokens_seen": 225116160
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000470591775325978,
      "loss": 3.2066,
      "theoretical_loss": 4.2917471846630075,
      "tokens_seen": 225181696
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004705817452357071,
      "loss": 2.9519,
      "theoretical_loss": 4.291593779692207,
      "tokens_seen": 225247232
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047057171514543634,
      "loss": 3.1154,
      "theoretical_loss": 4.291440431841513,
      "tokens_seen": 225312768
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004705616850551655,
      "loss": 3.181,
      "theoretical_loss": 4.291287141073053,
      "tokens_seen": 225378304
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004705516549648947,
      "loss": 2.9915,
      "theoretical_loss": 4.291133907348989,
      "tokens_seen": 225443840
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004705416248746239,
      "loss": 3.1199,
      "theoretical_loss": 4.29098073063152,
      "tokens_seen": 225509376
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004705315947843531,
      "loss": 3.3002,
      "theoretical_loss": 4.29082761088288,
      "tokens_seen": 225574912
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047052156469408224,
      "loss": 2.7099,
      "theoretical_loss": 4.290674548065338,
      "tokens_seen": 225640448
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004705115346038115,
      "loss": 3.1867,
      "theoretical_loss": 4.290521542141203,
      "tokens_seen": 225705984
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004705015045135406,
      "loss": 3.0885,
      "theoretical_loss": 4.290368593072817,
      "tokens_seen": 225771520
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047049147442326984,
      "loss": 3.1099,
      "theoretical_loss": 4.290215700822556,
      "tokens_seen": 225837056
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000470481444332999,
      "loss": 3.1279,
      "theoretical_loss": 4.290062865352837,
      "tokens_seen": 225902592
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004704714142427282,
      "loss": 3.0228,
      "theoretical_loss": 4.289910086626108,
      "tokens_seen": 225968128
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004704613841524574,
      "loss": 3.3652,
      "theoretical_loss": 4.289757364604855,
      "tokens_seen": 226033664
    },
    {
      "epoch": 0.07,
      "objective/train/docs_used": 288143,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.2444591522216797,
      "objective/train/theoretical_loss": 4.2896046992515995,
      "objective/train/tokens_used": 246559200,
      "theoretical_loss": 4.2896046992515995,
      "tokens_seen": 226099200
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047045135406218656,
      "loss": 3.2585,
      "theoretical_loss": 4.2896046992515995,
      "tokens_seen": 226099200
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047044132397191575,
      "loss": 2.841,
      "theoretical_loss": 4.289452090528897,
      "tokens_seen": 226164736
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000470431293881645,
      "loss": 3.1428,
      "theoretical_loss": 4.289299538399341,
      "tokens_seen": 226230272
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004704212637913741,
      "loss": 3.229,
      "theoretical_loss": 4.28914704282556,
      "tokens_seen": 226295808
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047041123370110334,
      "loss": 3.003,
      "theoretical_loss": 4.288994603770215,
      "tokens_seen": 226361344
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047040120361083247,
      "loss": 3.1008,
      "theoretical_loss": 4.288842221196007,
      "tokens_seen": 226426880
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004703911735205617,
      "loss": 3.0763,
      "theoretical_loss": 4.28868989506567,
      "tokens_seen": 226492416
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004703811434302909,
      "loss": 3.1798,
      "theoretical_loss": 4.288537625341974,
      "tokens_seen": 226557952
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047037111334002007,
      "loss": 3.5773,
      "theoretical_loss": 4.288385411987722,
      "tokens_seen": 226623488
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047036108324974925,
      "loss": 3.156,
      "theoretical_loss": 4.288233254965755,
      "tokens_seen": 226689024
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004703510531594785,
      "loss": 2.9742,
      "theoretical_loss": 4.2880811542389505,
      "tokens_seen": 226754560
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004703410230692076,
      "loss": 2.9553,
      "theoretical_loss": 4.287929109770217,
      "tokens_seen": 226820096
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047033099297893685,
      "loss": 3.1251,
      "theoretical_loss": 4.287777121522501,
      "tokens_seen": 226885632
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000470320962888666,
      "loss": 3.0942,
      "theoretical_loss": 4.287625189458781,
      "tokens_seen": 226951168
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004703109327983952,
      "loss": 3.2267,
      "theoretical_loss": 4.287473313542077,
      "tokens_seen": 227016704
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004703009027081244,
      "loss": 3.0715,
      "theoretical_loss": 4.287321493735438,
      "tokens_seen": 227082240
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047029087261785357,
      "loss": 3.2224,
      "theoretical_loss": 4.287169730001949,
      "tokens_seen": 227147776
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047028084252758275,
      "loss": 3.1082,
      "theoretical_loss": 4.287018022304733,
      "tokens_seen": 227213312
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047027081243731193,
      "loss": 3.2017,
      "theoretical_loss": 4.286866370606943,
      "tokens_seen": 227278848
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004702607823470411,
      "loss": 2.8765,
      "theoretical_loss": 4.286714774871772,
      "tokens_seen": 227344384
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047025075225677035,
      "loss": 2.7387,
      "theoretical_loss": 4.286563235062444,
      "tokens_seen": 227409920
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004702407221664995,
      "loss": 2.8634,
      "theoretical_loss": 4.28641175114222,
      "tokens_seen": 227475456
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004702306920762287,
      "loss": 2.7575,
      "theoretical_loss": 4.286260323074394,
      "tokens_seen": 227540992
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047022066198595784,
      "loss": 3.0389,
      "theoretical_loss": 4.286108950822296,
      "tokens_seen": 227606528
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004702106318956871,
      "loss": 3.0842,
      "theoretical_loss": 4.285957634349289,
      "tokens_seen": 227672064
    },
    {
      "epoch": 0.07,
      "objective/train/docs_used": 288537,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.2806308269500732,
      "objective/train/theoretical_loss": 4.285806373618774,
      "objective/train/tokens_used": 248197600,
      "theoretical_loss": 4.285806373618774,
      "tokens_seen": 227737600
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047020060180541626,
      "loss": 3.402,
      "theoretical_loss": 4.285806373618774,
      "tokens_seen": 227737600
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047019057171514544,
      "loss": 3.2526,
      "theoretical_loss": 4.285655168594182,
      "tokens_seen": 227803136
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004701805416248746,
      "loss": 3.2338,
      "theoretical_loss": 4.285504019238982,
      "tokens_seen": 227868672
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047017051153460385,
      "loss": 3.1311,
      "theoretical_loss": 4.285352925516676,
      "tokens_seen": 227934208
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000470160481444333,
      "loss": 3.2231,
      "theoretical_loss": 4.2852018873908,
      "tokens_seen": 227999744
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004701504513540622,
      "loss": 3.1156,
      "theoretical_loss": 4.285050904824925,
      "tokens_seen": 228065280
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047014042126379134,
      "loss": 3.2022,
      "theoretical_loss": 4.284899977782658,
      "tokens_seen": 228130816
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004701303911735206,
      "loss": 3.1321,
      "theoretical_loss": 4.284749106227636,
      "tokens_seen": 228196352
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047012036108324976,
      "loss": 3.1009,
      "theoretical_loss": 4.284598290123535,
      "tokens_seen": 228261888
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047011033099297894,
      "loss": 3.2566,
      "theoretical_loss": 4.284447529434061,
      "tokens_seen": 228327424
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004701003009027081,
      "loss": 3.1493,
      "theoretical_loss": 4.284296824122959,
      "tokens_seen": 228392960
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004700902708124373,
      "loss": 3.1941,
      "theoretical_loss": 4.284146174154003,
      "tokens_seen": 228458496
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004700802407221665,
      "loss": 2.9355,
      "theoretical_loss": 4.283995579491004,
      "tokens_seen": 228524032
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004700702106318957,
      "loss": 3.098,
      "theoretical_loss": 4.283845040097807,
      "tokens_seen": 228589568
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047006018054162484,
      "loss": 3.013,
      "theoretical_loss": 4.28369455593829,
      "tokens_seen": 228655104
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004700501504513541,
      "loss": 3.2876,
      "theoretical_loss": 4.2835441269763646,
      "tokens_seen": 228720640
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047004012036108326,
      "loss": 3.1562,
      "theoretical_loss": 4.283393753175979,
      "tokens_seen": 228786176
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047003009027081244,
      "loss": 3.1474,
      "theoretical_loss": 4.283243434501112,
      "tokens_seen": 228851712
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004700200601805416,
      "loss": 3.1408,
      "theoretical_loss": 4.283093170915778,
      "tokens_seen": 228917248
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004700100300902708,
      "loss": 3.0895,
      "theoretical_loss": 4.282942962384023,
      "tokens_seen": 228982784
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047,
      "loss": 3.1473,
      "theoretical_loss": 4.282792808869932,
      "tokens_seen": 229048320
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004699899699097292,
      "loss": 3.2697,
      "theoretical_loss": 4.282642710337618,
      "tokens_seen": 229113856
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046997993981945835,
      "loss": 2.9681,
      "theoretical_loss": 4.28249266675123,
      "tokens_seen": 229179392
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004699699097291876,
      "loss": 3.0361,
      "theoretical_loss": 4.282342678074951,
      "tokens_seen": 229244928
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046995987963891676,
      "loss": 3.0598,
      "theoretical_loss": 4.2821927442729955,
      "tokens_seen": 229310464
    },
    {
      "epoch": 0.07,
      "objective/train/docs_used": 289830,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.521411418914795,
      "objective/train/theoretical_loss": 4.282042865309616,
      "objective/train/tokens_used": 249836000,
      "theoretical_loss": 4.282042865309616,
      "tokens_seen": 229376000
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046994984954864595,
      "loss": 3.1125,
      "theoretical_loss": 4.282042865309616,
      "tokens_seen": 229376000
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004699398194583752,
      "loss": 2.9385,
      "theoretical_loss": 4.281893041149093,
      "tokens_seen": 229441536
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004699297893681043,
      "loss": 3.3297,
      "theoretical_loss": 4.2817432717557455,
      "tokens_seen": 229507072
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046991975927783354,
      "loss": 2.8433,
      "theoretical_loss": 4.28159355709392,
      "tokens_seen": 229572608
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046990972918756267,
      "loss": 3.0393,
      "theoretical_loss": 4.281443897128004,
      "tokens_seen": 229638144
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004698996990972919,
      "loss": 3.0533,
      "theoretical_loss": 4.2812942918224115,
      "tokens_seen": 229703680
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004698896690070211,
      "loss": 3.0605,
      "theoretical_loss": 4.281144741141593,
      "tokens_seen": 229769216
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046987963891675027,
      "loss": 3.2703,
      "theoretical_loss": 4.280995245050032,
      "tokens_seen": 229834752
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046986960882647945,
      "loss": 3.296,
      "theoretical_loss": 4.2808458035122445,
      "tokens_seen": 229900288
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004698595787362087,
      "loss": 3.1057,
      "theoretical_loss": 4.2806964164927805,
      "tokens_seen": 229965824
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004698495486459378,
      "loss": 3.2996,
      "theoretical_loss": 4.280547083956224,
      "tokens_seen": 230031360
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046983951855566705,
      "loss": 3.1414,
      "theoretical_loss": 4.280397805867188,
      "tokens_seen": 230096896
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004698294884653962,
      "loss": 2.9225,
      "theoretical_loss": 4.280248582190324,
      "tokens_seen": 230162432
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004698194583751254,
      "loss": 3.1511,
      "theoretical_loss": 4.280099412890312,
      "tokens_seen": 230227968
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004698094282848546,
      "loss": 3.36,
      "theoretical_loss": 4.279950297931869,
      "tokens_seen": 230293504
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046979939819458377,
      "loss": 3.1327,
      "theoretical_loss": 4.27980123727974,
      "tokens_seen": 230359040
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046978936810431295,
      "loss": 3.1655,
      "theoretical_loss": 4.279652230898709,
      "tokens_seen": 230424576
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046977933801404213,
      "loss": 2.9951,
      "theoretical_loss": 4.279503278753586,
      "tokens_seen": 230490112
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004697693079237713,
      "loss": 3.1871,
      "theoretical_loss": 4.27935438080922,
      "tokens_seen": 230555648
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046975927783350055,
      "loss": 3.413,
      "theoretical_loss": 4.27920553703049,
      "tokens_seen": 230621184
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004697492477432297,
      "loss": 3.0407,
      "theoretical_loss": 4.279056747382306,
      "tokens_seen": 230686720
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004697392176529589,
      "loss": 3.2621,
      "theoretical_loss": 4.278908011829613,
      "tokens_seen": 230752256
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046972918756268804,
      "loss": 3.0509,
      "theoretical_loss": 4.27875933033739,
      "tokens_seen": 230817792
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004697191574724173,
      "loss": 3.1669,
      "theoretical_loss": 4.278610702870646,
      "tokens_seen": 230883328
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046970912738214646,
      "loss": 3.0121,
      "theoretical_loss": 4.278462129394423,
      "tokens_seen": 230948864
    },
    {
      "epoch": 0.07,
      "objective/train/docs_used": 290404,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.064453125,
      "objective/train/theoretical_loss": 4.278313609873795,
      "objective/train/tokens_used": 251474400,
      "theoretical_loss": 4.278313609873795,
      "tokens_seen": 231014400
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046969909729187564,
      "loss": 3.1929,
      "theoretical_loss": 4.278313609873795,
      "tokens_seen": 231014400
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004696890672016048,
      "loss": 3.1449,
      "theoretical_loss": 4.278165144273871,
      "tokens_seen": 231079936
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046967903711133405,
      "loss": 3.1433,
      "theoretical_loss": 4.27801673255979,
      "tokens_seen": 231145472
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004696690070210632,
      "loss": 3.1905,
      "theoretical_loss": 4.277868374696725,
      "tokens_seen": 231211008
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004696589769307924,
      "loss": 3.2421,
      "theoretical_loss": 4.277720070649879,
      "tokens_seen": 231276544
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046964894684052154,
      "loss": 3.3793,
      "theoretical_loss": 4.277571820384491,
      "tokens_seen": 231342080
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004696389167502508,
      "loss": 3.0555,
      "theoretical_loss": 4.277423623865829,
      "tokens_seen": 231407616
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046962888665997996,
      "loss": 3.0292,
      "theoretical_loss": 4.277275481059195,
      "tokens_seen": 231473152
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046961885656970914,
      "loss": 3.2957,
      "theoretical_loss": 4.2771273919299215,
      "tokens_seen": 231538688
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004696088264794383,
      "loss": 3.2876,
      "theoretical_loss": 4.276979356443377,
      "tokens_seen": 231604224
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004695987963891675,
      "loss": 3.3123,
      "theoretical_loss": 4.276831374564957,
      "tokens_seen": 231669760
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004695887662988967,
      "loss": 3.0926,
      "theoretical_loss": 4.276683446260093,
      "tokens_seen": 231735296
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004695787362086259,
      "loss": 3.0816,
      "theoretical_loss": 4.276535571494247,
      "tokens_seen": 231800832
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046956870611835505,
      "loss": 2.8716,
      "theoretical_loss": 4.276387750232913,
      "tokens_seen": 231866368
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004695586760280843,
      "loss": 3.5923,
      "theoretical_loss": 4.276239982441617,
      "tokens_seen": 231931904
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046954864593781346,
      "loss": 3.1727,
      "theoretical_loss": 4.276092268085918,
      "tokens_seen": 231997440
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046953861584754264,
      "loss": 3.0979,
      "theoretical_loss": 4.275944607131406,
      "tokens_seen": 232062976
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004695285857572718,
      "loss": 3.179,
      "theoretical_loss": 4.275796999543703,
      "tokens_seen": 232128512
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000469518555667001,
      "loss": 3.0949,
      "theoretical_loss": 4.275649445288461,
      "tokens_seen": 232194048
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004695085255767302,
      "loss": 3.2259,
      "theoretical_loss": 4.275501944331367,
      "tokens_seen": 232259584
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004694984954864594,
      "loss": 3.2972,
      "theoretical_loss": 4.275354496638139,
      "tokens_seen": 232325120
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046948846539618855,
      "loss": 3.5002,
      "theoretical_loss": 4.275207102174525,
      "tokens_seen": 232390656
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004694784353059178,
      "loss": 2.9256,
      "theoretical_loss": 4.275059760906305,
      "tokens_seen": 232456192
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004694684052156469,
      "loss": 3.3357,
      "theoretical_loss": 4.2749124727992935,
      "tokens_seen": 232521728
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046945837512537615,
      "loss": 3.146,
      "theoretical_loss": 4.274765237819333,
      "tokens_seen": 232587264
    },
    {
      "epoch": 0.07,
      "objective/train/docs_used": 291491,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.0524933338165283,
      "objective/train/theoretical_loss": 4.274618055932298,
      "objective/train/tokens_used": 253112800,
      "theoretical_loss": 4.274618055932298,
      "tokens_seen": 232652800
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004694483450351053,
      "loss": 2.9909,
      "theoretical_loss": 4.274618055932298,
      "tokens_seen": 232652800
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004694383149448345,
      "loss": 3.1934,
      "theoretical_loss": 4.2744709271040975,
      "tokens_seen": 232718336
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004694282848545637,
      "loss": 3.075,
      "theoretical_loss": 4.27432385130067,
      "tokens_seen": 232783872
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046941825476429287,
      "loss": 3.3298,
      "theoretical_loss": 4.274176828487984,
      "tokens_seen": 232849408
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046940822467402205,
      "loss": 3.1214,
      "theoretical_loss": 4.2740298586320415,
      "tokens_seen": 232914944
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004693981945837513,
      "loss": 3.2709,
      "theoretical_loss": 4.273882941698876,
      "tokens_seen": 232980480
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004693881644934804,
      "loss": 3.351,
      "theoretical_loss": 4.27373607765455,
      "tokens_seen": 233046016
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046937813440320965,
      "loss": 3.24,
      "theoretical_loss": 4.2735892664651605,
      "tokens_seen": 233111552
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046936810431293883,
      "loss": 3.1293,
      "theoretical_loss": 4.273442508096833,
      "tokens_seen": 233177088
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000469358074222668,
      "loss": 3.2126,
      "theoretical_loss": 4.273295802515726,
      "tokens_seen": 233242624
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004693480441323972,
      "loss": 3.1292,
      "theoretical_loss": 4.273149149688028,
      "tokens_seen": 233308160
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004693380140421264,
      "loss": 3.0769,
      "theoretical_loss": 4.27300254957996,
      "tokens_seen": 233373696
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046932798395185555,
      "loss": 3.026,
      "theoretical_loss": 4.272856002157772,
      "tokens_seen": 233439232
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004693179538615848,
      "loss": 3.335,
      "theoretical_loss": 4.272709507387748,
      "tokens_seen": 233504768
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004693079237713139,
      "loss": 3.1597,
      "theoretical_loss": 4.2725630652362,
      "tokens_seen": 233570304
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046929789368104315,
      "loss": 3.1306,
      "theoretical_loss": 4.272416675669473,
      "tokens_seen": 233635840
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004692878635907723,
      "loss": 3.2993,
      "theoretical_loss": 4.272270338653942,
      "tokens_seen": 233701376
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004692778335005015,
      "loss": 3.2775,
      "theoretical_loss": 4.272124054156014,
      "tokens_seen": 233766912
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004692678034102307,
      "loss": 3.0821,
      "theoretical_loss": 4.271977822142125,
      "tokens_seen": 233832448
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004692577733199599,
      "loss": 3.1264,
      "theoretical_loss": 4.271831642578745,
      "tokens_seen": 233897984
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046924774322968906,
      "loss": 3.0504,
      "theoretical_loss": 4.27168551543237,
      "tokens_seen": 233963520
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046923771313941824,
      "loss": 3.29,
      "theoretical_loss": 4.271539440669532,
      "tokens_seen": 234029056
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004692276830491474,
      "loss": 3.2624,
      "theoretical_loss": 4.27139341825679,
      "tokens_seen": 234094592
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046921765295887666,
      "loss": 3.288,
      "theoretical_loss": 4.271247448160736,
      "tokens_seen": 234160128
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046920762286860584,
      "loss": 3.2345,
      "theoretical_loss": 4.27110153034799,
      "tokens_seen": 234225664
    },
    {
      "epoch": 0.07,
      "objective/train/docs_used": 292210,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.441500663757324,
      "objective/train/theoretical_loss": 4.270955664785207,
      "objective/train/tokens_used": 254751200,
      "theoretical_loss": 4.270955664785207,
      "tokens_seen": 234291200
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000469197592778335,
      "loss": 3.2181,
      "theoretical_loss": 4.270955664785207,
      "tokens_seen": 234291200
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046918756268806425,
      "loss": 3.0661,
      "theoretical_loss": 4.2708098514390676,
      "tokens_seen": 234356736
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004691775325977934,
      "loss": 3.0524,
      "theoretical_loss": 4.270664090276286,
      "tokens_seen": 234422272
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004691675025075226,
      "loss": 3.2508,
      "theoretical_loss": 4.2705183812636065,
      "tokens_seen": 234487808
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046915747241725174,
      "loss": 3.3243,
      "theoretical_loss": 4.270372724367803,
      "tokens_seen": 234553344
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000469147442326981,
      "loss": 3.2352,
      "theoretical_loss": 4.270227119555681,
      "tokens_seen": 234618880
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046913741223671016,
      "loss": 3.1664,
      "theoretical_loss": 4.270081566794076,
      "tokens_seen": 234684416
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046912738214643934,
      "loss": 3.2272,
      "theoretical_loss": 4.269936066049852,
      "tokens_seen": 234749952
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004691173520561685,
      "loss": 3.403,
      "theoretical_loss": 4.269790617289907,
      "tokens_seen": 234815488
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004691073219658977,
      "loss": 3.171,
      "theoretical_loss": 4.269645220481166,
      "tokens_seen": 234881024
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004690972918756269,
      "loss": 3.1504,
      "theoretical_loss": 4.269499875590587,
      "tokens_seen": 234946560
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004690872617853561,
      "loss": 3.3593,
      "theoretical_loss": 4.269354582585156,
      "tokens_seen": 235012096
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046907723169508525,
      "loss": 3.4094,
      "theoretical_loss": 4.269209341431889,
      "tokens_seen": 235077632
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004690672016048145,
      "loss": 3.0117,
      "theoretical_loss": 4.269064152097835,
      "tokens_seen": 235143168
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046905717151454366,
      "loss": 3.1555,
      "theoretical_loss": 4.26891901455007,
      "tokens_seen": 235208704
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046904714142427284,
      "loss": 3.3753,
      "theoretical_loss": 4.268773928755701,
      "tokens_seen": 235274240
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000469037111334002,
      "loss": 3.1484,
      "theoretical_loss": 4.268628894681868,
      "tokens_seen": 235339776
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004690270812437312,
      "loss": 3.2826,
      "theoretical_loss": 4.268483912295735,
      "tokens_seen": 235405312
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004690170511534604,
      "loss": 3.267,
      "theoretical_loss": 4.268338981564502,
      "tokens_seen": 235470848
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004690070210631896,
      "loss": 3.0986,
      "theoretical_loss": 4.268194102455395,
      "tokens_seen": 235536384
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046899699097291875,
      "loss": 3.0481,
      "theoretical_loss": 4.26804927493567,
      "tokens_seen": 235601920
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000468986960882648,
      "loss": 3.3586,
      "theoretical_loss": 4.267904498972618,
      "tokens_seen": 235667456
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004689769307923771,
      "loss": 3.1682,
      "theoretical_loss": 4.267759774533552,
      "tokens_seen": 235732992
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046896690070210635,
      "loss": 3.1601,
      "theoretical_loss": 4.267615101585821,
      "tokens_seen": 235798528
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046895687061183553,
      "loss": 3.3558,
      "theoretical_loss": 4.267470480096801,
      "tokens_seen": 235864064
    },
    {
      "epoch": 0.07,
      "objective/train/docs_used": 293604,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7816476821899414,
      "objective/train/theoretical_loss": 4.267325910033897,
      "objective/train/tokens_used": 256389600,
      "theoretical_loss": 4.267325910033897,
      "tokens_seen": 235929600
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004689468405215647,
      "loss": 3.0005,
      "theoretical_loss": 4.267325910033897,
      "tokens_seen": 235929600
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004689368104312939,
      "loss": 3.3922,
      "theoretical_loss": 4.267181391364547,
      "tokens_seen": 235995136
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046892678034102307,
      "loss": 3.2998,
      "theoretical_loss": 4.267036924056215,
      "tokens_seen": 236060672
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046891675025075225,
      "loss": 3.0282,
      "theoretical_loss": 4.266892508076397,
      "tokens_seen": 236126208
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004689067201604815,
      "loss": 3.2748,
      "theoretical_loss": 4.266748143392617,
      "tokens_seen": 236191744
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004688966900702106,
      "loss": 3.0778,
      "theoretical_loss": 4.26660382997243,
      "tokens_seen": 236257280
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046888665997993985,
      "loss": 3.3932,
      "theoretical_loss": 4.26645956778342,
      "tokens_seen": 236322816
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046887662988966903,
      "loss": 3.5158,
      "theoretical_loss": 4.2663153567932,
      "tokens_seen": 236388352
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004688665997993982,
      "loss": 2.9413,
      "theoretical_loss": 4.266171196969412,
      "tokens_seen": 236453888
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004688565697091274,
      "loss": 3.0881,
      "theoretical_loss": 4.2660270882797295,
      "tokens_seen": 236519424
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004688465396188566,
      "loss": 3.3507,
      "theoretical_loss": 4.265883030691853,
      "tokens_seen": 236584960
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046883650952858575,
      "loss": 3.2892,
      "theoretical_loss": 4.265739024173515,
      "tokens_seen": 236650496
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000468826479438315,
      "loss": 3.1898,
      "theoretical_loss": 4.265595068692473,
      "tokens_seen": 236716032
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004688164493480441,
      "loss": 3.0696,
      "theoretical_loss": 4.26545116421652,
      "tokens_seen": 236781568
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046880641925777335,
      "loss": 3.385,
      "theoretical_loss": 4.265307310713471,
      "tokens_seen": 236847104
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004687963891675025,
      "loss": 3.3006,
      "theoretical_loss": 4.2651635081511765,
      "tokens_seen": 236912640
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004687863590772317,
      "loss": 3.1971,
      "theoretical_loss": 4.265019756497512,
      "tokens_seen": 236978176
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004687763289869609,
      "loss": 3.0931,
      "theoretical_loss": 4.264876055720386,
      "tokens_seen": 237043712
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004687662988966901,
      "loss": 3.1088,
      "theoretical_loss": 4.264732405787731,
      "tokens_seen": 237109248
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046875626880641926,
      "loss": 3.0559,
      "theoretical_loss": 4.264588806667513,
      "tokens_seen": 237174784
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046874623871614844,
      "loss": 3.3792,
      "theoretical_loss": 4.264445258327724,
      "tokens_seen": 237240320
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004687362086258776,
      "loss": 3.2076,
      "theoretical_loss": 4.264301760736389,
      "tokens_seen": 237305856
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046872617853560686,
      "loss": 3.2832,
      "theoretical_loss": 4.264158313861557,
      "tokens_seen": 237371392
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000468716148445336,
      "loss": 3.3468,
      "theoretical_loss": 4.264014917671309,
      "tokens_seen": 237436928
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004687061183550652,
      "loss": 3.0554,
      "theoretical_loss": 4.2638715721337554,
      "tokens_seen": 237502464
    },
    {
      "epoch": 0.07,
      "objective/train/docs_used": 294906,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.298981189727783,
      "objective/train/theoretical_loss": 4.263728277217032,
      "objective/train/tokens_used": 258028000,
      "theoretical_loss": 4.263728277217032,
      "tokens_seen": 237568000
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004686960882647944,
      "loss": 3.2936,
      "theoretical_loss": 4.263728277217032,
      "tokens_seen": 237568000
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004686860581745236,
      "loss": 3.0715,
      "theoretical_loss": 4.263585032889306,
      "tokens_seen": 237633536
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046867602808425276,
      "loss": 3.2912,
      "theoretical_loss": 4.263441839118776,
      "tokens_seen": 237699072
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046866599799398194,
      "loss": 2.9441,
      "theoretical_loss": 4.2632986958736625,
      "tokens_seen": 237764608
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004686559679037111,
      "loss": 3.467,
      "theoretical_loss": 4.263155603122221,
      "tokens_seen": 237830144
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046864593781344036,
      "loss": 3.2158,
      "theoretical_loss": 4.263012560832733,
      "tokens_seen": 237895680
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004686359077231695,
      "loss": 3.2119,
      "theoretical_loss": 4.262869568973508,
      "tokens_seen": 237961216
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004686258776328987,
      "loss": 3.2893,
      "theoretical_loss": 4.262726627512886,
      "tokens_seen": 238026752
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046861584754262785,
      "loss": 3.4987,
      "theoretical_loss": 4.262583736419234,
      "tokens_seen": 238092288
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004686058174523571,
      "loss": 3.271,
      "theoretical_loss": 4.26244089566095,
      "tokens_seen": 238157824
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046859578736208626,
      "loss": 3.2722,
      "theoretical_loss": 4.262298105206456,
      "tokens_seen": 238223360
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046858575727181545,
      "loss": 3.3172,
      "theoretical_loss": 4.262155365024207,
      "tokens_seen": 238288896
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004685757271815446,
      "loss": 3.2976,
      "theoretical_loss": 4.262012675082685,
      "tokens_seen": 238354432
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004685656970912738,
      "loss": 3.229,
      "theoretical_loss": 4.261870035350399,
      "tokens_seen": 238419968
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000468555667001003,
      "loss": 3.5022,
      "theoretical_loss": 4.261727445795888,
      "tokens_seen": 238485504
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004685456369107322,
      "loss": 3.1887,
      "theoretical_loss": 4.26158490638772,
      "tokens_seen": 238551040
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046853560682046135,
      "loss": 3.1939,
      "theoretical_loss": 4.261442417094488,
      "tokens_seen": 238616576
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004685255767301906,
      "loss": 3.0544,
      "theoretical_loss": 4.261299977884816,
      "tokens_seen": 238682112
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046851554663991977,
      "loss": 3.1491,
      "theoretical_loss": 4.2611575887273565,
      "tokens_seen": 238747648
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046850551654964895,
      "loss": 3.215,
      "theoretical_loss": 4.261015249590789,
      "tokens_seen": 238813184
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046849548645937813,
      "loss": 3.2087,
      "theoretical_loss": 4.260872960443822,
      "tokens_seen": 238878720
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004684854563691073,
      "loss": 3.0708,
      "theoretical_loss": 4.260730721255191,
      "tokens_seen": 238944256
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004684754262788365,
      "loss": 2.872,
      "theoretical_loss": 4.260588531993662,
      "tokens_seen": 239009792
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046846539618856573,
      "loss": 3.329,
      "theoretical_loss": 4.260446392628026,
      "tokens_seen": 239075328
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004684553660982949,
      "loss": 3.2693,
      "theoretical_loss": 4.2603043031271035,
      "tokens_seen": 239140864
    },
    {
      "epoch": 0.07,
      "objective/train/docs_used": 295763,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9719858169555664,
      "objective/train/theoretical_loss": 4.260162263459744,
      "objective/train/tokens_used": 259666400,
      "theoretical_loss": 4.260162263459744,
      "tokens_seen": 239206400
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004684453360080241,
      "loss": 3.251,
      "theoretical_loss": 4.260162263459744,
      "tokens_seen": 239206400
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046843530591775327,
      "loss": 3.3018,
      "theoretical_loss": 4.260020273594824,
      "tokens_seen": 239271936
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046842527582748245,
      "loss": 2.9441,
      "theoretical_loss": 4.259878333501247,
      "tokens_seen": 239337472
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004684152457372117,
      "loss": 3.3198,
      "theoretical_loss": 4.259736443147946,
      "tokens_seen": 239403008
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004684052156469408,
      "loss": 3.1598,
      "theoretical_loss": 4.259594602503881,
      "tokens_seen": 239468544
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046839518555667005,
      "loss": 3.4569,
      "theoretical_loss": 4.259452811538041,
      "tokens_seen": 239534080
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046838515546639923,
      "loss": 3.3406,
      "theoretical_loss": 4.259311070219441,
      "tokens_seen": 239599616
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004683751253761284,
      "loss": 3.0996,
      "theoretical_loss": 4.259169378517125,
      "tokens_seen": 239665152
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004683650952858576,
      "loss": 3.2584,
      "theoretical_loss": 4.259027736400165,
      "tokens_seen": 239730688
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004683550651955868,
      "loss": 3.4028,
      "theoretical_loss": 4.258886143837661,
      "tokens_seen": 239796224
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046834503510531595,
      "loss": 3.2419,
      "theoretical_loss": 4.258744600798739,
      "tokens_seen": 239861760
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004683350050150452,
      "loss": 3.121,
      "theoretical_loss": 4.2586031072525525,
      "tokens_seen": 239927296
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004683249749247743,
      "loss": 3.1897,
      "theoretical_loss": 4.258461663168285,
      "tokens_seen": 239992832
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046831494483450355,
      "loss": 3.0282,
      "theoretical_loss": 4.258320268515147,
      "tokens_seen": 240058368
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004683049147442327,
      "loss": 3.1978,
      "theoretical_loss": 4.258178923262376,
      "tokens_seen": 240123904
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004682948846539619,
      "loss": 3.2214,
      "theoretical_loss": 4.258037627379235,
      "tokens_seen": 240189440
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004682848545636911,
      "loss": 3.3644,
      "theoretical_loss": 4.257896380835018,
      "tokens_seen": 240254976
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004682748244734203,
      "loss": 3.1584,
      "theoretical_loss": 4.257755183599045,
      "tokens_seen": 240320512
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046826479438314946,
      "loss": 3.1607,
      "theoretical_loss": 4.257614035640662,
      "tokens_seen": 240386048
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046825476429287864,
      "loss": 2.992,
      "theoretical_loss": 4.257472936929246,
      "tokens_seen": 240451584
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004682447342026078,
      "loss": 3.399,
      "theoretical_loss": 4.257331887434198,
      "tokens_seen": 240517120
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046823470411233706,
      "loss": 3.076,
      "theoretical_loss": 4.257190887124946,
      "tokens_seen": 240582656
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004682246740220662,
      "loss": 3.1499,
      "theoretical_loss": 4.25704993597095,
      "tokens_seen": 240648192
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004682146439317954,
      "loss": 3.3741,
      "theoretical_loss": 4.256909033941691,
      "tokens_seen": 240713728
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004682046138415246,
      "loss": 3.2284,
      "theoretical_loss": 4.256768181006683,
      "tokens_seen": 240779264
    },
    {
      "epoch": 0.07,
      "objective/train/docs_used": 296516,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.0105109214782715,
      "objective/train/theoretical_loss": 4.2566273771354615,
      "objective/train/tokens_used": 261304800,
      "theoretical_loss": 4.2566273771354615,
      "tokens_seen": 240844800
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004681945837512538,
      "loss": 3.02,
      "theoretical_loss": 4.2566273771354615,
      "tokens_seen": 240844800
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046818455366098296,
      "loss": 3.3288,
      "theoretical_loss": 4.256486622297595,
      "tokens_seen": 240910336
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046817452357071214,
      "loss": 3.397,
      "theoretical_loss": 4.256345916462674,
      "tokens_seen": 240975872
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004681644934804413,
      "loss": 3.0537,
      "theoretical_loss": 4.256205259600321,
      "tokens_seen": 241041408
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046815446339017056,
      "loss": 3.077,
      "theoretical_loss": 4.256064651680182,
      "tokens_seen": 241106944
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004681444332998997,
      "loss": 3.2307,
      "theoretical_loss": 4.255924092671931,
      "tokens_seen": 241172480
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004681344032096289,
      "loss": 2.9942,
      "theoretical_loss": 4.255783582545269,
      "tokens_seen": 241238016
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046812437311935805,
      "loss": 3.281,
      "theoretical_loss": 4.255643121269924,
      "tokens_seen": 241303552
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004681143430290873,
      "loss": 3.2357,
      "theoretical_loss": 4.255502708815651,
      "tokens_seen": 241369088
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046810431293881646,
      "loss": 3.098,
      "theoretical_loss": 4.255362345152234,
      "tokens_seen": 241434624
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046809428284854565,
      "loss": 3.0425,
      "theoretical_loss": 4.255222030249479,
      "tokens_seen": 241500160
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004680842527582748,
      "loss": 3.1521,
      "theoretical_loss": 4.255081764077224,
      "tokens_seen": 241565696
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000468074222668004,
      "loss": 3.3244,
      "theoretical_loss": 4.25494154660533,
      "tokens_seen": 241631232
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004680641925777332,
      "loss": 3.0535,
      "theoretical_loss": 4.254801377803689,
      "tokens_seen": 241696768
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004680541624874624,
      "loss": 3.0864,
      "theoretical_loss": 4.254661257642215,
      "tokens_seen": 241762304
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046804413239719155,
      "loss": 3.1507,
      "theoretical_loss": 4.254521186090852,
      "tokens_seen": 241827840
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004680341023069208,
      "loss": 3.1771,
      "theoretical_loss": 4.254381163119568,
      "tokens_seen": 241893376
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046802407221664997,
      "loss": 3.3228,
      "theoretical_loss": 4.254241188698361,
      "tokens_seen": 241958912
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046801404212637915,
      "loss": 3.1643,
      "theoretical_loss": 4.2541012627972545,
      "tokens_seen": 242024448
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046800401203610833,
      "loss": 3.0985,
      "theoretical_loss": 4.2539613853862965,
      "tokens_seen": 242089984
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004679939819458375,
      "loss": 3.1079,
      "theoretical_loss": 4.253821556435565,
      "tokens_seen": 242155520
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004679839518555667,
      "loss": 3.1036,
      "theoretical_loss": 4.253681775915161,
      "tokens_seen": 242221056
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046797392176529593,
      "loss": 3.1256,
      "theoretical_loss": 4.253542043795215,
      "tokens_seen": 242286592
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046796389167502505,
      "loss": 3.0556,
      "theoretical_loss": 4.253402360045882,
      "tokens_seen": 242352128
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004679538615847543,
      "loss": 3.2891,
      "theoretical_loss": 4.253262724637346,
      "tokens_seen": 242417664
    },
    {
      "epoch": 0.07,
      "objective/train/docs_used": 297983,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.877701997756958,
      "objective/train/theoretical_loss": 4.253123137539814,
      "objective/train/tokens_used": 262943200,
      "theoretical_loss": 4.253123137539814,
      "tokens_seen": 242483200
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004679438314944834,
      "loss": 3.1046,
      "theoretical_loss": 4.253123137539814,
      "tokens_seen": 242483200
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046793380140421265,
      "loss": 3.1167,
      "theoretical_loss": 4.252983598723521,
      "tokens_seen": 242548736
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046792377131394183,
      "loss": 3.4114,
      "theoretical_loss": 4.25284410815873,
      "tokens_seen": 242614272
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000467913741223671,
      "loss": 3.323,
      "theoretical_loss": 4.2527046658157275,
      "tokens_seen": 242679808
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004679037111334002,
      "loss": 2.809,
      "theoretical_loss": 4.252565271664828,
      "tokens_seen": 242745344
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046789368104312943,
      "loss": 3.1643,
      "theoretical_loss": 4.252425925676373,
      "tokens_seen": 242810880
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046788365095285856,
      "loss": 3.4548,
      "theoretical_loss": 4.252286627820727,
      "tokens_seen": 242876416
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004678736208625878,
      "loss": 3.2465,
      "theoretical_loss": 4.252147378068285,
      "tokens_seen": 242941952
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004678635907723169,
      "loss": 3.2615,
      "theoretical_loss": 4.252008176389465,
      "tokens_seen": 243007488
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046785356068204616,
      "loss": 3.1489,
      "theoretical_loss": 4.251869022754712,
      "tokens_seen": 243073024
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046784353059177534,
      "loss": 3.2578,
      "theoretical_loss": 4.251729917134498,
      "tokens_seen": 243138560
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004678335005015045,
      "loss": 3.0778,
      "theoretical_loss": 4.251590859499322,
      "tokens_seen": 243204096
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004678234704112337,
      "loss": 3.046,
      "theoretical_loss": 4.251451849819704,
      "tokens_seen": 243269632
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004678134403209629,
      "loss": 3.255,
      "theoretical_loss": 4.251312888066197,
      "tokens_seen": 243335168
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046780341023069206,
      "loss": 2.9979,
      "theoretical_loss": 4.251173974209375,
      "tokens_seen": 243400704
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004677933801404213,
      "loss": 3.0556,
      "theoretical_loss": 4.251035108219839,
      "tokens_seen": 243466240
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004677833500501504,
      "loss": 3.2338,
      "theoretical_loss": 4.250896290068218,
      "tokens_seen": 243531776
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046777331995987966,
      "loss": 3.1279,
      "theoretical_loss": 4.250757519725165,
      "tokens_seen": 243597312
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004677632898696088,
      "loss": 3.2705,
      "theoretical_loss": 4.25061879716136,
      "tokens_seen": 243662848
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000467753259779338,
      "loss": 3.1101,
      "theoretical_loss": 4.250480122347507,
      "tokens_seen": 243728384
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004677432296890672,
      "loss": 2.9749,
      "theoretical_loss": 4.250341495254337,
      "tokens_seen": 243793920
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004677331995987964,
      "loss": 3.0554,
      "theoretical_loss": 4.250202915852608,
      "tokens_seen": 243859456
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046772316950852556,
      "loss": 2.9862,
      "theoretical_loss": 4.250064384113102,
      "tokens_seen": 243924992
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004677131394182548,
      "loss": 2.8649,
      "theoretical_loss": 4.249925900006627,
      "tokens_seen": 243990528
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000467703109327984,
      "loss": 3.533,
      "theoretical_loss": 4.249787463504019,
      "tokens_seen": 244056064
    },
    {
      "epoch": 0.07,
      "objective/train/docs_used": 298552,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.992788553237915,
      "objective/train/theoretical_loss": 4.249649074576134,
      "objective/train/tokens_used": 264581600,
      "theoretical_loss": 4.249649074576134,
      "tokens_seen": 244121600
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046769307923771316,
      "loss": 3.0907,
      "theoretical_loss": 4.249649074576134,
      "tokens_seen": 244121600
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046768304914744234,
      "loss": 2.9822,
      "theoretical_loss": 4.249510733193862,
      "tokens_seen": 244187136
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004676730190571715,
      "loss": 3.1675,
      "theoretical_loss": 4.249372439328111,
      "tokens_seen": 244252672
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046766298896690076,
      "loss": 3.1947,
      "theoretical_loss": 4.249234192949818,
      "tokens_seen": 244318208
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004676529588766299,
      "loss": 3.1614,
      "theoretical_loss": 4.249095994029947,
      "tokens_seen": 244383744
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004676429287863591,
      "loss": 3.2011,
      "theoretical_loss": 4.248957842539484,
      "tokens_seen": 244449280
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046763289869608825,
      "loss": 3.1561,
      "theoretical_loss": 4.248819738449442,
      "tokens_seen": 244514816
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004676228686058175,
      "loss": 3.3678,
      "theoretical_loss": 4.2486816817308615,
      "tokens_seen": 244580352
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046761283851554666,
      "loss": 3.1089,
      "theoretical_loss": 4.248543672354805,
      "tokens_seen": 244645888
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046760280842527585,
      "loss": 3.2738,
      "theoretical_loss": 4.248405710292364,
      "tokens_seen": 244711424
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000467592778335005,
      "loss": 3.3828,
      "theoretical_loss": 4.248267795514652,
      "tokens_seen": 244776960
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004675827482447342,
      "loss": 3.1759,
      "theoretical_loss": 4.248129927992808,
      "tokens_seen": 244842496
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004675727181544634,
      "loss": 3.004,
      "theoretical_loss": 4.247992107698002,
      "tokens_seen": 244908032
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004675626880641926,
      "loss": 3.2486,
      "theoretical_loss": 4.247854334601421,
      "tokens_seen": 244973568
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046755265797392175,
      "loss": 3.2047,
      "theoretical_loss": 4.247716608674283,
      "tokens_seen": 245039104
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000467542627883651,
      "loss": 3.0704,
      "theoretical_loss": 4.247578929887829,
      "tokens_seen": 245104640
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046753259779338017,
      "loss": 2.9287,
      "theoretical_loss": 4.247441298213326,
      "tokens_seen": 245170176
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046752256770310935,
      "loss": 3.0956,
      "theoretical_loss": 4.247303713622067,
      "tokens_seen": 245235712
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046751253761283853,
      "loss": 3.0942,
      "theoretical_loss": 4.247166176085367,
      "tokens_seen": 245301248
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004675025075225677,
      "loss": 3.0354,
      "theoretical_loss": 4.247028685574569,
      "tokens_seen": 245366784
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004674924774322969,
      "loss": 3.135,
      "theoretical_loss": 4.246891242061041,
      "tokens_seen": 245432320
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046748244734202613,
      "loss": 3.0614,
      "theoretical_loss": 4.246753845516174,
      "tokens_seen": 245497856
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046747241725175525,
      "loss": 3.1953,
      "theoretical_loss": 4.246616495911388,
      "tokens_seen": 245563392
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004674623871614845,
      "loss": 2.9679,
      "theoretical_loss": 4.246479193218123,
      "tokens_seen": 245628928
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004674523570712136,
      "loss": 3.3495,
      "theoretical_loss": 4.246341937407848,
      "tokens_seen": 245694464
    },
    {
      "epoch": 0.07,
      "objective/train/docs_used": 299755,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.3899707794189453,
      "objective/train/theoretical_loss": 4.246204728452055,
      "objective/train/tokens_used": 266220000,
      "theoretical_loss": 4.246204728452055,
      "tokens_seen": 245760000
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046744232698094285,
      "loss": 3.138,
      "theoretical_loss": 4.246204728452055,
      "tokens_seen": 245760000
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046743229689067203,
      "loss": 3.3381,
      "theoretical_loss": 4.246067566322259,
      "tokens_seen": 245825536
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004674222668004012,
      "loss": 2.9391,
      "theoretical_loss": 4.245930450990007,
      "tokens_seen": 245891072
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004674122367101304,
      "loss": 3.2446,
      "theoretical_loss": 4.245793382426861,
      "tokens_seen": 245956608
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046740220661985963,
      "loss": 3.1792,
      "theoretical_loss": 4.245656360604417,
      "tokens_seen": 246022144
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046739217652958876,
      "loss": 3.3674,
      "theoretical_loss": 4.24551938549429,
      "tokens_seen": 246087680
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000467382146439318,
      "loss": 3.1692,
      "theoretical_loss": 4.2453824570681205,
      "tokens_seen": 246153216
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004673721163490471,
      "loss": 3.3341,
      "theoretical_loss": 4.245245575297577,
      "tokens_seen": 246218752
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046736208625877636,
      "loss": 3.3756,
      "theoretical_loss": 4.2451087401543495,
      "tokens_seen": 246284288
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046735205616850554,
      "loss": 3.2973,
      "theoretical_loss": 4.244971951610154,
      "tokens_seen": 246349824
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004673420260782347,
      "loss": 3.376,
      "theoretical_loss": 4.24483520963673,
      "tokens_seen": 246415360
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004673319959879639,
      "loss": 2.9486,
      "theoretical_loss": 4.244698514205844,
      "tokens_seen": 246480896
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004673219658976931,
      "loss": 3.0086,
      "theoretical_loss": 4.244561865289285,
      "tokens_seen": 246546432
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046731193580742226,
      "loss": 3.1113,
      "theoretical_loss": 4.244425262858867,
      "tokens_seen": 246611968
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004673019057171515,
      "loss": 3.2665,
      "theoretical_loss": 4.2442887068864295,
      "tokens_seen": 246677504
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004672918756268806,
      "loss": 3.1233,
      "theoretical_loss": 4.244152197343835,
      "tokens_seen": 246743040
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046728184553660986,
      "loss": 3.3192,
      "theoretical_loss": 4.244015734202973,
      "tokens_seen": 246808576
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000467271815446339,
      "loss": 3.3794,
      "theoretical_loss": 4.243879317435755,
      "tokens_seen": 246874112
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004672617853560682,
      "loss": 2.863,
      "theoretical_loss": 4.243742947014117,
      "tokens_seen": 246939648
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004672517552657974,
      "loss": 3.3378,
      "theoretical_loss": 4.243606622910021,
      "tokens_seen": 247005184
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004672417251755266,
      "loss": 2.9326,
      "theoretical_loss": 4.243470345095453,
      "tokens_seen": 247070720
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046723169508525576,
      "loss": 3.1262,
      "theoretical_loss": 4.2433341135424225,
      "tokens_seen": 247136256
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000467221664994985,
      "loss": 3.2178,
      "theoretical_loss": 4.243197928222964,
      "tokens_seen": 247201792
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004672116349047141,
      "loss": 3.2335,
      "theoretical_loss": 4.243061789109136,
      "tokens_seen": 247267328
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046720160481444336,
      "loss": 3.1028,
      "theoretical_loss": 4.242925696173021,
      "tokens_seen": 247332864
    },
    {
      "epoch": 0.07,
      "objective/train/docs_used": 300199,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.2579872608184814,
      "objective/train/theoretical_loss": 4.2427896493867285,
      "objective/train/tokens_used": 267858400,
      "theoretical_loss": 4.2427896493867285,
      "tokens_seen": 247398400
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004671915747241725,
      "loss": 3.0951,
      "theoretical_loss": 4.2427896493867285,
      "tokens_seen": 247398400
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004671815446339017,
      "loss": 3.0132,
      "theoretical_loss": 4.242653648722387,
      "tokens_seen": 247463936
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004671715145436309,
      "loss": 3.3596,
      "theoretical_loss": 4.242517694152154,
      "tokens_seen": 247529472
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004671614844533601,
      "loss": 2.881,
      "theoretical_loss": 4.24238178564821,
      "tokens_seen": 247595008
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046715145436308927,
      "loss": 3.3289,
      "theoretical_loss": 4.242245923182756,
      "tokens_seen": 247660544
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046714142427281845,
      "loss": 3.1006,
      "theoretical_loss": 4.242110106728022,
      "tokens_seen": 247726080
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046713139418254763,
      "loss": 3.3098,
      "theoretical_loss": 4.241974336256261,
      "tokens_seen": 247791616
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046712136409227686,
      "loss": 3.4344,
      "theoretical_loss": 4.241838611739748,
      "tokens_seen": 247857152
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000467111334002006,
      "loss": 3.3127,
      "theoretical_loss": 4.241702933150783,
      "tokens_seen": 247922688
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004671013039117352,
      "loss": 3.2709,
      "theoretical_loss": 4.241567300461693,
      "tokens_seen": 247988224
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046709127382146435,
      "loss": 3.2324,
      "theoretical_loss": 4.241431713644823,
      "tokens_seen": 248053760
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004670812437311936,
      "loss": 2.8616,
      "theoretical_loss": 4.241296172672547,
      "tokens_seen": 248119296
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046707121364092277,
      "loss": 3.1879,
      "theoretical_loss": 4.24116067751726,
      "tokens_seen": 248184832
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046706118355065195,
      "loss": 3.2202,
      "theoretical_loss": 4.241025228151383,
      "tokens_seen": 248250368
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046705115346038113,
      "loss": 3.2631,
      "theoretical_loss": 4.24088982454736,
      "tokens_seen": 248315904
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046704112337011037,
      "loss": 3.0812,
      "theoretical_loss": 4.240754466677659,
      "tokens_seen": 248381440
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004670310932798395,
      "loss": 3.2234,
      "theoretical_loss": 4.240619154514771,
      "tokens_seen": 248446976
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046702106318956873,
      "loss": 3.2165,
      "theoretical_loss": 4.240483888031212,
      "tokens_seen": 248512512
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046701103309929786,
      "loss": 2.9284,
      "theoretical_loss": 4.240348667199521,
      "tokens_seen": 248578048
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004670010030090271,
      "loss": 3.1256,
      "theoretical_loss": 4.240213491992261,
      "tokens_seen": 248643584
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004669909729187563,
      "loss": 3.1306,
      "theoretical_loss": 4.240078362382019,
      "tokens_seen": 248709120
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046698094282848545,
      "loss": 3.1289,
      "theoretical_loss": 4.239943278341404,
      "tokens_seen": 248774656
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046697091273821464,
      "loss": 3.2737,
      "theoretical_loss": 4.239808239843052,
      "tokens_seen": 248840192
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004669608826479438,
      "loss": 3.1408,
      "theoretical_loss": 4.239673246859619,
      "tokens_seen": 248905728
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046695085255767305,
      "loss": 3.2095,
      "theoretical_loss": 4.239538299363788,
      "tokens_seen": 248971264
    },
    {
      "epoch": 0.08,
      "objective/train/docs_used": 301486,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.3037164211273193,
      "objective/train/theoretical_loss": 4.239403397328261,
      "objective/train/tokens_used": 269496800,
      "theoretical_loss": 4.239403397328261,
      "tokens_seen": 249036800
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046694082246740223,
      "loss": 3.109,
      "theoretical_loss": 4.239403397328261,
      "tokens_seen": 249036800
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004669307923771314,
      "loss": 3.2687,
      "theoretical_loss": 4.239268540725769,
      "tokens_seen": 249102336
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004669207622868606,
      "loss": 3.2338,
      "theoretical_loss": 4.239133729529064,
      "tokens_seen": 249167872
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046691073219658983,
      "loss": 3.023,
      "theoretical_loss": 4.2389989637109196,
      "tokens_seen": 249233408
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046690070210631896,
      "loss": 3.2841,
      "theoretical_loss": 4.2388642432441355,
      "tokens_seen": 249298944
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004668906720160482,
      "loss": 3.4195,
      "theoretical_loss": 4.238729568101535,
      "tokens_seen": 249364480
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004668806419257773,
      "loss": 3.0559,
      "theoretical_loss": 4.238594938255963,
      "tokens_seen": 249430016
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046687061183550656,
      "loss": 3.1747,
      "theoretical_loss": 4.2384603536802885,
      "tokens_seen": 249495552
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046686058174523574,
      "loss": 3.1241,
      "theoretical_loss": 4.238325814347404,
      "tokens_seen": 249561088
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004668505516549649,
      "loss": 3.3317,
      "theoretical_loss": 4.238191320230227,
      "tokens_seen": 249626624
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004668405215646941,
      "loss": 3.127,
      "theoretical_loss": 4.238056871301695,
      "tokens_seen": 249692160
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004668304914744233,
      "loss": 3.1163,
      "theoretical_loss": 4.237922467534771,
      "tokens_seen": 249757696
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046682046138415246,
      "loss": 3.1401,
      "theoretical_loss": 4.237788108902441,
      "tokens_seen": 249823232
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004668104312938817,
      "loss": 3.285,
      "theoretical_loss": 4.237653795377714,
      "tokens_seen": 249888768
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004668004012036108,
      "loss": 3.4912,
      "theoretical_loss": 4.237519526933622,
      "tokens_seen": 249954304
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046679037111334006,
      "loss": 2.9408,
      "theoretical_loss": 4.2373853035432205,
      "tokens_seen": 250019840
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004667803410230692,
      "loss": 3.312,
      "theoretical_loss": 4.237251125179588,
      "tokens_seen": 250085376
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004667703109327984,
      "loss": 3.14,
      "theoretical_loss": 4.237116991815826,
      "tokens_seen": 250150912
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004667602808425276,
      "loss": 3.1893,
      "theoretical_loss": 4.23698290342506,
      "tokens_seen": 250216448
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004667502507522568,
      "loss": 3.1478,
      "theoretical_loss": 4.236848859980437,
      "tokens_seen": 250281984
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046674022066198596,
      "loss": 3.0303,
      "theoretical_loss": 4.23671486145513,
      "tokens_seen": 250347520
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004667301905717152,
      "loss": 3.1047,
      "theoretical_loss": 4.236580907822331,
      "tokens_seen": 250413056
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004667201604814443,
      "loss": 3.0878,
      "theoretical_loss": 4.236446999055257,
      "tokens_seen": 250478592
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046671013039117356,
      "loss": 3.123,
      "theoretical_loss": 4.2363131351271495,
      "tokens_seen": 250544128
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004667001003009027,
      "loss": 3.1892,
      "theoretical_loss": 4.2361793160112695,
      "tokens_seen": 250609664
    },
    {
      "epoch": 0.08,
      "objective/train/docs_used": 302298,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.1085546016693115,
      "objective/train/theoretical_loss": 4.236045541680905,
      "objective/train/tokens_used": 271135200,
      "theoretical_loss": 4.236045541680905,
      "tokens_seen": 250675200
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004666900702106319,
      "loss": 3.3128,
      "theoretical_loss": 4.236045541680905,
      "tokens_seen": 250675200
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004666800401203611,
      "loss": 3.097,
      "theoretical_loss": 4.235911812109363,
      "tokens_seen": 250740736
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004666700100300903,
      "loss": 3.054,
      "theoretical_loss": 4.235778127269976,
      "tokens_seen": 250806272
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046665997993981947,
      "loss": 3.2811,
      "theoretical_loss": 4.235644487136098,
      "tokens_seen": 250871808
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046664994984954865,
      "loss": 3.2177,
      "theoretical_loss": 4.235510891681108,
      "tokens_seen": 250937344
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046663991975927783,
      "loss": 3.1458,
      "theoretical_loss": 4.235377340878404,
      "tokens_seen": 251002880
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046662988966900706,
      "loss": 3.2887,
      "theoretical_loss": 4.23524383470141,
      "tokens_seen": 251068416
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004666198595787362,
      "loss": 3.5902,
      "theoretical_loss": 4.235110373123572,
      "tokens_seen": 251133952
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046660982948846543,
      "loss": 3.2736,
      "theoretical_loss": 4.2349769561183574,
      "tokens_seen": 251199488
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046659979939819455,
      "loss": 3.3576,
      "theoretical_loss": 4.2348435836592575,
      "tokens_seen": 251265024
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004665897693079238,
      "loss": 3.0621,
      "theoretical_loss": 4.234710255719786,
      "tokens_seen": 251330560
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046657973921765297,
      "loss": 3.2782,
      "theoretical_loss": 4.234576972273481,
      "tokens_seen": 251396096
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046656970912738215,
      "loss": 3.1883,
      "theoretical_loss": 4.234443733293899,
      "tokens_seen": 251461632
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046655967903711133,
      "loss": 3.2245,
      "theoretical_loss": 4.234310538754624,
      "tokens_seen": 251527168
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046654964894684057,
      "loss": 3.2696,
      "theoretical_loss": 4.2341773886292575,
      "tokens_seen": 251592704
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004665396188565697,
      "loss": 3.1524,
      "theoretical_loss": 4.234044282891429,
      "tokens_seen": 251658240
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046652958876629893,
      "loss": 3.0471,
      "theoretical_loss": 4.233911221514787,
      "tokens_seen": 251723776
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046651955867602806,
      "loss": 3.1227,
      "theoretical_loss": 4.233778204473002,
      "tokens_seen": 251789312
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004665095285857573,
      "loss": 3.2876,
      "theoretical_loss": 4.23364523173977,
      "tokens_seen": 251854848
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004664994984954865,
      "loss": 3.2065,
      "theoretical_loss": 4.233512303288807,
      "tokens_seen": 251920384
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046648946840521565,
      "loss": 2.8681,
      "theoretical_loss": 4.233379419093851,
      "tokens_seen": 251985920
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046647943831494484,
      "loss": 3.0819,
      "theoretical_loss": 4.233246579128666,
      "tokens_seen": 252051456
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000466469408224674,
      "loss": 3.1822,
      "theoretical_loss": 4.233113783367033,
      "tokens_seen": 252116992
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004664593781344032,
      "loss": 3.1466,
      "theoretical_loss": 4.232981031782761,
      "tokens_seen": 252182528
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046644934804413243,
      "loss": 3.3105,
      "theoretical_loss": 4.232848324349677,
      "tokens_seen": 252248064
    },
    {
      "epoch": 0.08,
      "objective/train/docs_used": 303567,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.1599833965301514,
      "objective/train/theoretical_loss": 4.232715661041632,
      "objective/train/tokens_used": 272773600,
      "theoretical_loss": 4.232715661041632,
      "tokens_seen": 252313600
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046643931795386156,
      "loss": 3.1736,
      "theoretical_loss": 4.232715661041632,
      "tokens_seen": 252313600
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004664292878635908,
      "loss": 3.1568,
      "theoretical_loss": 4.232583041832499,
      "tokens_seen": 252379136
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004664192577733199,
      "loss": 3.1369,
      "theoretical_loss": 4.232450466696174,
      "tokens_seen": 252444672
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046640922768304916,
      "loss": 3.0532,
      "theoretical_loss": 4.2323179356065745,
      "tokens_seen": 252510208
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046639919759277834,
      "loss": 2.9697,
      "theoretical_loss": 4.23218544853764,
      "tokens_seen": 252575744
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004663891675025075,
      "loss": 3.31,
      "theoretical_loss": 4.232053005463333,
      "tokens_seen": 252641280
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004663791374122367,
      "loss": 3.2688,
      "theoretical_loss": 4.231920606357638,
      "tokens_seen": 252706816
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046636910732196594,
      "loss": 3.0768,
      "theoretical_loss": 4.231788251194559,
      "tokens_seen": 252772352
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046635907723169506,
      "loss": 3.1853,
      "theoretical_loss": 4.231655939948127,
      "tokens_seen": 252837888
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004663490471414243,
      "loss": 3.0185,
      "theoretical_loss": 4.231523672592392,
      "tokens_seen": 252903424
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004663390170511534,
      "loss": 3.2061,
      "theoretical_loss": 4.231391449101425,
      "tokens_seen": 252968960
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046632898696088266,
      "loss": 3.0254,
      "theoretical_loss": 4.231259269449322,
      "tokens_seen": 253034496
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046631895687061184,
      "loss": 3.4604,
      "theoretical_loss": 4.231127133610198,
      "tokens_seen": 253100032
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000466308926780341,
      "loss": 3.1453,
      "theoretical_loss": 4.230995041558194,
      "tokens_seen": 253165568
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004662988966900702,
      "loss": 3.1521,
      "theoretical_loss": 4.230862993267468,
      "tokens_seen": 253231104
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004662888665997994,
      "loss": 3.2638,
      "theoretical_loss": 4.230730988712205,
      "tokens_seen": 253296640
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046627883650952857,
      "loss": 3.3752,
      "theoretical_loss": 4.230599027866606,
      "tokens_seen": 253362176
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004662688064192578,
      "loss": 3.1583,
      "theoretical_loss": 4.2304671107048994,
      "tokens_seen": 253427712
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046625877632898693,
      "loss": 3.2989,
      "theoretical_loss": 4.2303352372013325,
      "tokens_seen": 253493248
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046624874623871616,
      "loss": 3.2402,
      "theoretical_loss": 4.230203407330176,
      "tokens_seen": 253558784
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046623871614844535,
      "loss": 3.1902,
      "theoretical_loss": 4.230071621065721,
      "tokens_seen": 253624320
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004662286860581745,
      "loss": 2.866,
      "theoretical_loss": 4.2299398783822815,
      "tokens_seen": 253689856
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004662186559679037,
      "loss": 3.2554,
      "theoretical_loss": 4.229808179254192,
      "tokens_seen": 253755392
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004662086258776329,
      "loss": 3.2293,
      "theoretical_loss": 4.22967652365581,
      "tokens_seen": 253820928
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004661985957873621,
      "loss": 2.9234,
      "theoretical_loss": 4.229544911561513,
      "tokens_seen": 253886464
    },
    {
      "epoch": 0.08,
      "objective/train/docs_used": 304269,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.1997804641723633,
      "objective/train/theoretical_loss": 4.229413342945703,
      "objective/train/tokens_used": 274412000,
      "theoretical_loss": 4.229413342945703,
      "tokens_seen": 253952000
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004661885656970913,
      "loss": 3.353,
      "theoretical_loss": 4.229413342945703,
      "tokens_seen": 253952000
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004661785356068205,
      "loss": 3.1267,
      "theoretical_loss": 4.229281817782801,
      "tokens_seen": 254017536
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046616850551654967,
      "loss": 3.2084,
      "theoretical_loss": 4.229150336047251,
      "tokens_seen": 254083072
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046615847542627885,
      "loss": 3.3788,
      "theoretical_loss": 4.229018897713519,
      "tokens_seen": 254148608
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046614844533600803,
      "loss": 3.0708,
      "theoretical_loss": 4.22888750275609,
      "tokens_seen": 254214144
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046613841524573727,
      "loss": 3.0028,
      "theoretical_loss": 4.228756151149475,
      "tokens_seen": 254279680
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004661283851554664,
      "loss": 3.3021,
      "theoretical_loss": 4.228624842868202,
      "tokens_seen": 254345216
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046611835506519563,
      "loss": 3.0294,
      "theoretical_loss": 4.228493577886824,
      "tokens_seen": 254410752
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046610832497492475,
      "loss": 3.1577,
      "theoretical_loss": 4.228362356179913,
      "tokens_seen": 254476288
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000466098294884654,
      "loss": 3.0022,
      "theoretical_loss": 4.228231177722063,
      "tokens_seen": 254541824
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046608826479438317,
      "loss": 3.0895,
      "theoretical_loss": 4.228100042487892,
      "tokens_seen": 254607360
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046607823470411235,
      "loss": 3.1647,
      "theoretical_loss": 4.227968950452035,
      "tokens_seen": 254672896
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046606820461384153,
      "loss": 3.2423,
      "theoretical_loss": 4.227837901589153,
      "tokens_seen": 254738432
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046605817452357077,
      "loss": 3.056,
      "theoretical_loss": 4.227706895873924,
      "tokens_seen": 254803968
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004660481444332999,
      "loss": 2.8991,
      "theoretical_loss": 4.227575933281051,
      "tokens_seen": 254869504
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046603811434302913,
      "loss": 3.0,
      "theoretical_loss": 4.227445013785257,
      "tokens_seen": 254935040
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046602808425275826,
      "loss": 3.0826,
      "theoretical_loss": 4.227314137361285,
      "tokens_seen": 255000576
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004660180541624875,
      "loss": 3.2037,
      "theoretical_loss": 4.227183303983901,
      "tokens_seen": 255066112
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004660080240722167,
      "loss": 3.2881,
      "theoretical_loss": 4.227052513627893,
      "tokens_seen": 255131648
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046599799398194586,
      "loss": 3.1299,
      "theoretical_loss": 4.226921766268067,
      "tokens_seen": 255197184
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046598796389167504,
      "loss": 3.0684,
      "theoretical_loss": 4.226791061879253,
      "tokens_seen": 255262720
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004659779338014042,
      "loss": 3.0469,
      "theoretical_loss": 4.226660400436302,
      "tokens_seen": 255328256
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004659679037111334,
      "loss": 3.0472,
      "theoretical_loss": 4.226529781914084,
      "tokens_seen": 255393792
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046595787362086263,
      "loss": 3.2079,
      "theoretical_loss": 4.226399206287493,
      "tokens_seen": 255459328
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046594784353059176,
      "loss": 3.1358,
      "theoretical_loss": 4.226268673531442,
      "tokens_seen": 255524864
    },
    {
      "epoch": 0.08,
      "objective/train/docs_used": 305849,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.743330240249634,
      "objective/train/theoretical_loss": 4.226138183620867,
      "objective/train/tokens_used": 276050400,
      "theoretical_loss": 4.226138183620867,
      "tokens_seen": 255590400
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000465937813440321,
      "loss": 3.1548,
      "theoretical_loss": 4.226138183620867,
      "tokens_seen": 255590400
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004659277833500501,
      "loss": 3.0213,
      "theoretical_loss": 4.226007736530723,
      "tokens_seen": 255655936
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046591775325977936,
      "loss": 3.1984,
      "theoretical_loss": 4.225877332235987,
      "tokens_seen": 255721472
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046590772316950854,
      "loss": 3.1091,
      "theoretical_loss": 4.225746970711657,
      "tokens_seen": 255787008
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004658976930792377,
      "loss": 3.2381,
      "theoretical_loss": 4.225616651932753,
      "tokens_seen": 255852544
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004658876629889669,
      "loss": 3.2222,
      "theoretical_loss": 4.225486375874315,
      "tokens_seen": 255918080
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046587763289869614,
      "loss": 3.1889,
      "theoretical_loss": 4.225356142511402,
      "tokens_seen": 255983616
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046586760280842526,
      "loss": 3.0673,
      "theoretical_loss": 4.225225951819099,
      "tokens_seen": 256049152
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004658575727181545,
      "loss": 3.1606,
      "theoretical_loss": 4.225095803772507,
      "tokens_seen": 256114688
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004658475426278836,
      "loss": 3.2571,
      "theoretical_loss": 4.22496569834675,
      "tokens_seen": 256180224
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046583751253761286,
      "loss": 3.3425,
      "theoretical_loss": 4.224835635516973,
      "tokens_seen": 256245760
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046582748244734204,
      "loss": 3.0092,
      "theoretical_loss": 4.224705615258341,
      "tokens_seen": 256311296
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004658174523570712,
      "loss": 3.2458,
      "theoretical_loss": 4.224575637546041,
      "tokens_seen": 256376832
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004658074222668004,
      "loss": 2.9137,
      "theoretical_loss": 4.224445702355279,
      "tokens_seen": 256442368
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004657973921765296,
      "loss": 3.1672,
      "theoretical_loss": 4.2243158096612845,
      "tokens_seen": 256507904
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046578736208625877,
      "loss": 3.1599,
      "theoretical_loss": 4.224185959439305,
      "tokens_seen": 256573440
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000465777331995988,
      "loss": 3.1255,
      "theoretical_loss": 4.22405615166461,
      "tokens_seen": 256638976
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046576730190571713,
      "loss": 3.1789,
      "theoretical_loss": 4.22392638631249,
      "tokens_seen": 256704512
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046575727181544636,
      "loss": 3.1087,
      "theoretical_loss": 4.223796663358255,
      "tokens_seen": 256770048
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046574724172517555,
      "loss": 3.2464,
      "theoretical_loss": 4.223666982777237,
      "tokens_seen": 256835584
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004657372116349047,
      "loss": 3.327,
      "theoretical_loss": 4.223537344544788,
      "tokens_seen": 256901120
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004657271815446339,
      "loss": 3.1453,
      "theoretical_loss": 4.223407748636282,
      "tokens_seen": 256966656
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004657171514543631,
      "loss": 3.0814,
      "theoretical_loss": 4.22327819502711,
      "tokens_seen": 257032192
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046570712136409227,
      "loss": 3.1326,
      "theoretical_loss": 4.223148683692687,
      "tokens_seen": 257097728
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004656970912738215,
      "loss": 3.2236,
      "theoretical_loss": 4.223019214608446,
      "tokens_seen": 257163264
    },
    {
      "epoch": 0.08,
      "objective/train/docs_used": 306364,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.629127264022827,
      "objective/train/theoretical_loss": 4.222889787749845,
      "objective/train/tokens_used": 277688800,
      "theoretical_loss": 4.222889787749845,
      "tokens_seen": 257228800
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046568706118355063,
      "loss": 3.0022,
      "theoretical_loss": 4.222889787749845,
      "tokens_seen": 257228800
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046567703109327987,
      "loss": 3.2723,
      "theoretical_loss": 4.222760403092358,
      "tokens_seen": 257294336
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000465667001003009,
      "loss": 3.1032,
      "theoretical_loss": 4.22263106061148,
      "tokens_seen": 257359872
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046565697091273823,
      "loss": 3.2668,
      "theoretical_loss": 4.222501760282729,
      "tokens_seen": 257425408
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004656469408224674,
      "loss": 3.3694,
      "theoretical_loss": 4.22237250208164,
      "tokens_seen": 257490944
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004656369107321966,
      "loss": 3.0323,
      "theoretical_loss": 4.222243285983772,
      "tokens_seen": 257556480
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004656268806419258,
      "loss": 3.0638,
      "theoretical_loss": 4.222114111964703,
      "tokens_seen": 257622016
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046561685055165495,
      "loss": 3.2751,
      "theoretical_loss": 4.221984980000029,
      "tokens_seen": 257687552
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046560682046138414,
      "loss": 3.1497,
      "theoretical_loss": 4.2218558900653695,
      "tokens_seen": 257753088
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046559679037111337,
      "loss": 3.2393,
      "theoretical_loss": 4.221726842136364,
      "tokens_seen": 257818624
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004655867602808425,
      "loss": 3.3169,
      "theoretical_loss": 4.2215978361886695,
      "tokens_seen": 257884160
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046557673019057173,
      "loss": 3.1882,
      "theoretical_loss": 4.221468872197967,
      "tokens_seen": 257949696
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004655667001003009,
      "loss": 3.0225,
      "theoretical_loss": 4.221339950139956,
      "tokens_seen": 258015232
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004655566700100301,
      "loss": 3.1899,
      "theoretical_loss": 4.221211069990357,
      "tokens_seen": 258080768
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004655466399197593,
      "loss": 3.2241,
      "theoretical_loss": 4.221082231724908,
      "tokens_seen": 258146304
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046553660982948846,
      "loss": 3.1808,
      "theoretical_loss": 4.22095343531937,
      "tokens_seen": 258211840
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046552657973921764,
      "loss": 3.4274,
      "theoretical_loss": 4.220824680749525,
      "tokens_seen": 258277376
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004655165496489469,
      "loss": 3.0687,
      "theoretical_loss": 4.220695967991171,
      "tokens_seen": 258342912
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000465506519558676,
      "loss": 3.1397,
      "theoretical_loss": 4.220567297020131,
      "tokens_seen": 258408448
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046549648946840524,
      "loss": 3.1805,
      "theoretical_loss": 4.220438667812244,
      "tokens_seen": 258473984
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046548645937813436,
      "loss": 3.1492,
      "theoretical_loss": 4.220310080343373,
      "tokens_seen": 258539520
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004654764292878636,
      "loss": 2.7064,
      "theoretical_loss": 4.220181534589398,
      "tokens_seen": 258605056
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004654663991975928,
      "loss": 3.0719,
      "theoretical_loss": 4.22005303052622,
      "tokens_seen": 258670592
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046545636910732196,
      "loss": 3.2042,
      "theoretical_loss": 4.219924568129759,
      "tokens_seen": 258736128
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004654463390170512,
      "loss": 3.0768,
      "theoretical_loss": 4.219796147375957,
      "tokens_seen": 258801664
    },
    {
      "epoch": 0.08,
      "objective/train/docs_used": 307092,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.476236343383789,
      "objective/train/theoretical_loss": 4.219667768240775,
      "objective/train/tokens_used": 279327200,
      "theoretical_loss": 4.219667768240775,
      "tokens_seen": 258867200
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004654363089267803,
      "loss": 3.1775,
      "theoretical_loss": 4.219667768240775,
      "tokens_seen": 258867200
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046542627883650956,
      "loss": 2.9703,
      "theoretical_loss": 4.219539430700195,
      "tokens_seen": 258932736
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046541624874623874,
      "loss": 3.0814,
      "theoretical_loss": 4.2194111347302155,
      "tokens_seen": 258998272
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004654062186559679,
      "loss": 3.124,
      "theoretical_loss": 4.219282880306859,
      "tokens_seen": 259063808
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004653961885656971,
      "loss": 2.9946,
      "theoretical_loss": 4.219154667406166,
      "tokens_seen": 259129344
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046538615847542634,
      "loss": 3.1049,
      "theoretical_loss": 4.219026496004198,
      "tokens_seen": 259194880
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046537612838515546,
      "loss": 3.06,
      "theoretical_loss": 4.218898366077035,
      "tokens_seen": 259260416
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004653660982948847,
      "loss": 3.0953,
      "theoretical_loss": 4.218770277600775,
      "tokens_seen": 259325952
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004653560682046138,
      "loss": 2.8445,
      "theoretical_loss": 4.218642230551541,
      "tokens_seen": 259391488
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046534603811434306,
      "loss": 3.0274,
      "theoretical_loss": 4.218514224905472,
      "tokens_seen": 259457024
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046533600802407224,
      "loss": 2.9485,
      "theoretical_loss": 4.218386260638727,
      "tokens_seen": 259522560
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004653259779338014,
      "loss": 3.1876,
      "theoretical_loss": 4.2182583377274865,
      "tokens_seen": 259588096
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004653159478435306,
      "loss": 3.1376,
      "theoretical_loss": 4.218130456147948,
      "tokens_seen": 259653632
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004653059177532598,
      "loss": 3.1214,
      "theoretical_loss": 4.218002615876332,
      "tokens_seen": 259719168
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046529588766298897,
      "loss": 3.2295,
      "theoretical_loss": 4.217874816888877,
      "tokens_seen": 259784704
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004652858575727182,
      "loss": 3.0723,
      "theoretical_loss": 4.217747059161839,
      "tokens_seen": 259850240
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046527582748244733,
      "loss": 2.9696,
      "theoretical_loss": 4.217619342671498,
      "tokens_seen": 259915776
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046526579739217656,
      "loss": 3.0124,
      "theoretical_loss": 4.2174916673941505,
      "tokens_seen": 259981312
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046525576730190575,
      "loss": 2.8966,
      "theoretical_loss": 4.217364033306113,
      "tokens_seen": 260046848
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004652457372116349,
      "loss": 3.1619,
      "theoretical_loss": 4.217236440383724,
      "tokens_seen": 260112384
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004652357071213641,
      "loss": 2.9974,
      "theoretical_loss": 4.217108888603337,
      "tokens_seen": 260177920
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004652256770310933,
      "loss": 2.9334,
      "theoretical_loss": 4.21698137794133,
      "tokens_seen": 260243456
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046521564694082247,
      "loss": 3.2643,
      "theoretical_loss": 4.216853908374097,
      "tokens_seen": 260308992
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004652056168505517,
      "loss": 3.3784,
      "theoretical_loss": 4.216726479878052,
      "tokens_seen": 260374528
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046519558676028083,
      "loss": 3.1801,
      "theoretical_loss": 4.216599092429631,
      "tokens_seen": 260440064
    },
    {
      "epoch": 0.08,
      "objective/train/docs_used": 308211,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.457490921020508,
      "objective/train/theoretical_loss": 4.216471746005286,
      "objective/train/tokens_used": 280965600,
      "theoretical_loss": 4.216471746005286,
      "tokens_seen": 260505600
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046518555667001007,
      "loss": 3.1286,
      "theoretical_loss": 4.216471746005286,
      "tokens_seen": 260505600
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004651755265797392,
      "loss": 2.9917,
      "theoretical_loss": 4.216344440581491,
      "tokens_seen": 260571136
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046516549648946843,
      "loss": 3.0801,
      "theoretical_loss": 4.2162171761347365,
      "tokens_seen": 260636672
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004651554663991976,
      "loss": 3.1188,
      "theoretical_loss": 4.2160899526415365,
      "tokens_seen": 260702208
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004651454363089268,
      "loss": 3.0944,
      "theoretical_loss": 4.215962770078422,
      "tokens_seen": 260767744
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000465135406218656,
      "loss": 3.3147,
      "theoretical_loss": 4.215835628421942,
      "tokens_seen": 260833280
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046512537612838515,
      "loss": 3.1019,
      "theoretical_loss": 4.215708527648667,
      "tokens_seen": 260898816
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046511534603811434,
      "loss": 2.8503,
      "theoretical_loss": 4.215581467735187,
      "tokens_seen": 260964352
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046510531594784357,
      "loss": 3.2366,
      "theoretical_loss": 4.215454448658109,
      "tokens_seen": 261029888
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004650952858575727,
      "loss": 3.1674,
      "theoretical_loss": 4.215327470394062,
      "tokens_seen": 261095424
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046508525576730193,
      "loss": 3.1104,
      "theoretical_loss": 4.215200532919691,
      "tokens_seen": 261160960
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004650752256770311,
      "loss": 3.197,
      "theoretical_loss": 4.215073636211664,
      "tokens_seen": 261226496
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004650651955867603,
      "loss": 3.1214,
      "theoretical_loss": 4.214946780246666,
      "tokens_seen": 261292032
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004650551654964895,
      "loss": 3.0783,
      "theoretical_loss": 4.214819965001401,
      "tokens_seen": 261357568
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046504513540621866,
      "loss": 3.0267,
      "theoretical_loss": 4.214693190452593,
      "tokens_seen": 261423104
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046503510531594784,
      "loss": 3.0135,
      "theoretical_loss": 4.214566456576984,
      "tokens_seen": 261488640
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004650250752256771,
      "loss": 2.9927,
      "theoretical_loss": 4.214439763351336,
      "tokens_seen": 261554176
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004650150451354062,
      "loss": 2.9392,
      "theoretical_loss": 4.214313110752431,
      "tokens_seen": 261619712
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046500501504513544,
      "loss": 2.9189,
      "theoretical_loss": 4.214186498757069,
      "tokens_seen": 261685248
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046499498495486456,
      "loss": 3.0795,
      "theoretical_loss": 4.214059927342068,
      "tokens_seen": 261750784
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004649849548645938,
      "loss": 3.0373,
      "theoretical_loss": 4.213933396484267,
      "tokens_seen": 261816320
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000464974924774323,
      "loss": 3.1087,
      "theoretical_loss": 4.213806906160523,
      "tokens_seen": 261881856
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046496489468405216,
      "loss": 3.2205,
      "theoretical_loss": 4.213680456347712,
      "tokens_seen": 261947392
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046495486459378134,
      "loss": 3.438,
      "theoretical_loss": 4.213554047022729,
      "tokens_seen": 262012928
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004649448345035105,
      "loss": 3.1792,
      "theoretical_loss": 4.213427678162489,
      "tokens_seen": 262078464
    },
    {
      "epoch": 0.08,
      "objective/train/docs_used": 308640,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.578385353088379,
      "objective/train/theoretical_loss": 4.213301349743924,
      "objective/train/tokens_used": 282604000,
      "theoretical_loss": 4.213301349743924,
      "tokens_seen": 262144000
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004649348044132397,
      "loss": 3.2087,
      "theoretical_loss": 4.213301349743924,
      "tokens_seen": 262144000
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046492477432296894,
      "loss": 3.0893,
      "theoretical_loss": 4.2131750617439865,
      "tokens_seen": 262209536
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046491474423269807,
      "loss": 3.1618,
      "theoretical_loss": 4.213048814139647,
      "tokens_seen": 262275072
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004649047141424273,
      "loss": 2.9118,
      "theoretical_loss": 4.212922606907895,
      "tokens_seen": 262340608
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004648946840521565,
      "loss": 3.2177,
      "theoretical_loss": 4.21279644002574,
      "tokens_seen": 262406144
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046488465396188566,
      "loss": 3.1045,
      "theoretical_loss": 4.212670313470209,
      "tokens_seen": 262471680
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046487462387161485,
      "loss": 3.0042,
      "theoretical_loss": 4.212544227218347,
      "tokens_seen": 262537216
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000464864593781344,
      "loss": 3.0721,
      "theoretical_loss": 4.21241818124722,
      "tokens_seen": 262602752
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004648545636910732,
      "loss": 3.1759,
      "theoretical_loss": 4.212292175533912,
      "tokens_seen": 262668288
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046484453360080244,
      "loss": 3.2897,
      "theoretical_loss": 4.212166210055526,
      "tokens_seen": 262733824
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046483450351053157,
      "loss": 3.1411,
      "theoretical_loss": 4.212040284789181,
      "tokens_seen": 262799360
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004648244734202608,
      "loss": 3.0484,
      "theoretical_loss": 4.211914399712019,
      "tokens_seen": 262864896
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046481444332998993,
      "loss": 3.0421,
      "theoretical_loss": 4.211788554801198,
      "tokens_seen": 262930432
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046480441323971917,
      "loss": 3.1647,
      "theoretical_loss": 4.211662750033895,
      "tokens_seen": 262995968
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046479438314944835,
      "loss": 2.9247,
      "theoretical_loss": 4.211536985387307,
      "tokens_seen": 263061504
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046478435305917753,
      "loss": 3.1853,
      "theoretical_loss": 4.211411260838647,
      "tokens_seen": 263127040
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004647743229689067,
      "loss": 3.0719,
      "theoretical_loss": 4.2112855763651496,
      "tokens_seen": 263192576
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046476429287863595,
      "loss": 3.1743,
      "theoretical_loss": 4.211159931944065,
      "tokens_seen": 263258112
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004647542627883651,
      "loss": 3.3044,
      "theoretical_loss": 4.211034327552666,
      "tokens_seen": 263323648
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004647442326980943,
      "loss": 3.172,
      "theoretical_loss": 4.210908763168239,
      "tokens_seen": 263389184
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046473420260782344,
      "loss": 2.8232,
      "theoretical_loss": 4.210783238768093,
      "tokens_seen": 263454720
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046472417251755267,
      "loss": 2.8838,
      "theoretical_loss": 4.210657754329553,
      "tokens_seen": 263520256
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004647141424272819,
      "loss": 3.1246,
      "theoretical_loss": 4.210532309829965,
      "tokens_seen": 263585792
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046470411233701103,
      "loss": 3.0826,
      "theoretical_loss": 4.21040690524669,
      "tokens_seen": 263651328
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046469408224674027,
      "loss": 3.0629,
      "theoretical_loss": 4.21028154055711,
      "tokens_seen": 263716864
    },
    {
      "epoch": 0.08,
      "objective/train/docs_used": 309875,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5037975311279297,
      "objective/train/theoretical_loss": 4.2101562157386265,
      "objective/train/tokens_used": 284242400,
      "theoretical_loss": 4.2101562157386265,
      "tokens_seen": 263782400
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004646840521564694,
      "loss": 2.8095,
      "theoretical_loss": 4.2101562157386265,
      "tokens_seen": 263782400
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046467402206619863,
      "loss": 3.2129,
      "theoretical_loss": 4.210030930768655,
      "tokens_seen": 263847936
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004646639919759278,
      "loss": 3.3535,
      "theoretical_loss": 4.2099056856246335,
      "tokens_seen": 263913472
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000464653961885657,
      "loss": 3.0754,
      "theoretical_loss": 4.209780480284017,
      "tokens_seen": 263979008
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004646439317953862,
      "loss": 3.2122,
      "theoretical_loss": 4.209655314724279,
      "tokens_seen": 264044544
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046463390170511535,
      "loss": 2.9998,
      "theoretical_loss": 4.209530188922911,
      "tokens_seen": 264110080
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046462387161484454,
      "loss": 3.2189,
      "theoretical_loss": 4.209405102857422,
      "tokens_seen": 264175616
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046461384152457377,
      "loss": 3.1448,
      "theoretical_loss": 4.209280056505342,
      "tokens_seen": 264241152
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004646038114343029,
      "loss": 3.1553,
      "theoretical_loss": 4.209155049844217,
      "tokens_seen": 264306688
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046459378134403213,
      "loss": 3.0602,
      "theoretical_loss": 4.209030082851612,
      "tokens_seen": 264372224
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004645837512537613,
      "loss": 3.1091,
      "theoretical_loss": 4.208905155505109,
      "tokens_seen": 264437760
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004645737211634905,
      "loss": 2.961,
      "theoretical_loss": 4.20878026778231,
      "tokens_seen": 264503296
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004645636910732197,
      "loss": 2.9389,
      "theoretical_loss": 4.208655419660834,
      "tokens_seen": 264568832
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046455366098294886,
      "loss": 3.2368,
      "theoretical_loss": 4.208530611118321,
      "tokens_seen": 264634368
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046454363089267804,
      "loss": 3.2025,
      "theoretical_loss": 4.208405842132423,
      "tokens_seen": 264699904
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004645336008024073,
      "loss": 3.3101,
      "theoretical_loss": 4.208281112680817,
      "tokens_seen": 264765440
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004645235707121364,
      "loss": 2.9905,
      "theoretical_loss": 4.208156422741195,
      "tokens_seen": 264830976
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046451354062186564,
      "loss": 3.3088,
      "theoretical_loss": 4.208031772291265,
      "tokens_seen": 264896512
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046450351053159476,
      "loss": 3.1124,
      "theoretical_loss": 4.207907161308757,
      "tokens_seen": 264962048
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000464493480441324,
      "loss": 3.011,
      "theoretical_loss": 4.2077825897714165,
      "tokens_seen": 265027584
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004644834503510532,
      "loss": 3.1281,
      "theoretical_loss": 4.207658057657008,
      "tokens_seen": 265093120
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046447342026078236,
      "loss": 3.2965,
      "theoretical_loss": 4.207533564943316,
      "tokens_seen": 265158656
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046446339017051154,
      "loss": 3.0647,
      "theoretical_loss": 4.207409111608138,
      "tokens_seen": 265224192
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004644533600802407,
      "loss": 2.9901,
      "theoretical_loss": 4.2072846976292935,
      "tokens_seen": 265289728
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004644433299899699,
      "loss": 3.0768,
      "theoretical_loss": 4.2071603229846195,
      "tokens_seen": 265355264
    },
    {
      "epoch": 0.08,
      "objective/train/docs_used": 311046,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.1395962238311768,
      "objective/train/theoretical_loss": 4.20703598765197,
      "objective/train/tokens_used": 285880800,
      "theoretical_loss": 4.20703598765197,
      "tokens_seen": 265420800
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046443329989969914,
      "loss": 3.1445,
      "theoretical_loss": 4.20703598765197,
      "tokens_seen": 265420800
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046442326980942827,
      "loss": 3.0687,
      "theoretical_loss": 4.206911691609217,
      "tokens_seen": 265486336
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004644132397191575,
      "loss": 2.8642,
      "theoretical_loss": 4.206787434834251,
      "tokens_seen": 265551872
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004644032096288867,
      "loss": 3.3262,
      "theoretical_loss": 4.20666321730498,
      "tokens_seen": 265617408
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046439317953861586,
      "loss": 3.095,
      "theoretical_loss": 4.206539038999329,
      "tokens_seen": 265682944
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046438314944834505,
      "loss": 2.9848,
      "theoretical_loss": 4.206414899895244,
      "tokens_seen": 265748480
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004643731193580742,
      "loss": 3.1623,
      "theoretical_loss": 4.206290799970685,
      "tokens_seen": 265814016
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004643630892678034,
      "loss": 3.1056,
      "theoretical_loss": 4.206166739203632,
      "tokens_seen": 265879552
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046435305917753264,
      "loss": 3.2523,
      "theoretical_loss": 4.206042717572082,
      "tokens_seen": 265945088
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046434302908726177,
      "loss": 2.9739,
      "theoretical_loss": 4.20591873505405,
      "tokens_seen": 266010624
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000464332998996991,
      "loss": 3.0009,
      "theoretical_loss": 4.20579479162757,
      "tokens_seen": 266076160
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046432296890672013,
      "loss": 3.1793,
      "theoretical_loss": 4.205670887270691,
      "tokens_seen": 266141696
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046431293881644937,
      "loss": 3.0394,
      "theoretical_loss": 4.205547021961482,
      "tokens_seen": 266207232
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046430290872617855,
      "loss": 3.04,
      "theoretical_loss": 4.205423195678029,
      "tokens_seen": 266272768
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046429287863590773,
      "loss": 2.7836,
      "theoretical_loss": 4.205299408398435,
      "tokens_seen": 266338304
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004642828485456369,
      "loss": 3.4339,
      "theoretical_loss": 4.2051756601008226,
      "tokens_seen": 266403840
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046427281845536615,
      "loss": 3.2731,
      "theoretical_loss": 4.20505195076333,
      "tokens_seen": 266469376
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004642627883650953,
      "loss": 2.9399,
      "theoretical_loss": 4.204928280364115,
      "tokens_seen": 266534912
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004642527582748245,
      "loss": 2.6865,
      "theoretical_loss": 4.20480464888135,
      "tokens_seen": 266600448
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046424272818455364,
      "loss": 3.1693,
      "theoretical_loss": 4.204681056293228,
      "tokens_seen": 266665984
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046423269809428287,
      "loss": 3.178,
      "theoretical_loss": 4.204557502577957,
      "tokens_seen": 266731520
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046422266800401205,
      "loss": 3.0357,
      "theoretical_loss": 4.204433987713767,
      "tokens_seen": 266797056
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046421263791374123,
      "loss": 3.1383,
      "theoretical_loss": 4.2043105116789,
      "tokens_seen": 266862592
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004642026078234704,
      "loss": 3.1161,
      "theoretical_loss": 4.204187074451617,
      "tokens_seen": 266928128
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004641925777331996,
      "loss": 3.0181,
      "theoretical_loss": 4.204063676010202,
      "tokens_seen": 266993664
    },
    {
      "epoch": 0.08,
      "objective/train/docs_used": 311732,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.2471554279327393,
      "objective/train/theoretical_loss": 4.203940316332948,
      "objective/train/tokens_used": 287519200,
      "theoretical_loss": 4.203940316332948,
      "tokens_seen": 267059200
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004641825476429288,
      "loss": 3.2104,
      "theoretical_loss": 4.203940316332948,
      "tokens_seen": 267059200
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000464172517552658,
      "loss": 3.0556,
      "theoretical_loss": 4.203816995398171,
      "tokens_seen": 267124736
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046416248746238714,
      "loss": 2.944,
      "theoretical_loss": 4.203693713184203,
      "tokens_seen": 267190272
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004641524573721164,
      "loss": 3.212,
      "theoretical_loss": 4.203570469669392,
      "tokens_seen": 267255808
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004641424272818455,
      "loss": 3.2094,
      "theoretical_loss": 4.203447264832107,
      "tokens_seen": 267321344
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046413239719157474,
      "loss": 3.4161,
      "theoretical_loss": 4.203324098650731,
      "tokens_seen": 267386880
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004641223671013039,
      "loss": 3.1592,
      "theoretical_loss": 4.203200971103666,
      "tokens_seen": 267452416
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004641123370110331,
      "loss": 3.1786,
      "theoretical_loss": 4.20307788216933,
      "tokens_seen": 267517952
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004641023069207623,
      "loss": 3.2507,
      "theoretical_loss": 4.202954831826159,
      "tokens_seen": 267583488
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004640922768304915,
      "loss": 3.132,
      "theoretical_loss": 4.202831820052609,
      "tokens_seen": 267649024
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046408224674022064,
      "loss": 3.032,
      "theoretical_loss": 4.202708846827148,
      "tokens_seen": 267714560
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004640722166499499,
      "loss": 2.9754,
      "theoretical_loss": 4.202585912128266,
      "tokens_seen": 267780096
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000464062186559679,
      "loss": 3.0707,
      "theoretical_loss": 4.202463015934468,
      "tokens_seen": 267845632
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046405215646940824,
      "loss": 2.8317,
      "theoretical_loss": 4.202340158224277,
      "tokens_seen": 267911168
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004640421263791374,
      "loss": 3.2203,
      "theoretical_loss": 4.202217338976231,
      "tokens_seen": 267976704
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004640320962888666,
      "loss": 2.7621,
      "theoretical_loss": 4.2020945581688895,
      "tokens_seen": 268042240
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004640220661985958,
      "loss": 2.9985,
      "theoretical_loss": 4.201971815780826,
      "tokens_seen": 268107776
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046401203610832496,
      "loss": 3.1052,
      "theoretical_loss": 4.201849111790631,
      "tokens_seen": 268173312
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046400200601805414,
      "loss": 3.1588,
      "theoretical_loss": 4.201726446176915,
      "tokens_seen": 268238848
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004639919759277834,
      "loss": 3.1762,
      "theoretical_loss": 4.201603818918302,
      "tokens_seen": 268304384
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004639819458375125,
      "loss": 2.9964,
      "theoretical_loss": 4.201481229993435,
      "tokens_seen": 268369920
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046397191574724174,
      "loss": 2.9145,
      "theoretical_loss": 4.201358679380976,
      "tokens_seen": 268435456
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004639618856569709,
      "loss": 3.1938,
      "theoretical_loss": 4.201236167059601,
      "tokens_seen": 268500992
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004639518555667001,
      "loss": 2.9321,
      "theoretical_loss": 4.201113693008002,
      "tokens_seen": 268566528
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046394182547642934,
      "loss": 2.843,
      "theoretical_loss": 4.200991257204894,
      "tokens_seen": 268632064
    },
    {
      "epoch": 0.08,
      "objective/train/docs_used": 312511,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.1893441677093506,
      "objective/train/theoretical_loss": 4.2008688596290025,
      "objective/train/tokens_used": 289157600,
      "theoretical_loss": 4.2008688596290025,
      "tokens_seen": 268697600
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046393179538615847,
      "loss": 3.3237,
      "theoretical_loss": 4.2008688596290025,
      "tokens_seen": 268697600
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004639217652958877,
      "loss": 3.1753,
      "theoretical_loss": 4.200746500259073,
      "tokens_seen": 268763136
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004639117352056169,
      "loss": 2.9467,
      "theoretical_loss": 4.200624179073869,
      "tokens_seen": 268828672
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046390170511534606,
      "loss": 3.0959,
      "theoretical_loss": 4.2005018960521685,
      "tokens_seen": 268894208
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046389167502507525,
      "loss": 3.1224,
      "theoretical_loss": 4.200379651172769,
      "tokens_seen": 268959744
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004638816449348044,
      "loss": 3.1511,
      "theoretical_loss": 4.200257444414483,
      "tokens_seen": 269025280
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004638716148445336,
      "loss": 3.2735,
      "theoretical_loss": 4.200135275756139,
      "tokens_seen": 269090816
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046386158475426284,
      "loss": 3.0709,
      "theoretical_loss": 4.200013145176587,
      "tokens_seen": 269156352
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046385155466399197,
      "loss": 3.3529,
      "theoretical_loss": 4.199891052654689,
      "tokens_seen": 269221888
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004638415245737212,
      "loss": 3.1699,
      "theoretical_loss": 4.199768998169326,
      "tokens_seen": 269287424
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046383149448345033,
      "loss": 3.1224,
      "theoretical_loss": 4.199646981699395,
      "tokens_seen": 269352960
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046382146439317957,
      "loss": 3.1028,
      "theoretical_loss": 4.199525003223812,
      "tokens_seen": 269418496
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046381143430290875,
      "loss": 2.9704,
      "theoretical_loss": 4.199403062721506,
      "tokens_seen": 269484032
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046380140421263793,
      "loss": 3.1699,
      "theoretical_loss": 4.199281160171427,
      "tokens_seen": 269549568
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004637913741223671,
      "loss": 2.773,
      "theoretical_loss": 4.1991592955525405,
      "tokens_seen": 269615104
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046378134403209635,
      "loss": 3.1007,
      "theoretical_loss": 4.199037468843825,
      "tokens_seen": 269680640
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004637713139418255,
      "loss": 3.3052,
      "theoretical_loss": 4.198915680024282,
      "tokens_seen": 269746176
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004637612838515547,
      "loss": 2.9844,
      "theoretical_loss": 4.198793929072925,
      "tokens_seen": 269811712
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046375125376128384,
      "loss": 3.1632,
      "theoretical_loss": 4.198672215968785,
      "tokens_seen": 269877248
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046374122367101307,
      "loss": 3.0686,
      "theoretical_loss": 4.198550540690912,
      "tokens_seen": 269942784
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046373119358074225,
      "loss": 2.9668,
      "theoretical_loss": 4.198428903218371,
      "tokens_seen": 270008320
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046372116349047143,
      "loss": 2.9512,
      "theoretical_loss": 4.198307303530243,
      "tokens_seen": 270073856
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004637111334002006,
      "loss": 3.3384,
      "theoretical_loss": 4.198185741605628,
      "tokens_seen": 270139392
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004637011033099298,
      "loss": 3.2329,
      "theoretical_loss": 4.19806421742364,
      "tokens_seen": 270204928
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000463691073219659,
      "loss": 2.7934,
      "theoretical_loss": 4.197942730963412,
      "tokens_seen": 270270464
    },
    {
      "epoch": 0.08,
      "objective/train/docs_used": 313588,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.1681175231933594,
      "objective/train/theoretical_loss": 4.19782128220409,
      "objective/train/tokens_used": 290796000,
      "theoretical_loss": 4.19782128220409,
      "tokens_seen": 270336000
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004636810431293882,
      "loss": 3.1274,
      "theoretical_loss": 4.19782128220409,
      "tokens_seen": 270336000
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046367101303911734,
      "loss": 3.231,
      "theoretical_loss": 4.19769987112484,
      "tokens_seen": 270401536
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004636609829488466,
      "loss": 2.9321,
      "theoretical_loss": 4.1975784977048445,
      "tokens_seen": 270467072
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004636509528585757,
      "loss": 3.1377,
      "theoretical_loss": 4.1974571619233,
      "tokens_seen": 270532608
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046364092276830494,
      "loss": 2.8866,
      "theoretical_loss": 4.197335863759422,
      "tokens_seen": 270598144
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004636308926780341,
      "loss": 3.0662,
      "theoretical_loss": 4.1972146031924416,
      "tokens_seen": 270663680
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004636208625877633,
      "loss": 2.9822,
      "theoretical_loss": 4.197093380201606,
      "tokens_seen": 270729216
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004636108324974925,
      "loss": 3.1604,
      "theoretical_loss": 4.196972194766179,
      "tokens_seen": 270794752
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004636008024072217,
      "loss": 2.7832,
      "theoretical_loss": 4.196851046865442,
      "tokens_seen": 270860288
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046359077231695084,
      "loss": 3.0355,
      "theoretical_loss": 4.1967299364786905,
      "tokens_seen": 270925824
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004635807422266801,
      "loss": 3.1407,
      "theoretical_loss": 4.196608863585239,
      "tokens_seen": 270991360
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004635707121364092,
      "loss": 2.861,
      "theoretical_loss": 4.1964878281644165,
      "tokens_seen": 271056896
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046356068204613844,
      "loss": 3.2408,
      "theoretical_loss": 4.19636683019557,
      "tokens_seen": 271122432
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004635506519558676,
      "loss": 3.0655,
      "theoretical_loss": 4.196245869658061,
      "tokens_seen": 271187968
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004635406218655968,
      "loss": 2.9336,
      "theoretical_loss": 4.1961249465312696,
      "tokens_seen": 271253504
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000463530591775326,
      "loss": 2.9904,
      "theoretical_loss": 4.196004060794589,
      "tokens_seen": 271319040
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046352056168505516,
      "loss": 2.8734,
      "theoretical_loss": 4.195883212427433,
      "tokens_seen": 271384576
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046351053159478434,
      "loss": 3.0506,
      "theoretical_loss": 4.195762401409229,
      "tokens_seen": 271450112
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004635005015045136,
      "loss": 2.8748,
      "theoretical_loss": 4.19564162771942,
      "tokens_seen": 271515648
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004634904714142427,
      "loss": 3.0623,
      "theoretical_loss": 4.195520891337466,
      "tokens_seen": 271581184
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046348044132397194,
      "loss": 3.0601,
      "theoretical_loss": 4.195400192242845,
      "tokens_seen": 271646720
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046347041123370107,
      "loss": 2.7883,
      "theoretical_loss": 4.19527953041505,
      "tokens_seen": 271712256
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004634603811434303,
      "loss": 3.1026,
      "theoretical_loss": 4.19515890583359,
      "tokens_seen": 271777792
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004634503510531595,
      "loss": 3.1229,
      "theoretical_loss": 4.195038318477989,
      "tokens_seen": 271843328
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046344032096288867,
      "loss": 2.9915,
      "theoretical_loss": 4.194917768327789,
      "tokens_seen": 271908864
    },
    {
      "epoch": 0.08,
      "objective/train/docs_used": 314289,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.050889253616333,
      "objective/train/theoretical_loss": 4.194797255362549,
      "objective/train/tokens_used": 292434400,
      "theoretical_loss": 4.194797255362549,
      "tokens_seen": 271974400
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046343029087261785,
      "loss": 3.1228,
      "theoretical_loss": 4.194797255362549,
      "tokens_seen": 271974400
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004634202607823471,
      "loss": 3.1203,
      "theoretical_loss": 4.194676779561841,
      "tokens_seen": 272039936
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004634102306920762,
      "loss": 2.7881,
      "theoretical_loss": 4.194556340905256,
      "tokens_seen": 272105472
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046340020060180545,
      "loss": 3.0799,
      "theoretical_loss": 4.194435939372401,
      "tokens_seen": 272171008
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046339017051153457,
      "loss": 3.032,
      "theoretical_loss": 4.194315574942896,
      "tokens_seen": 272236544
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004633801404212638,
      "loss": 3.2427,
      "theoretical_loss": 4.194195247596381,
      "tokens_seen": 272302080
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000463370110330993,
      "loss": 2.9247,
      "theoretical_loss": 4.19407495731251,
      "tokens_seen": 272367616
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046336008024072217,
      "loss": 2.9747,
      "theoretical_loss": 4.193954704070952,
      "tokens_seen": 272433152
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046335005015045135,
      "loss": 3.1666,
      "theoretical_loss": 4.193834487851396,
      "tokens_seen": 272498688
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046334002006018053,
      "loss": 3.0714,
      "theoretical_loss": 4.193714308633542,
      "tokens_seen": 272564224
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004633299899699097,
      "loss": 3.0147,
      "theoretical_loss": 4.1935941663971095,
      "tokens_seen": 272629760
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046331995987963895,
      "loss": 3.1997,
      "theoretical_loss": 4.193474061121833,
      "tokens_seen": 272695296
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004633099297893681,
      "loss": 2.8764,
      "theoretical_loss": 4.193353992787463,
      "tokens_seen": 272760832
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004632998996990973,
      "loss": 3.2184,
      "theoretical_loss": 4.193233961373766,
      "tokens_seen": 272826368
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046328986960882644,
      "loss": 3.1254,
      "theoretical_loss": 4.1931139668605235,
      "tokens_seen": 272891904
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004632798395185557,
      "loss": 3.0326,
      "theoretical_loss": 4.192994009227535,
      "tokens_seen": 272957440
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046326980942828485,
      "loss": 3.0574,
      "theoretical_loss": 4.192874088454613,
      "tokens_seen": 273022976
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046325977933801404,
      "loss": 3.1175,
      "theoretical_loss": 4.19275420452159,
      "tokens_seen": 273088512
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004632497492477432,
      "loss": 3.0634,
      "theoretical_loss": 4.192634357408309,
      "tokens_seen": 273154048
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046323971915747245,
      "loss": 2.7333,
      "theoretical_loss": 4.192514547094634,
      "tokens_seen": 273219584
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004632296890672016,
      "loss": 2.9853,
      "theoretical_loss": 4.192394773560441,
      "tokens_seen": 273285120
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004632196589769308,
      "loss": 3.1364,
      "theoretical_loss": 4.192275036785625,
      "tokens_seen": 273350656
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046320962888666,
      "loss": 2.8319,
      "theoretical_loss": 4.192155336750094,
      "tokens_seen": 273416192
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004631995987963892,
      "loss": 3.0046,
      "theoretical_loss": 4.192035673433773,
      "tokens_seen": 273481728
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004631895687061184,
      "loss": 3.1155,
      "theoretical_loss": 4.191916046816605,
      "tokens_seen": 273547264
    },
    {
      "epoch": 0.08,
      "objective/train/docs_used": 314822,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.1304068565368652,
      "objective/train/theoretical_loss": 4.191796456878544,
      "objective/train/tokens_used": 294072800,
      "theoretical_loss": 4.191796456878544,
      "tokens_seen": 273612800
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046317953861584754,
      "loss": 2.943,
      "theoretical_loss": 4.191796456878544,
      "tokens_seen": 273612800
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004631695085255768,
      "loss": 2.8587,
      "theoretical_loss": 4.191676903599563,
      "tokens_seen": 273678336
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004631594784353059,
      "loss": 3.1101,
      "theoretical_loss": 4.191557386959651,
      "tokens_seen": 273743872
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046314944834503514,
      "loss": 3.1475,
      "theoretical_loss": 4.191437906938811,
      "tokens_seen": 273809408
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004631394182547643,
      "loss": 3.0439,
      "theoretical_loss": 4.191318463517062,
      "tokens_seen": 273874944
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004631293881644935,
      "loss": 3.1469,
      "theoretical_loss": 4.19119905667444,
      "tokens_seen": 273940480
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004631193580742227,
      "loss": 3.1943,
      "theoretical_loss": 4.191079686390996,
      "tokens_seen": 274006016
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004631093279839519,
      "loss": 3.1121,
      "theoretical_loss": 4.190960352646796,
      "tokens_seen": 274071552
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046309929789368104,
      "loss": 3.0344,
      "theoretical_loss": 4.190841055421921,
      "tokens_seen": 274137088
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004630892678034103,
      "loss": 3.0578,
      "theoretical_loss": 4.19072179469647,
      "tokens_seen": 274202624
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004630792377131394,
      "loss": 3.1027,
      "theoretical_loss": 4.190602570450556,
      "tokens_seen": 274268160
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046306920762286864,
      "loss": 3.0957,
      "theoretical_loss": 4.190483382664308,
      "tokens_seen": 274333696
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004630591775325978,
      "loss": 2.8938,
      "theoretical_loss": 4.19036423131787,
      "tokens_seen": 274399232
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000463049147442327,
      "loss": 2.8262,
      "theoretical_loss": 4.190245116391403,
      "tokens_seen": 274464768
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004630391173520562,
      "loss": 2.9228,
      "theoretical_loss": 4.190126037865082,
      "tokens_seen": 274530304
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046302908726178536,
      "loss": 3.0006,
      "theoretical_loss": 4.190006995719098,
      "tokens_seen": 274595840
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046301905717151455,
      "loss": 2.9758,
      "theoretical_loss": 4.1898879899336565,
      "tokens_seen": 274661376
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004630090270812438,
      "loss": 3.2112,
      "theoretical_loss": 4.189769020488981,
      "tokens_seen": 274726912
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004629989969909729,
      "loss": 3.0473,
      "theoretical_loss": 4.189650087365309,
      "tokens_seen": 274792448
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046298896690070214,
      "loss": 2.8983,
      "theoretical_loss": 4.189531190542893,
      "tokens_seen": 274857984
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046297893681043127,
      "loss": 2.7118,
      "theoretical_loss": 4.189412330002001,
      "tokens_seen": 274923520
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004629689067201605,
      "loss": 2.9403,
      "theoretical_loss": 4.189293505722918,
      "tokens_seen": 274989056
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004629588766298897,
      "loss": 2.9515,
      "theoretical_loss": 4.189174717685942,
      "tokens_seen": 275054592
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046294884653961887,
      "loss": 2.7063,
      "theoretical_loss": 4.189055965871389,
      "tokens_seen": 275120128
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046293881644934805,
      "loss": 3.0041,
      "theoretical_loss": 4.188937250259587,
      "tokens_seen": 275185664
    },
    {
      "epoch": 0.08,
      "objective/train/docs_used": 318374,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.3687500953674316,
      "objective/train/theoretical_loss": 4.188818570830883,
      "objective/train/tokens_used": 295711200,
      "theoretical_loss": 4.188818570830883,
      "tokens_seen": 275251200
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004629287863590773,
      "loss": 2.9334,
      "theoretical_loss": 4.188818570830883,
      "tokens_seen": 275251200
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004629187562688064,
      "loss": 2.9519,
      "theoretical_loss": 4.188699927565638,
      "tokens_seen": 275316736
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046290872617853565,
      "loss": 3.24,
      "theoretical_loss": 4.188581320444228,
      "tokens_seen": 275382272
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046289869608826477,
      "loss": 3.0205,
      "theoretical_loss": 4.1884627494470426,
      "tokens_seen": 275447808
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000462888665997994,
      "loss": 3.1389,
      "theoretical_loss": 4.1883442145544905,
      "tokens_seen": 275513344
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004628786359077232,
      "loss": 3.117,
      "theoretical_loss": 4.188225715746992,
      "tokens_seen": 275578880
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046286860581745237,
      "loss": 3.0967,
      "theoretical_loss": 4.188107253004986,
      "tokens_seen": 275644416
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046285857572718155,
      "loss": 3.2164,
      "theoretical_loss": 4.187988826308925,
      "tokens_seen": 275709952
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046284854563691073,
      "loss": 3.0597,
      "theoretical_loss": 4.187870435639275,
      "tokens_seen": 275775488
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004628385155466399,
      "loss": 2.9352,
      "theoretical_loss": 4.18775208097652,
      "tokens_seen": 275841024
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046282848545636915,
      "loss": 3.2597,
      "theoretical_loss": 4.187633762301159,
      "tokens_seen": 275906560
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004628184553660983,
      "loss": 3.0023,
      "theoretical_loss": 4.187515479593704,
      "tokens_seen": 275972096
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004628084252758275,
      "loss": 2.9841,
      "theoretical_loss": 4.187397232834683,
      "tokens_seen": 276037632
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046279839518555664,
      "loss": 3.3956,
      "theoretical_loss": 4.187279022004642,
      "tokens_seen": 276103168
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004627883650952859,
      "loss": 3.2086,
      "theoretical_loss": 4.1871608470841375,
      "tokens_seen": 276168704
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046277833500501505,
      "loss": 2.8913,
      "theoretical_loss": 4.1870427080537445,
      "tokens_seen": 276234240
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046276830491474424,
      "loss": 2.9785,
      "theoretical_loss": 4.1869246048940525,
      "tokens_seen": 276299776
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004627582748244734,
      "loss": 3.138,
      "theoretical_loss": 4.186806537585666,
      "tokens_seen": 276365312
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046274824473420265,
      "loss": 2.9597,
      "theoretical_loss": 4.186688506109202,
      "tokens_seen": 276430848
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004627382146439318,
      "loss": 3.1554,
      "theoretical_loss": 4.186570510445296,
      "tokens_seen": 276496384
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000462728184553661,
      "loss": 3.1408,
      "theoretical_loss": 4.186452550574599,
      "tokens_seen": 276561920
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046271815446339014,
      "loss": 2.8797,
      "theoretical_loss": 4.186334626477774,
      "tokens_seen": 276627456
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004627081243731194,
      "loss": 2.892,
      "theoretical_loss": 4.186216738135501,
      "tokens_seen": 276692992
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046269809428284856,
      "loss": 2.6408,
      "theoretical_loss": 4.186098885528473,
      "tokens_seen": 276758528
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046268806419257774,
      "loss": 3.0981,
      "theoretical_loss": 4.185981068637401,
      "tokens_seen": 276824064
    },
    {
      "epoch": 0.08,
      "objective/train/docs_used": 320606,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.31486439704895,
      "objective/train/theoretical_loss": 4.185863287443008,
      "objective/train/tokens_used": 296742368,
      "theoretical_loss": 4.185863287443008,
      "tokens_seen": 276889600
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004626780341023069,
      "loss": 3.0755,
      "theoretical_loss": 4.185863287443008,
      "tokens_seen": 276889600
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004626680040120361,
      "loss": 2.9726,
      "theoretical_loss": 4.185745541926035,
      "tokens_seen": 276955136
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004626579739217653,
      "loss": 3.1442,
      "theoretical_loss": 4.185627832067237,
      "tokens_seen": 277020672
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004626479438314945,
      "loss": 3.17,
      "theoretical_loss": 4.1855101578473795,
      "tokens_seen": 277086208
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046263791374122364,
      "loss": 3.0287,
      "theoretical_loss": 4.18539251924725,
      "tokens_seen": 277151744
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004626278836509529,
      "loss": 2.8911,
      "theoretical_loss": 4.185274916247646,
      "tokens_seen": 277217280
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000462617853560682,
      "loss": 2.869,
      "theoretical_loss": 4.185157348829383,
      "tokens_seen": 277282816
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046260782347041124,
      "loss": 3.1912,
      "theoretical_loss": 4.185039816973289,
      "tokens_seen": 277348352
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004625977933801404,
      "loss": 2.8488,
      "theoretical_loss": 4.184922320660207,
      "tokens_seen": 277413888
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004625877632898696,
      "loss": 3.0904,
      "theoretical_loss": 4.184804859870997,
      "tokens_seen": 277479424
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004625777331995988,
      "loss": 3.0076,
      "theoretical_loss": 4.184687434586531,
      "tokens_seen": 277544960
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000462567703109328,
      "loss": 2.8916,
      "theoretical_loss": 4.184570044787698,
      "tokens_seen": 277610496
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046255767301905715,
      "loss": 2.8376,
      "theoretical_loss": 4.1844526904554,
      "tokens_seen": 277676032
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004625476429287864,
      "loss": 2.6536,
      "theoretical_loss": 4.184335371570556,
      "tokens_seen": 277741568
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004625376128385155,
      "loss": 2.97,
      "theoretical_loss": 4.184218088114097,
      "tokens_seen": 277807104
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046252758274824475,
      "loss": 3.0261,
      "theoretical_loss": 4.1841008400669715,
      "tokens_seen": 277872640
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004625175526579739,
      "loss": 3.0633,
      "theoretical_loss": 4.183983627410142,
      "tokens_seen": 277938176
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004625075225677031,
      "loss": 3.0716,
      "theoretical_loss": 4.183866450124584,
      "tokens_seen": 278003712
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004624974924774323,
      "loss": 3.0718,
      "theoretical_loss": 4.18374930819129,
      "tokens_seen": 278069248
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046248746238716147,
      "loss": 3.0414,
      "theoretical_loss": 4.183632201591264,
      "tokens_seen": 278134784
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046247743229689065,
      "loss": 2.8162,
      "theoretical_loss": 4.18351513030553,
      "tokens_seen": 278200320
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004624674022066199,
      "loss": 3.1109,
      "theoretical_loss": 4.1833980943151206,
      "tokens_seen": 278265856
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046245737211634907,
      "loss": 3.0844,
      "theoretical_loss": 4.183281093601087,
      "tokens_seen": 278331392
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046244734202607825,
      "loss": 3.2941,
      "theoretical_loss": 4.183164128144495,
      "tokens_seen": 278396928
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004624373119358075,
      "loss": 3.0747,
      "theoretical_loss": 4.183047197926422,
      "tokens_seen": 278462464
    },
    {
      "epoch": 0.08,
      "objective/train/docs_used": 320606,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.0842607021331787,
      "objective/train/theoretical_loss": 4.182930302927963,
      "objective/train/tokens_used": 296742368,
      "theoretical_loss": 4.182930302927963,
      "tokens_seen": 278528000
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004624272818455366,
      "loss": 2.8917,
      "theoretical_loss": 4.182930302927963,
      "tokens_seen": 278528000
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046241725175526585,
      "loss": 3.085,
      "theoretical_loss": 4.182813443130227,
      "tokens_seen": 278593536
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000462407221664995,
      "loss": 2.8507,
      "theoretical_loss": 4.182696618514337,
      "tokens_seen": 278659072
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004623971915747242,
      "loss": 2.9847,
      "theoretical_loss": 4.18257982906143,
      "tokens_seen": 278724608
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004623871614844534,
      "loss": 3.0537,
      "theoretical_loss": 4.1824630747526585,
      "tokens_seen": 278790144
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046237713139418257,
      "loss": 2.8442,
      "theoretical_loss": 4.182346355569189,
      "tokens_seen": 278855680
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046236710130391175,
      "loss": 3.2378,
      "theoretical_loss": 4.182229671492204,
      "tokens_seen": 278921216
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046235707121364093,
      "loss": 2.8921,
      "theoretical_loss": 4.1821130225028975,
      "tokens_seen": 278986752
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004623470411233701,
      "loss": 2.9071,
      "theoretical_loss": 4.1819964085824815,
      "tokens_seen": 279052288
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046233701103309935,
      "loss": 2.8913,
      "theoretical_loss": 4.181879829712178,
      "tokens_seen": 279117824
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004623269809428285,
      "loss": 3.2138,
      "theoretical_loss": 4.181763285873231,
      "tokens_seen": 279183360
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004623169508525577,
      "loss": 3.0489,
      "theoretical_loss": 4.181646777046889,
      "tokens_seen": 279248896
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046230692076228684,
      "loss": 2.8127,
      "theoretical_loss": 4.181530303214423,
      "tokens_seen": 279314432
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004622968906720161,
      "loss": 3.1112,
      "theoretical_loss": 4.181413864357115,
      "tokens_seen": 279379968
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046228686058174525,
      "loss": 2.9245,
      "theoretical_loss": 4.181297460456262,
      "tokens_seen": 279445504
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046227683049147444,
      "loss": 2.6736,
      "theoretical_loss": 4.181181091493174,
      "tokens_seen": 279511040
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004622668004012036,
      "loss": 3.0354,
      "theoretical_loss": 4.181064757449178,
      "tokens_seen": 279576576
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046225677031093285,
      "loss": 3.3787,
      "theoretical_loss": 4.180948458305615,
      "tokens_seen": 279642112
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000462246740220662,
      "loss": 2.9473,
      "theoretical_loss": 4.180832194043836,
      "tokens_seen": 279707648
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004622367101303912,
      "loss": 2.9305,
      "theoretical_loss": 4.180715964645213,
      "tokens_seen": 279773184
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046222668004012034,
      "loss": 3.0333,
      "theoretical_loss": 4.180599770091126,
      "tokens_seen": 279838720
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004622166499498496,
      "loss": 2.9703,
      "theoretical_loss": 4.180483610362975,
      "tokens_seen": 279904256
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046220661985957876,
      "loss": 2.7939,
      "theoretical_loss": 4.18036748544217,
      "tokens_seen": 279969792
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046219658976930794,
      "loss": 3.0377,
      "theoretical_loss": 4.180251395310137,
      "tokens_seen": 280035328
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004621865596790371,
      "loss": 3.0576,
      "theoretical_loss": 4.1801353399483165,
      "tokens_seen": 280100864
    },
    {
      "epoch": 0.08,
      "objective/train/docs_used": 320606,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.94958758354187,
      "objective/train/theoretical_loss": 4.180019319338163,
      "objective/train/tokens_used": 296742368,
      "theoretical_loss": 4.180019319338163,
      "tokens_seen": 280166400
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004621765295887663,
      "loss": 3.0546,
      "theoretical_loss": 4.180019319338163,
      "tokens_seen": 280166400
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004621664994984955,
      "loss": 2.875,
      "theoretical_loss": 4.179903333461144,
      "tokens_seen": 280231936
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004621564694082247,
      "loss": 3.0677,
      "theoretical_loss": 4.179787382298744,
      "tokens_seen": 280297472
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046214643931795384,
      "loss": 3.1748,
      "theoretical_loss": 4.179671465832458,
      "tokens_seen": 280363008
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004621364092276831,
      "loss": 3.0775,
      "theoretical_loss": 4.179555584043799,
      "tokens_seen": 280428544
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004621263791374122,
      "loss": 3.1528,
      "theoretical_loss": 4.17943973691429,
      "tokens_seen": 280494080
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046211634904714144,
      "loss": 2.9587,
      "theoretical_loss": 4.179323924425472,
      "tokens_seen": 280559616
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004621063189568706,
      "loss": 2.9175,
      "theoretical_loss": 4.179208146558899,
      "tokens_seen": 280625152
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004620962888665998,
      "loss": 2.9214,
      "theoretical_loss": 4.1790924032961385,
      "tokens_seen": 280690688
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000462086258776329,
      "loss": 2.9464,
      "theoretical_loss": 4.178976694618772,
      "tokens_seen": 280756224
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004620762286860582,
      "loss": 3.1425,
      "theoretical_loss": 4.178861020508395,
      "tokens_seen": 280821760
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046206619859578735,
      "loss": 2.8949,
      "theoretical_loss": 4.178745380946619,
      "tokens_seen": 280887296
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004620561685055166,
      "loss": 2.9406,
      "theoretical_loss": 4.178629775915066,
      "tokens_seen": 280952832
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004620461384152457,
      "loss": 2.9719,
      "theoretical_loss": 4.178514205395376,
      "tokens_seen": 281018368
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046203610832497495,
      "loss": 2.9724,
      "theoretical_loss": 4.178398669369201,
      "tokens_seen": 281083904
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004620260782347041,
      "loss": 3.0159,
      "theoretical_loss": 4.178283167818206,
      "tokens_seen": 281149440
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004620160481444333,
      "loss": 3.1142,
      "theoretical_loss": 4.178167700724073,
      "tokens_seen": 281214976
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004620060180541625,
      "loss": 3.1041,
      "theoretical_loss": 4.178052268068494,
      "tokens_seen": 281280512
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046199598796389167,
      "loss": 2.9121,
      "theoretical_loss": 4.177936869833179,
      "tokens_seen": 281346048
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046198595787362085,
      "loss": 2.8584,
      "theoretical_loss": 4.17782150599985,
      "tokens_seen": 281411584
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004619759277833501,
      "loss": 2.9394,
      "theoretical_loss": 4.1777061765502435,
      "tokens_seen": 281477120
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004619658976930792,
      "loss": 3.0848,
      "theoretical_loss": 4.1775908814661085,
      "tokens_seen": 281542656
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046195586760280845,
      "loss": 2.7208,
      "theoretical_loss": 4.17747562072921,
      "tokens_seen": 281608192
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046194583751253763,
      "loss": 3.0835,
      "theoretical_loss": 4.177360394321325,
      "tokens_seen": 281673728
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004619358074222668,
      "loss": 3.1928,
      "theoretical_loss": 4.177245202224246,
      "tokens_seen": 281739264
    },
    {
      "epoch": 0.09,
      "objective/train/docs_used": 320606,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.1719119548797607,
      "objective/train/theoretical_loss": 4.17713004441978,
      "objective/train/tokens_used": 296742368,
      "theoretical_loss": 4.17713004441978,
      "tokens_seen": 281804800
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000461925777331996,
      "loss": 2.9915,
      "theoretical_loss": 4.17713004441978,
      "tokens_seen": 281804800
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004619157472417252,
      "loss": 2.9478,
      "theoretical_loss": 4.177014920889745,
      "tokens_seen": 281870336
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046190571715145435,
      "loss": 3.0954,
      "theoretical_loss": 4.176899831615974,
      "tokens_seen": 281935872
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004618956870611836,
      "loss": 2.9708,
      "theoretical_loss": 4.176784776580316,
      "tokens_seen": 282001408
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004618856569709127,
      "loss": 3.0337,
      "theoretical_loss": 4.176669755764632,
      "tokens_seen": 282066944
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046187562688064195,
      "loss": 3.261,
      "theoretical_loss": 4.176554769150796,
      "tokens_seen": 282132480
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004618655967903711,
      "loss": 3.0503,
      "theoretical_loss": 4.176439816720697,
      "tokens_seen": 282198016
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004618555667001003,
      "loss": 2.8879,
      "theoretical_loss": 4.1763248984562376,
      "tokens_seen": 282263552
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004618455366098295,
      "loss": 2.9648,
      "theoretical_loss": 4.176210014339335,
      "tokens_seen": 282329088
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004618355065195587,
      "loss": 2.673,
      "theoretical_loss": 4.17609516435192,
      "tokens_seen": 282394624
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046182547642928786,
      "loss": 3.2389,
      "theoretical_loss": 4.1759803484759335,
      "tokens_seen": 282460160
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046181544633901704,
      "loss": 2.9818,
      "theoretical_loss": 4.175865566693336,
      "tokens_seen": 282525696
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004618054162487462,
      "loss": 2.7571,
      "theoretical_loss": 4.175750818986098,
      "tokens_seen": 282591232
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046179538615847545,
      "loss": 2.9559,
      "theoretical_loss": 4.1756361053362046,
      "tokens_seen": 282656768
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004617853560682046,
      "loss": 3.1227,
      "theoretical_loss": 4.1755214257256545,
      "tokens_seen": 282722304
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004617753259779338,
      "loss": 3.1455,
      "theoretical_loss": 4.17540678013646,
      "tokens_seen": 282787840
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000461765295887663,
      "loss": 3.044,
      "theoretical_loss": 4.175292168550648,
      "tokens_seen": 282853376
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004617552657973922,
      "loss": 2.7818,
      "theoretical_loss": 4.175177590950257,
      "tokens_seen": 282918912
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046174523570712136,
      "loss": 2.7973,
      "theoretical_loss": 4.175063047317342,
      "tokens_seen": 282984448
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046173520561685054,
      "loss": 3.1903,
      "theoretical_loss": 4.174948537633968,
      "tokens_seen": 283049984
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004617251755265797,
      "loss": 3.2353,
      "theoretical_loss": 4.174834061882218,
      "tokens_seen": 283115520
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046171514543630896,
      "loss": 3.0471,
      "theoretical_loss": 4.1747196200441845,
      "tokens_seen": 283181056
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046170511534603814,
      "loss": 2.8341,
      "theoretical_loss": 4.174605212101977,
      "tokens_seen": 283246592
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004616950852557673,
      "loss": 2.8812,
      "theoretical_loss": 4.174490838037716,
      "tokens_seen": 283312128
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004616850551654965,
      "loss": 2.9857,
      "theoretical_loss": 4.174376497833537,
      "tokens_seen": 283377664
    },
    {
      "epoch": 0.09,
      "objective/train/docs_used": 320606,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5323517322540283,
      "objective/train/theoretical_loss": 4.174262191471587,
      "objective/train/tokens_used": 296742368,
      "theoretical_loss": 4.174262191471587,
      "tokens_seen": 283443200
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004616750250752257,
      "loss": 2.8143,
      "theoretical_loss": 4.174262191471587,
      "tokens_seen": 283443200
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004616649949849549,
      "loss": 3.11,
      "theoretical_loss": 4.17414791893403,
      "tokens_seen": 283508736
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046165496489468404,
      "loss": 3.0364,
      "theoretical_loss": 4.17403368020304,
      "tokens_seen": 283574272
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004616449348044133,
      "loss": 3.2199,
      "theoretical_loss": 4.173919475260808,
      "tokens_seen": 283639808
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004616349047141424,
      "loss": 3.0457,
      "theoretical_loss": 4.173805304089536,
      "tokens_seen": 283705344
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046162487462387164,
      "loss": 2.9339,
      "theoretical_loss": 4.173691166671439,
      "tokens_seen": 283770880
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004616148445336008,
      "loss": 3.1206,
      "theoretical_loss": 4.173577062988748,
      "tokens_seen": 283836416
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046160481444333,
      "loss": 2.7604,
      "theoretical_loss": 4.173462993023706,
      "tokens_seen": 283901952
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004615947843530592,
      "loss": 2.7994,
      "theoretical_loss": 4.173348956758568,
      "tokens_seen": 283967488
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004615847542627884,
      "loss": 2.885,
      "theoretical_loss": 4.173234954175605,
      "tokens_seen": 284033024
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046157472417251755,
      "loss": 2.7621,
      "theoretical_loss": 4.173120985257102,
      "tokens_seen": 284098560
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004615646940822468,
      "loss": 3.1411,
      "theoretical_loss": 4.173007049985352,
      "tokens_seen": 284164096
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004615546639919759,
      "loss": 2.7966,
      "theoretical_loss": 4.172893148342667,
      "tokens_seen": 284229632
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046154463390170515,
      "loss": 2.9813,
      "theoretical_loss": 4.172779280311372,
      "tokens_seen": 284295168
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004615346038114343,
      "loss": 3.0168,
      "theoretical_loss": 4.172665445873801,
      "tokens_seen": 284360704
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004615245737211635,
      "loss": 2.9192,
      "theoretical_loss": 4.172551645012307,
      "tokens_seen": 284426240
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004615145436308927,
      "loss": 2.8819,
      "theoretical_loss": 4.1724378777092515,
      "tokens_seen": 284491776
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046150451354062187,
      "loss": 2.5965,
      "theoretical_loss": 4.172324143947012,
      "tokens_seen": 284557312
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046149448345035105,
      "loss": 3.0194,
      "theoretical_loss": 4.172210443707979,
      "tokens_seen": 284622848
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004614844533600803,
      "loss": 2.7609,
      "theoretical_loss": 4.1720967769745565,
      "tokens_seen": 284688384
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004614744232698094,
      "loss": 2.9768,
      "theoretical_loss": 4.171983143729159,
      "tokens_seen": 284753920
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046146439317953865,
      "loss": 2.7417,
      "theoretical_loss": 4.1718695439542195,
      "tokens_seen": 284819456
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046145436308926783,
      "loss": 3.1363,
      "theoretical_loss": 4.17175597763218,
      "tokens_seen": 284884992
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000461444332998997,
      "loss": 2.9471,
      "theoretical_loss": 4.171642444745497,
      "tokens_seen": 284950528
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004614343029087262,
      "loss": 3.2626,
      "theoretical_loss": 4.1715289452766395,
      "tokens_seen": 285016064
    },
    {
      "epoch": 0.09,
      "objective/train/docs_used": 320606,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.09801983833313,
      "objective/train/theoretical_loss": 4.1714154792080915,
      "objective/train/tokens_used": 296742368,
      "theoretical_loss": 4.1714154792080915,
      "tokens_seen": 285081600
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004614242728184554,
      "loss": 3.0522,
      "theoretical_loss": 4.1714154792080915,
      "tokens_seen": 285081600
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046141424272818455,
      "loss": 2.7098,
      "theoretical_loss": 4.171302046522349,
      "tokens_seen": 285147136
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004614042126379138,
      "loss": 3.1704,
      "theoretical_loss": 4.171188647201921,
      "tokens_seen": 285212672
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004613941825476429,
      "loss": 3.0845,
      "theoretical_loss": 4.1710752812293315,
      "tokens_seen": 285278208
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046138415245737215,
      "loss": 2.8987,
      "theoretical_loss": 4.170961948587115,
      "tokens_seen": 285343744
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004613741223671013,
      "loss": 2.7445,
      "theoretical_loss": 4.17084864925782,
      "tokens_seen": 285409280
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004613640922768305,
      "loss": 3.0298,
      "theoretical_loss": 4.1707353832240095,
      "tokens_seen": 285474816
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004613540621865597,
      "loss": 2.9432,
      "theoretical_loss": 4.170622150468258,
      "tokens_seen": 285540352
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004613440320962889,
      "loss": 3.1768,
      "theoretical_loss": 4.170508950973154,
      "tokens_seen": 285605888
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046133400200601806,
      "loss": 2.7724,
      "theoretical_loss": 4.1703957847213,
      "tokens_seen": 285671424
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046132397191574724,
      "loss": 2.7251,
      "theoretical_loss": 4.170282651695308,
      "tokens_seen": 285736960
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004613139418254764,
      "loss": 2.9493,
      "theoretical_loss": 4.170169551877808,
      "tokens_seen": 285802496
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046130391173520566,
      "loss": 2.8802,
      "theoretical_loss": 4.170056485251439,
      "tokens_seen": 285868032
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004612938816449348,
      "loss": 3.0992,
      "theoretical_loss": 4.169943451798856,
      "tokens_seen": 285933568
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000461283851554664,
      "loss": 2.9054,
      "theoretical_loss": 4.169830451502724,
      "tokens_seen": 285999104
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004612738214643932,
      "loss": 2.8996,
      "theoretical_loss": 4.169717484345725,
      "tokens_seen": 286064640
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004612637913741224,
      "loss": 2.7219,
      "theoretical_loss": 4.1696045503105506,
      "tokens_seen": 286130176
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046125376128385156,
      "loss": 3.0844,
      "theoretical_loss": 4.169491649379905,
      "tokens_seen": 286195712
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046124373119358074,
      "loss": 3.1184,
      "theoretical_loss": 4.169378781536509,
      "tokens_seen": 286261248
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004612337011033099,
      "loss": 2.7949,
      "theoretical_loss": 4.169265946763095,
      "tokens_seen": 286326784
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046122367101303916,
      "loss": 3.0171,
      "theoretical_loss": 4.169153145042405,
      "tokens_seen": 286392320
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004612136409227683,
      "loss": 3.0261,
      "theoretical_loss": 4.169040376357199,
      "tokens_seen": 286457856
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004612036108324975,
      "loss": 2.9589,
      "theoretical_loss": 4.168927640690246,
      "tokens_seen": 286523392
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046119358074222665,
      "loss": 2.9269,
      "theoretical_loss": 4.16881493802433,
      "tokens_seen": 286588928
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004611835506519559,
      "loss": 2.916,
      "theoretical_loss": 4.168702268342248,
      "tokens_seen": 286654464
    },
    {
      "epoch": 0.09,
      "objective/train/docs_used": 320606,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.830076217651367,
      "objective/train/theoretical_loss": 4.168589631626808,
      "objective/train/tokens_used": 296742368,
      "theoretical_loss": 4.168589631626808,
      "tokens_seen": 286720000
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046117352056168506,
      "loss": 2.9636,
      "theoretical_loss": 4.168589631626808,
      "tokens_seen": 286720000
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046116349047141425,
      "loss": 2.7699,
      "theoretical_loss": 4.168477027860833,
      "tokens_seen": 286785536
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004611534603811434,
      "loss": 3.0801,
      "theoretical_loss": 4.168364457027158,
      "tokens_seen": 286851072
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004611434302908726,
      "loss": 3.0063,
      "theoretical_loss": 4.168251919108632,
      "tokens_seen": 286916608
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004611334002006018,
      "loss": 2.8081,
      "theoretical_loss": 4.168139414088113,
      "tokens_seen": 286982144
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000461123370110331,
      "loss": 3.0765,
      "theoretical_loss": 4.168026941948478,
      "tokens_seen": 287047680
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046111334002006015,
      "loss": 3.1996,
      "theoretical_loss": 4.167914502672611,
      "tokens_seen": 287113216
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004611033099297894,
      "loss": 2.7884,
      "theoretical_loss": 4.1678020962434115,
      "tokens_seen": 287178752
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046109327983951857,
      "loss": 2.9895,
      "theoretical_loss": 4.167689722643792,
      "tokens_seen": 287244288
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046108324974924775,
      "loss": 3.0916,
      "theoretical_loss": 4.1675773818566775,
      "tokens_seen": 287309824
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046107321965897693,
      "loss": 2.9913,
      "theoretical_loss": 4.167465073865006,
      "tokens_seen": 287375360
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004610631895687061,
      "loss": 2.8553,
      "theoretical_loss": 4.167352798651726,
      "tokens_seen": 287440896
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004610531594784353,
      "loss": 3.0485,
      "theoretical_loss": 4.167240556199802,
      "tokens_seen": 287506432
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004610431293881645,
      "loss": 3.1886,
      "theoretical_loss": 4.167128346492211,
      "tokens_seen": 287571968
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046103309929789365,
      "loss": 2.9215,
      "theoretical_loss": 4.16701616951194,
      "tokens_seen": 287637504
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004610230692076229,
      "loss": 3.0773,
      "theoretical_loss": 4.1669040252419896,
      "tokens_seen": 287703040
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000461013039117352,
      "loss": 3.2776,
      "theoretical_loss": 4.166791913665375,
      "tokens_seen": 287768576
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046100300902708125,
      "loss": 3.0554,
      "theoretical_loss": 4.166679834765123,
      "tokens_seen": 287834112
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046099297893681043,
      "loss": 2.9973,
      "theoretical_loss": 4.166567788524272,
      "tokens_seen": 287899648
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004609829488465396,
      "loss": 3.2166,
      "theoretical_loss": 4.166455774925875,
      "tokens_seen": 287965184
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004609729187562688,
      "loss": 3.1155,
      "theoretical_loss": 4.166343793952995,
      "tokens_seen": 288030720
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046096288866599803,
      "loss": 2.9107,
      "theoretical_loss": 4.166231845588712,
      "tokens_seen": 288096256
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004609528585757272,
      "loss": 2.8911,
      "theoretical_loss": 4.166119929816113,
      "tokens_seen": 288161792
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004609428284854564,
      "loss": 3.0025,
      "theoretical_loss": 4.166008046618303,
      "tokens_seen": 288227328
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004609327983951856,
      "loss": 2.9673,
      "theoretical_loss": 4.1658961959783944,
      "tokens_seen": 288292864
    },
    {
      "epoch": 0.09,
      "objective/train/docs_used": 320606,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.154595136642456,
      "objective/train/theoretical_loss": 4.165784377879517,
      "objective/train/tokens_used": 296742368,
      "theoretical_loss": 4.165784377879517,
      "tokens_seen": 288358400
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046092276830491475,
      "loss": 3.198,
      "theoretical_loss": 4.165784377879517,
      "tokens_seen": 288358400
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000460912738214644,
      "loss": 3.1049,
      "theoretical_loss": 4.165672592304811,
      "tokens_seen": 288423936
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004609027081243731,
      "loss": 2.9952,
      "theoretical_loss": 4.165560839237429,
      "tokens_seen": 288489472
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046089267803410235,
      "loss": 3.0699,
      "theoretical_loss": 4.165449118660536,
      "tokens_seen": 288555008
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004608826479438315,
      "loss": 3.0712,
      "theoretical_loss": 4.16533743055731,
      "tokens_seen": 288620544
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004608726178535607,
      "loss": 3.1647,
      "theoretical_loss": 4.165225774910941,
      "tokens_seen": 288686080
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004608625877632899,
      "loss": 3.2739,
      "theoretical_loss": 4.165114151704634,
      "tokens_seen": 288751616
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004608525576730191,
      "loss": 3.0161,
      "theoretical_loss": 4.165002560921601,
      "tokens_seen": 288817152
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046084252758274826,
      "loss": 3.1432,
      "theoretical_loss": 4.164891002545073,
      "tokens_seen": 288882688
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046083249749247744,
      "loss": 2.9712,
      "theoretical_loss": 4.16477947655829,
      "tokens_seen": 288948224
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004608224674022066,
      "loss": 2.8887,
      "theoretical_loss": 4.164667982944504,
      "tokens_seen": 289013760
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046081243731193586,
      "loss": 2.7605,
      "theoretical_loss": 4.164556521686981,
      "tokens_seen": 289079296
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000460802407221665,
      "loss": 3.1795,
      "theoretical_loss": 4.1644450927689975,
      "tokens_seen": 289144832
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004607923771313942,
      "loss": 2.902,
      "theoretical_loss": 4.164333696173846,
      "tokens_seen": 289210368
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004607823470411234,
      "loss": 3.0938,
      "theoretical_loss": 4.164222331884827,
      "tokens_seen": 289275904
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004607723169508526,
      "loss": 2.9628,
      "theoretical_loss": 4.164110999885256,
      "tokens_seen": 289341440
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046076228686058176,
      "loss": 2.8941,
      "theoretical_loss": 4.163999700158462,
      "tokens_seen": 289406976
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046075225677031094,
      "loss": 3.0364,
      "theoretical_loss": 4.163888432687784,
      "tokens_seen": 289472512
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004607422266800401,
      "loss": 3.1758,
      "theoretical_loss": 4.163777197456573,
      "tokens_seen": 289538048
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046073219658976936,
      "loss": 2.7997,
      "theoretical_loss": 4.163665994448197,
      "tokens_seen": 289603584
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004607221664994985,
      "loss": 2.9896,
      "theoretical_loss": 4.163554823646027,
      "tokens_seen": 289669120
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004607121364092277,
      "loss": 2.9085,
      "theoretical_loss": 4.163443685033458,
      "tokens_seen": 289734656
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046070210631895685,
      "loss": 2.8829,
      "theoretical_loss": 4.163332578593889,
      "tokens_seen": 289800192
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004606920762286861,
      "loss": 2.846,
      "theoretical_loss": 4.163221504310734,
      "tokens_seen": 289865728
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046068204613841526,
      "loss": 2.9586,
      "theoretical_loss": 4.1631104621674195,
      "tokens_seen": 289931264
    },
    {
      "epoch": 0.09,
      "objective/train/docs_used": 320606,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.6529836654663086,
      "objective/train/theoretical_loss": 4.162999452147384,
      "objective/train/tokens_used": 296742368,
      "theoretical_loss": 4.162999452147384,
      "tokens_seen": 289996800
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046067201604814445,
      "loss": 2.9531,
      "theoretical_loss": 4.162999452147384,
      "tokens_seen": 289996800
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004606619859578736,
      "loss": 3.323,
      "theoretical_loss": 4.1628884742340775,
      "tokens_seen": 290062336
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004606519558676028,
      "loss": 3.1529,
      "theoretical_loss": 4.162777528410963,
      "tokens_seen": 290127872
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000460641925777332,
      "loss": 3.1177,
      "theoretical_loss": 4.162666614661518,
      "tokens_seen": 290193408
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004606318956870612,
      "loss": 2.821,
      "theoretical_loss": 4.162555732969227,
      "tokens_seen": 290258944
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046062186559679035,
      "loss": 2.7999,
      "theoretical_loss": 4.162444883317591,
      "tokens_seen": 290324480
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004606118355065196,
      "loss": 3.0117,
      "theoretical_loss": 4.162334065690123,
      "tokens_seen": 290390016
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046060180541624877,
      "loss": 2.6194,
      "theoretical_loss": 4.162223280070345,
      "tokens_seen": 290455552
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046059177532597795,
      "loss": 2.7548,
      "theoretical_loss": 4.1621125264417955,
      "tokens_seen": 290521088
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046058174523570713,
      "loss": 2.8914,
      "theoretical_loss": 4.162001804788021,
      "tokens_seen": 290586624
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004605717151454363,
      "loss": 2.9791,
      "theoretical_loss": 4.161891115092583,
      "tokens_seen": 290652160
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004605616850551655,
      "loss": 3.3375,
      "theoretical_loss": 4.161780457339055,
      "tokens_seen": 290717696
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004605516549648947,
      "loss": 3.075,
      "theoretical_loss": 4.161669831511022,
      "tokens_seen": 290783232
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046054162487462385,
      "loss": 3.0684,
      "theoretical_loss": 4.16155923759208,
      "tokens_seen": 290848768
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004605315947843531,
      "loss": 3.2013,
      "theoretical_loss": 4.161448675565838,
      "tokens_seen": 290914304
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004605215646940822,
      "loss": 2.8147,
      "theoretical_loss": 4.161338145415918,
      "tokens_seen": 290979840
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046051153460381145,
      "loss": 2.8885,
      "theoretical_loss": 4.161227647125955,
      "tokens_seen": 291045376
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046050150451354063,
      "loss": 2.9984,
      "theoretical_loss": 4.161117180679591,
      "tokens_seen": 291110912
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004604914744232698,
      "loss": 2.9891,
      "theoretical_loss": 4.161006746060488,
      "tokens_seen": 291176448
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000460481444332999,
      "loss": 2.8484,
      "theoretical_loss": 4.160896343252311,
      "tokens_seen": 291241984
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046047141424272823,
      "loss": 2.7959,
      "theoretical_loss": 4.160785972238745,
      "tokens_seen": 291307520
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046046138415245736,
      "loss": 3.0276,
      "theoretical_loss": 4.160675633003484,
      "tokens_seen": 291373056
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004604513540621866,
      "loss": 2.9596,
      "theoretical_loss": 4.16056532553023,
      "tokens_seen": 291438592
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004604413239719157,
      "loss": 2.9352,
      "theoretical_loss": 4.160455049802706,
      "tokens_seen": 291504128
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046043129388164495,
      "loss": 3.0643,
      "theoretical_loss": 4.1603448058046375,
      "tokens_seen": 291569664
    },
    {
      "epoch": 0.09,
      "objective/train/docs_used": 320606,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.340557098388672,
      "objective/train/theoretical_loss": 4.160234593519768,
      "objective/train/tokens_used": 296742368,
      "theoretical_loss": 4.160234593519768,
      "tokens_seen": 291635200
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046042126379137414,
      "loss": 3.2364,
      "theoretical_loss": 4.160234593519768,
      "tokens_seen": 291635200
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004604112337011033,
      "loss": 3.0605,
      "theoretical_loss": 4.160124412931852,
      "tokens_seen": 291700736
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004604012036108325,
      "loss": 2.8302,
      "theoretical_loss": 4.160014264024654,
      "tokens_seen": 291766272
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004603911735205617,
      "loss": 2.7241,
      "theoretical_loss": 4.159904146781952,
      "tokens_seen": 291831808
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046038114343029086,
      "loss": 2.7337,
      "theoretical_loss": 4.159794061187536,
      "tokens_seen": 291897344
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004603711133400201,
      "loss": 3.0597,
      "theoretical_loss": 4.1596840072252075,
      "tokens_seen": 291962880
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004603610832497492,
      "loss": 2.8511,
      "theoretical_loss": 4.159573984878779,
      "tokens_seen": 292028416
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046035105315947846,
      "loss": 3.1022,
      "theoretical_loss": 4.159463994132079,
      "tokens_seen": 292093952
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004603410230692076,
      "loss": 3.2257,
      "theoretical_loss": 4.15935403496894,
      "tokens_seen": 292159488
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004603309929789368,
      "loss": 3.1807,
      "theoretical_loss": 4.159244107373215,
      "tokens_seen": 292225024
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000460320962888666,
      "loss": 2.9728,
      "theoretical_loss": 4.159134211328765,
      "tokens_seen": 292290560
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004603109327983952,
      "loss": 2.9467,
      "theoretical_loss": 4.159024346819461,
      "tokens_seen": 292356096
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046030090270812436,
      "loss": 3.1047,
      "theoretical_loss": 4.158914513829189,
      "tokens_seen": 292421632
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004602908726178536,
      "loss": 3.202,
      "theoretical_loss": 4.158804712341845,
      "tokens_seen": 292487168
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004602808425275827,
      "loss": 3.202,
      "theoretical_loss": 4.158694942341338,
      "tokens_seen": 292552704
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046027081243731196,
      "loss": 3.0828,
      "theoretical_loss": 4.1585852038115885,
      "tokens_seen": 292618240
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004602607823470411,
      "loss": 2.9317,
      "theoretical_loss": 4.1584754967365285,
      "tokens_seen": 292683776
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004602507522567703,
      "loss": 2.9311,
      "theoretical_loss": 4.1583658211001016,
      "tokens_seen": 292749312
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004602407221664995,
      "loss": 3.1576,
      "theoretical_loss": 4.158256176886264,
      "tokens_seen": 292814848
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004602306920762287,
      "loss": 2.8076,
      "theoretical_loss": 4.158146564078982,
      "tokens_seen": 292880384
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046022066198595787,
      "loss": 3.1682,
      "theoretical_loss": 4.158036982662237,
      "tokens_seen": 292945920
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046021063189568705,
      "loss": 2.8877,
      "theoretical_loss": 4.157927432620018,
      "tokens_seen": 293011456
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004602006018054163,
      "loss": 2.8661,
      "theoretical_loss": 4.157817913936329,
      "tokens_seen": 293076992
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046019057171514546,
      "loss": 3.0495,
      "theoretical_loss": 4.157708426595184,
      "tokens_seen": 293142528
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046018054162487465,
      "loss": 3.0307,
      "theoretical_loss": 4.157598970580608,
      "tokens_seen": 293208064
    },
    {
      "epoch": 0.09,
      "objective/train/docs_used": 320606,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6987125873565674,
      "objective/train/theoretical_loss": 4.157489545876642,
      "objective/train/tokens_used": 296742368,
      "theoretical_loss": 4.157489545876642,
      "tokens_seen": 293273600
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004601705115346038,
      "loss": 2.8924,
      "theoretical_loss": 4.157489545876642,
      "tokens_seen": 293273600
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000460160481444333,
      "loss": 2.7811,
      "theoretical_loss": 4.157380152467333,
      "tokens_seen": 293339136
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004601504513540622,
      "loss": 2.929,
      "theoretical_loss": 4.157270790336742,
      "tokens_seen": 293404672
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004601404212637914,
      "loss": 3.0652,
      "theoretical_loss": 4.157161459468944,
      "tokens_seen": 293470208
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046013039117352055,
      "loss": 3.0149,
      "theoretical_loss": 4.157052159848023,
      "tokens_seen": 293535744
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004601203610832498,
      "loss": 2.8461,
      "theoretical_loss": 4.156942891458074,
      "tokens_seen": 293601280
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046011033099297897,
      "loss": 2.9769,
      "theoretical_loss": 4.156833654283207,
      "tokens_seen": 293666816
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046010030090270815,
      "loss": 3.1259,
      "theoretical_loss": 4.15672444830754,
      "tokens_seen": 293732352
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046009027081243733,
      "loss": 3.0044,
      "theoretical_loss": 4.156615273515205,
      "tokens_seen": 293797888
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004600802407221665,
      "loss": 3.1242,
      "theoretical_loss": 4.156506129890344,
      "tokens_seen": 293863424
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004600702106318957,
      "loss": 2.8938,
      "theoretical_loss": 4.156397017417111,
      "tokens_seen": 293928960
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046006018054162493,
      "loss": 2.9038,
      "theoretical_loss": 4.156287936079675,
      "tokens_seen": 293994496
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046005015045135405,
      "loss": 2.9051,
      "theoretical_loss": 4.156178885862209,
      "tokens_seen": 294060032
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004600401203610833,
      "loss": 2.9019,
      "theoretical_loss": 4.156069866748906,
      "tokens_seen": 294125568
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004600300902708124,
      "loss": 2.8916,
      "theoretical_loss": 4.155960878723965,
      "tokens_seen": 294191104
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046002006018054165,
      "loss": 2.8845,
      "theoretical_loss": 4.155851921771598,
      "tokens_seen": 294256640
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046001003009027083,
      "loss": 2.8652,
      "theoretical_loss": 4.155742995876029,
      "tokens_seen": 294322176
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046,
      "loss": 2.8854,
      "theoretical_loss": 4.155634101021494,
      "tokens_seen": 294387712
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004599899699097292,
      "loss": 3.0241,
      "theoretical_loss": 4.155525237192238,
      "tokens_seen": 294453248
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045997993981945843,
      "loss": 3.2217,
      "theoretical_loss": 4.155416404372522,
      "tokens_seen": 294518784
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045996990972918756,
      "loss": 2.8677,
      "theoretical_loss": 4.155307602546614,
      "tokens_seen": 294584320
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004599598796389168,
      "loss": 2.9117,
      "theoretical_loss": 4.155198831698795,
      "tokens_seen": 294649856
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004599498495486459,
      "loss": 2.8769,
      "theoretical_loss": 4.155090091813358,
      "tokens_seen": 294715392
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045993981945837515,
      "loss": 3.2191,
      "theoretical_loss": 4.154981382874608,
      "tokens_seen": 294780928
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045992978936810434,
      "loss": 2.9602,
      "theoretical_loss": 4.154872704866859,
      "tokens_seen": 294846464
    },
    {
      "epoch": 0.09,
      "objective/train/docs_used": 320606,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.4592936038970947,
      "objective/train/theoretical_loss": 4.15476405777444,
      "objective/train/tokens_used": 296742368,
      "theoretical_loss": 4.15476405777444,
      "tokens_seen": 294912000
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004599197592778335,
      "loss": 3.0343,
      "theoretical_loss": 4.15476405777444,
      "tokens_seen": 294912000
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004599097291875627,
      "loss": 2.9556,
      "theoretical_loss": 4.154655441581687,
      "tokens_seen": 294977536
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004598996990972919,
      "loss": 2.9162,
      "theoretical_loss": 4.154546856272952,
      "tokens_seen": 295043072
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045988966900702106,
      "loss": 2.8154,
      "theoretical_loss": 4.154438301832596,
      "tokens_seen": 295108608
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004598796389167503,
      "loss": 3.1434,
      "theoretical_loss": 4.154329778244991,
      "tokens_seen": 295174144
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004598696088264794,
      "loss": 2.7265,
      "theoretical_loss": 4.154221285494521,
      "tokens_seen": 295239680
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045985957873620866,
      "loss": 3.1865,
      "theoretical_loss": 4.154112823565582,
      "tokens_seen": 295305216
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004598495486459378,
      "loss": 2.9111,
      "theoretical_loss": 4.15400439244258,
      "tokens_seen": 295370752
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000459839518555667,
      "loss": 2.9867,
      "theoretical_loss": 4.153895992109935,
      "tokens_seen": 295436288
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004598294884653962,
      "loss": 2.9011,
      "theoretical_loss": 4.153787622552073,
      "tokens_seen": 295501824
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004598194583751254,
      "loss": 3.0471,
      "theoretical_loss": 4.153679283753439,
      "tokens_seen": 295567360
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045980942828485456,
      "loss": 2.909,
      "theoretical_loss": 4.15357097569848,
      "tokens_seen": 295632896
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004597993981945838,
      "loss": 2.7348,
      "theoretical_loss": 4.153462698371665,
      "tokens_seen": 295698432
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004597893681043129,
      "loss": 3.0335,
      "theoretical_loss": 4.1533544517574645,
      "tokens_seen": 295763968
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045977933801404216,
      "loss": 2.9854,
      "theoretical_loss": 4.153246235840367,
      "tokens_seen": 295829504
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004597693079237713,
      "loss": 2.8577,
      "theoretical_loss": 4.153138050604868,
      "tokens_seen": 295895040
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004597592778335005,
      "loss": 3.1334,
      "theoretical_loss": 4.153029896035476,
      "tokens_seen": 295960576
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004597492477432297,
      "loss": 3.1153,
      "theoretical_loss": 4.152921772116712,
      "tokens_seen": 296026112
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004597392176529589,
      "loss": 3.0102,
      "theoretical_loss": 4.152813678833106,
      "tokens_seen": 296091648
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045972918756268807,
      "loss": 2.846,
      "theoretical_loss": 4.152705616169202,
      "tokens_seen": 296157184
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045971915747241725,
      "loss": 3.0878,
      "theoretical_loss": 4.15259758410955,
      "tokens_seen": 296222720
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045970912738214643,
      "loss": 3.0873,
      "theoretical_loss": 4.152489582638719,
      "tokens_seen": 296288256
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045969909729187566,
      "loss": 3.111,
      "theoretical_loss": 4.152381611741281,
      "tokens_seen": 296353792
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004596890672016048,
      "loss": 2.9683,
      "theoretical_loss": 4.152273671401824,
      "tokens_seen": 296419328
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000459679037111334,
      "loss": 3.2746,
      "theoretical_loss": 4.152165761604948,
      "tokens_seen": 296484864
    },
    {
      "epoch": 0.09,
      "objective/train/docs_used": 320606,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.3823139667510986,
      "objective/train/theoretical_loss": 4.152057882335261,
      "objective/train/tokens_used": 296742368,
      "theoretical_loss": 4.152057882335261,
      "tokens_seen": 296550400
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045966900702106315,
      "loss": 2.9545,
      "theoretical_loss": 4.152057882335261,
      "tokens_seen": 296550400
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004596589769307924,
      "loss": 2.8693,
      "theoretical_loss": 4.151950033577383,
      "tokens_seen": 296615936
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045964894684052157,
      "loss": 3.2989,
      "theoretical_loss": 4.151842215315947,
      "tokens_seen": 296681472
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045963891675025075,
      "loss": 3.0708,
      "theoretical_loss": 4.151734427535594,
      "tokens_seen": 296747008
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00045962888665997993,
      "loss": 3.9868,
      "theoretical_loss": 4.151601419005685,
      "tokens_seen": 296827904
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00045961885656970917,
      "loss": 3.0976,
      "theoretical_loss": 4.151493699276069,
      "tokens_seen": 296893440
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0004596088264794383,
      "loss": 3.2042,
      "theoretical_loss": 4.151386009977943,
      "tokens_seen": 296958976
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00045959879638916753,
      "loss": 3.1598,
      "theoretical_loss": 4.151278351095997,
      "tokens_seen": 297024512
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00045958876629889666,
      "loss": 3.2069,
      "theoretical_loss": 4.15117072261493,
      "tokens_seen": 297090048
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0004595787362086259,
      "loss": 2.9252,
      "theoretical_loss": 4.151063124519455,
      "tokens_seen": 297155584
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0004595687061183551,
      "loss": 3.1758,
      "theoretical_loss": 4.150955556794295,
      "tokens_seen": 297221120
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00045955867602808425,
      "loss": 3.1888,
      "theoretical_loss": 4.150848019424184,
      "tokens_seen": 297286656
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00045954864593781344,
      "loss": 3.1973,
      "theoretical_loss": 4.150740512393868,
      "tokens_seen": 297352192
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0004595386158475426,
      "loss": 3.1925,
      "theoretical_loss": 4.1506330356881005,
      "tokens_seen": 297417728
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0004595285857572718,
      "loss": 3.2762,
      "theoretical_loss": 4.150525589291652,
      "tokens_seen": 297483264
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00045951855566700103,
      "loss": 3.2021,
      "theoretical_loss": 4.150418173189299,
      "tokens_seen": 297548800
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00045950852557673016,
      "loss": 3.1805,
      "theoretical_loss": 4.1503107873658305,
      "tokens_seen": 297614336
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0004594984954864594,
      "loss": 3.3438,
      "theoretical_loss": 4.150203431806046,
      "tokens_seen": 297679872
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0004594884653961885,
      "loss": 3.1938,
      "theoretical_loss": 4.150096106494758,
      "tokens_seen": 297745408
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00045947843530591776,
      "loss": 3.1626,
      "theoretical_loss": 4.149988811416788,
      "tokens_seen": 297810944
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00045946840521564694,
      "loss": 3.1176,
      "theoretical_loss": 4.149881546556971,
      "tokens_seen": 297876480
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0004594583751253761,
      "loss": 3.0244,
      "theoretical_loss": 4.149774311900147,
      "tokens_seen": 297942016
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00045944834503510536,
      "loss": 3.0722,
      "theoretical_loss": 4.149667107431174,
      "tokens_seen": 298007552
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00045943831494483454,
      "loss": 3.0816,
      "theoretical_loss": 4.149559933134916,
      "tokens_seen": 298073088
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0004594282848545637,
      "loss": 3.2045,
      "theoretical_loss": 4.149452788996252,
      "tokens_seen": 298138624
    },
    {
      "epoch": 1.0,
      "objective/train/docs_used": 388086,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.5071983337402344,
      "objective/train/theoretical_loss": 4.149372450674081,
      "objective/train/tokens_used": 318647776,
      "theoretical_loss": 4.149372450674081,
      "tokens_seen": 298187776
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0004594182547642929,
      "loss": 3.0893,
      "theoretical_loss": 4.149345675000067,
      "tokens_seen": 298204160
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0004594082246740221,
      "loss": 3.2211,
      "theoretical_loss": 4.149238591131261,
      "tokens_seen": 298269696
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00045939819458375126,
      "loss": 3.1047,
      "theoretical_loss": 4.149131537374743,
      "tokens_seen": 298335232
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0004593881644934805,
      "loss": 3.1427,
      "theoretical_loss": 4.149024513715434,
      "tokens_seen": 298400768
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0004593781344032096,
      "loss": 3.2841,
      "theoretical_loss": 4.148917520138264,
      "tokens_seen": 298466304
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00045936810431293886,
      "loss": 3.1813,
      "theoretical_loss": 4.148810556628176,
      "tokens_seen": 298531840
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.000459358074222668,
      "loss": 3.0694,
      "theoretical_loss": 4.148703623170123,
      "tokens_seen": 298597376
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0004593480441323972,
      "loss": 3.0848,
      "theoretical_loss": 4.148596719749067,
      "tokens_seen": 298662912
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0004593380140421264,
      "loss": 3.1606,
      "theoretical_loss": 4.148489846349984,
      "tokens_seen": 298728448
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0004593279839518556,
      "loss": 3.001,
      "theoretical_loss": 4.14838300295786,
      "tokens_seen": 298793984
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00045931795386158476,
      "loss": 3.2776,
      "theoretical_loss": 4.148276189557689,
      "tokens_seen": 298859520
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.000459307923771314,
      "loss": 3.1101,
      "theoretical_loss": 4.148169406134479,
      "tokens_seen": 298925056
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0004592978936810431,
      "loss": 3.2241,
      "theoretical_loss": 4.148062652673248,
      "tokens_seen": 298990592
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00045928786359077236,
      "loss": 3.2768,
      "theoretical_loss": 4.147955929159024,
      "tokens_seen": 299056128
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0004592778335005015,
      "loss": 3.3648,
      "theoretical_loss": 4.147849235576846,
      "tokens_seen": 299121664
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0004592678034102307,
      "loss": 3.039,
      "theoretical_loss": 4.147742571911765,
      "tokens_seen": 299187200
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0004592577733199599,
      "loss": 3.0152,
      "theoretical_loss": 4.147635938148841,
      "tokens_seen": 299252736
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0004592477432296891,
      "loss": 3.1572,
      "theoretical_loss": 4.147529334273145,
      "tokens_seen": 299318272
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00045923771313941827,
      "loss": 3.2587,
      "theoretical_loss": 4.147422760269759,
      "tokens_seen": 299383808
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00045922768304914745,
      "loss": 3.0774,
      "theoretical_loss": 4.147316216123777,
      "tokens_seen": 299449344
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00045921765295887663,
      "loss": 3.2984,
      "theoretical_loss": 4.147209701820302,
      "tokens_seen": 299514880
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00045920762286860586,
      "loss": 3.0177,
      "theoretical_loss": 4.147103217344448,
      "tokens_seen": 299580416
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.000459197592778335,
      "loss": 3.1661,
      "theoretical_loss": 4.14699676268134,
      "tokens_seen": 299645952
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0004591875626880642,
      "loss": 2.9458,
      "theoretical_loss": 4.146890337816114,
      "tokens_seen": 299711488
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00045917753259779335,
      "loss": 2.9969,
      "theoretical_loss": 4.146783942733915,
      "tokens_seen": 299777024
    },
    {
      "epoch": 1.0,
      "objective/train/docs_used": 393339,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.2248926162719727,
      "objective/train/theoretical_loss": 4.1467041659584485,
      "objective/train/tokens_used": 320286176,
      "theoretical_loss": 4.1467041659584485,
      "tokens_seen": 299826176
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0004591675025075226,
      "loss": 3.0643,
      "theoretical_loss": 4.146677577419902,
      "tokens_seen": 299842560
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00045915747241725177,
      "loss": 3.0779,
      "theoretical_loss": 4.14657124185924,
      "tokens_seen": 299908096
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00045914744232698095,
      "loss": 3.1616,
      "theoretical_loss": 4.14646493603711,
      "tokens_seen": 299973632
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00045913741223671013,
      "loss": 3.0127,
      "theoretical_loss": 4.146358659938699,
      "tokens_seen": 300039168
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00045912738214643937,
      "loss": 3.0727,
      "theoretical_loss": 4.146252413549207,
      "tokens_seen": 300104704
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0004591173520561685,
      "loss": 3.2627,
      "theoretical_loss": 4.146146196853843,
      "tokens_seen": 300170240
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00045910732196589773,
      "loss": 3.1219,
      "theoretical_loss": 4.146040009837829,
      "tokens_seen": 300235776
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00045909729187562686,
      "loss": 3.2786,
      "theoretical_loss": 4.145933852486396,
      "tokens_seen": 300301312
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0004590872617853561,
      "loss": 3.0808,
      "theoretical_loss": 4.145827724784784,
      "tokens_seen": 300366848
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0004590772316950853,
      "loss": 3.054,
      "theoretical_loss": 4.145721626718247,
      "tokens_seen": 300432384
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00045906720160481445,
      "loss": 3.0884,
      "theoretical_loss": 4.1456155582720475,
      "tokens_seen": 300497920
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00045905717151454364,
      "loss": 3.0052,
      "theoretical_loss": 4.145509519431459,
      "tokens_seen": 300563456
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0004590471414242728,
      "loss": 2.9324,
      "theoretical_loss": 4.145403510181765,
      "tokens_seen": 300628992
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.000459037111334002,
      "loss": 3.1987,
      "theoretical_loss": 4.145297530508261,
      "tokens_seen": 300694528
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00045902708124373123,
      "loss": 3.245,
      "theoretical_loss": 4.145191580396251,
      "tokens_seen": 300760064
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00045901705115346036,
      "loss": 3.2895,
      "theoretical_loss": 4.145085659831049,
      "tokens_seen": 300825600
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0004590070210631896,
      "loss": 3.3486,
      "theoretical_loss": 4.144979768797985,
      "tokens_seen": 300891136
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0004589969909729187,
      "loss": 3.0806,
      "theoretical_loss": 4.14487390728239,
      "tokens_seen": 300956672
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00045898696088264796,
      "loss": 2.9642,
      "theoretical_loss": 4.144768075269616,
      "tokens_seen": 301022208
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00045897693079237714,
      "loss": 3.1238,
      "theoretical_loss": 4.144662272745018,
      "tokens_seen": 301087744
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0004589669007021063,
      "loss": 3.0449,
      "theoretical_loss": 4.144556499693964,
      "tokens_seen": 301153280
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0004589568706118355,
      "loss": 3.1233,
      "theoretical_loss": 4.144450756101832,
      "tokens_seen": 301218816
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00045894684052156474,
      "loss": 3.1373,
      "theoretical_loss": 4.144345041954011,
      "tokens_seen": 301284352
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00045893681043129386,
      "loss": 3.047,
      "theoretical_loss": 4.1442393572359,
      "tokens_seen": 301349888
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0004589267803410231,
      "loss": 3.0807,
      "theoretical_loss": 4.14413370193291,
      "tokens_seen": 301415424
    },
    {
      "epoch": 1.0,
      "objective/train/docs_used": 398212,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.0288240909576416,
      "objective/train/theoretical_loss": 4.144054479750567,
      "objective/train/tokens_used": 321924576,
      "theoretical_loss": 4.144054479750567,
      "tokens_seen": 301464576
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0004589167502507522,
      "loss": 3.0242,
      "theoretical_loss": 4.144028076030458,
      "tokens_seen": 301480960
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00045890672016048146,
      "loss": 3.2372,
      "theoretical_loss": 4.143922479513977,
      "tokens_seen": 301546496
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00045889669007021064,
      "loss": 3.141,
      "theoretical_loss": 4.143816912368906,
      "tokens_seen": 301612032
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0004588866599799398,
      "loss": 3.2979,
      "theoretical_loss": 4.143711374580697,
      "tokens_seen": 301677568
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.000458876629889669,
      "loss": 3.1576,
      "theoretical_loss": 4.143605866134811,
      "tokens_seen": 301743104
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0004588665997993982,
      "loss": 3.1577,
      "theoretical_loss": 4.1435003870167195,
      "tokens_seen": 301808640
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00045885656970912737,
      "loss": 3.1194,
      "theoretical_loss": 4.143394937211906,
      "tokens_seen": 301874176
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0004588465396188566,
      "loss": 3.0641,
      "theoretical_loss": 4.143289516705861,
      "tokens_seen": 301939712
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00045883650952858573,
      "loss": 3.1287,
      "theoretical_loss": 4.14318412548409,
      "tokens_seen": 302005248
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00045882647943831496,
      "loss": 3.1244,
      "theoretical_loss": 4.143078763532104,
      "tokens_seen": 302070784
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00045881644934804415,
      "loss": 3.2705,
      "theoretical_loss": 4.1429734308354265,
      "tokens_seen": 302136320
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0004588064192577733,
      "loss": 3.2505,
      "theoretical_loss": 4.142868127379592,
      "tokens_seen": 302201856
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0004587963891675025,
      "loss": 3.0527,
      "theoretical_loss": 4.142762853150145,
      "tokens_seen": 302267392
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0004587863590772317,
      "loss": 3.0989,
      "theoretical_loss": 4.142657608132638,
      "tokens_seen": 302332928
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00045877632898696087,
      "loss": 3.3695,
      "theoretical_loss": 4.142552392312638,
      "tokens_seen": 302398464
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0004587662988966901,
      "loss": 3.1914,
      "theoretical_loss": 4.142447205675717,
      "tokens_seen": 302464000
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00045875626880641923,
      "loss": 3.3743,
      "theoretical_loss": 4.142342048207462,
      "tokens_seen": 302529536
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00045874623871614847,
      "loss": 3.2538,
      "theoretical_loss": 4.1422369198934685,
      "tokens_seen": 302595072
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0004587362086258776,
      "loss": 3.2982,
      "theoretical_loss": 4.142131820719342,
      "tokens_seen": 302660608
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00045872617853560683,
      "loss": 3.1992,
      "theoretical_loss": 4.142026750670697,
      "tokens_seen": 302726144
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.000458716148445336,
      "loss": 3.1915,
      "theoretical_loss": 4.14192170973316,
      "tokens_seen": 302791680
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0004587061183550652,
      "loss": 3.3022,
      "theoretical_loss": 4.141816697892368,
      "tokens_seen": 302857216
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0004586960882647944,
      "loss": 3.347,
      "theoretical_loss": 4.141711715133967,
      "tokens_seen": 302922752
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00045868605817452355,
      "loss": 3.2762,
      "theoretical_loss": 4.1416067614436125,
      "tokens_seen": 302988288
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0004586760280842528,
      "loss": 3.1287,
      "theoretical_loss": 4.141501836806973,
      "tokens_seen": 303053824
    },
    {
      "epoch": 1.0,
      "objective/train/docs_used": 403119,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8905088901519775,
      "objective/train/theoretical_loss": 4.141423162387376,
      "objective/train/tokens_used": 323562976,
      "theoretical_loss": 4.141423162387376,
      "tokens_seen": 303102976
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00045866599799398197,
      "loss": 3.0808,
      "theoretical_loss": 4.141396941209724,
      "tokens_seen": 303119360
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00045865596790371115,
      "loss": 3.2722,
      "theoretical_loss": 4.141292074637554,
      "tokens_seen": 303184896
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00045864593781344033,
      "loss": 3.4675,
      "theoretical_loss": 4.141187237076158,
      "tokens_seen": 303250432
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00045863590772316957,
      "loss": 3.2572,
      "theoretical_loss": 4.141082428511247,
      "tokens_seen": 303315968
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0004586258776328987,
      "loss": 3.2468,
      "theoretical_loss": 4.140977648928534,
      "tokens_seen": 303381504
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00045861584754262793,
      "loss": 3.025,
      "theoretical_loss": 4.14087289831375,
      "tokens_seen": 303447040
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00045860581745235706,
      "loss": 3.2632,
      "theoretical_loss": 4.140768176652632,
      "tokens_seen": 303512576
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0004585957873620863,
      "loss": 3.4513,
      "theoretical_loss": 4.1406634839309255,
      "tokens_seen": 303578112
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0004585857572718155,
      "loss": 3.307,
      "theoretical_loss": 4.140558820134391,
      "tokens_seen": 303643648
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00045857572718154465,
      "loss": 3.2101,
      "theoretical_loss": 4.140454185248797,
      "tokens_seen": 303709184
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00045856569709127384,
      "loss": 3.3302,
      "theoretical_loss": 4.140349579259919,
      "tokens_seen": 303774720
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.000458555667001003,
      "loss": 3.2633,
      "theoretical_loss": 4.140245002153547,
      "tokens_seen": 303840256
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0004585456369107322,
      "loss": 3.0562,
      "theoretical_loss": 4.140140453915478,
      "tokens_seen": 303905792
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00045853560682046143,
      "loss": 3.1859,
      "theoretical_loss": 4.140035934531521,
      "tokens_seen": 303971328
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00045852557673019056,
      "loss": 3.1359,
      "theoretical_loss": 4.139931443987494,
      "tokens_seen": 304036864
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0004585155466399198,
      "loss": 3.4099,
      "theoretical_loss": 4.139826982269225,
      "tokens_seen": 304102400
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0004585055165496489,
      "loss": 3.2253,
      "theoretical_loss": 4.139722549362553,
      "tokens_seen": 304167936
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00045849548645937816,
      "loss": 3.1415,
      "theoretical_loss": 4.139618145253326,
      "tokens_seen": 304233472
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00045848545636910734,
      "loss": 3.0069,
      "theoretical_loss": 4.139513769927402,
      "tokens_seen": 304299008
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0004584754262788365,
      "loss": 3.3609,
      "theoretical_loss": 4.139409423370649,
      "tokens_seen": 304364544
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0004584653961885657,
      "loss": 3.3372,
      "theoretical_loss": 4.139305105568946,
      "tokens_seen": 304430080
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00045845536609829494,
      "loss": 3.124,
      "theoretical_loss": 4.139200816508181,
      "tokens_seen": 304495616
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00045844533600802406,
      "loss": 3.1771,
      "theoretical_loss": 4.139096556174252,
      "tokens_seen": 304561152
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0004584353059177533,
      "loss": 3.2137,
      "theoretical_loss": 4.138992324553068,
      "tokens_seen": 304626688
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0004584252758274824,
      "loss": 3.2045,
      "theoretical_loss": 4.138888121630545,
      "tokens_seen": 304692224
    },
    {
      "epoch": 1.0,
      "objective/train/docs_used": 408173,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.1989989280700684,
      "objective/train/theoretical_loss": 4.138809988263685,
      "objective/train/tokens_used": 325201376,
      "theoretical_loss": 4.138809988263685,
      "tokens_seen": 304741376
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00045841524573721166,
      "loss": 3.3398,
      "theoretical_loss": 4.138783947392613,
      "tokens_seen": 304757760
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00045840521564694084,
      "loss": 3.1126,
      "theoretical_loss": 4.13867980182521,
      "tokens_seen": 304823296
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00045839518555667,
      "loss": 3.1801,
      "theoretical_loss": 4.138575684914282,
      "tokens_seen": 304888832
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0004583851554663992,
      "loss": 3.2703,
      "theoretical_loss": 4.138471596645789,
      "tokens_seen": 304954368
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0004583751253761284,
      "loss": 3.2164,
      "theoretical_loss": 4.138367537005697,
      "tokens_seen": 305019904
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00045836509528585757,
      "loss": 3.3049,
      "theoretical_loss": 4.1382635059799835,
      "tokens_seen": 305085440
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0004583550651955868,
      "loss": 2.9696,
      "theoretical_loss": 4.138159503554638,
      "tokens_seen": 305150976
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00045834503510531593,
      "loss": 3.3098,
      "theoretical_loss": 4.138055529715655,
      "tokens_seen": 305216512
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00045833500501504516,
      "loss": 3.1387,
      "theoretical_loss": 4.137951584449044,
      "tokens_seen": 305282048
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00045832497492477435,
      "loss": 3.1771,
      "theoretical_loss": 4.13784766774082,
      "tokens_seen": 305347584
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0004583149448345035,
      "loss": 3.171,
      "theoretical_loss": 4.137743779577011,
      "tokens_seen": 305413120
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0004583049147442327,
      "loss": 3.009,
      "theoretical_loss": 4.137639919943655,
      "tokens_seen": 305478656
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0004582948846539619,
      "loss": 3.0046,
      "theoretical_loss": 4.137536088826796,
      "tokens_seen": 305544192
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00045828485456369107,
      "loss": 3.0266,
      "theoretical_loss": 4.1374322862124915,
      "tokens_seen": 305609728
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0004582748244734203,
      "loss": 2.9813,
      "theoretical_loss": 4.137328512086807,
      "tokens_seen": 305675264
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00045826479438314943,
      "loss": 3.0888,
      "theoretical_loss": 4.1372247664358195,
      "tokens_seen": 305740800
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00045825476429287867,
      "loss": 2.9889,
      "theoretical_loss": 4.137121049245614,
      "tokens_seen": 305806336
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0004582447342026078,
      "loss": 2.9691,
      "theoretical_loss": 4.137017360502286,
      "tokens_seen": 305871872
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00045823470411233703,
      "loss": 2.893,
      "theoretical_loss": 4.136913700191942,
      "tokens_seen": 305937408
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0004582246740220662,
      "loss": 3.1183,
      "theoretical_loss": 4.136810068300694,
      "tokens_seen": 306002944
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0004582146439317954,
      "loss": 3.2618,
      "theoretical_loss": 4.13670646481467,
      "tokens_seen": 306068480
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00045820461384152457,
      "loss": 3.2482,
      "theoretical_loss": 4.1366028897200025,
      "tokens_seen": 306134016
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00045819458375125375,
      "loss": 3.1096,
      "theoretical_loss": 4.1364993430028365,
      "tokens_seen": 306199552
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00045818455366098294,
      "loss": 3.4132,
      "theoretical_loss": 4.136395824649327,
      "tokens_seen": 306265088
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00045817452357071217,
      "loss": 3.1493,
      "theoretical_loss": 4.136292334645636,
      "tokens_seen": 306330624
    },
    {
      "epoch": 1.0,
      "objective/train/docs_used": 413344,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.146451950073242,
      "objective/train/theoretical_loss": 4.136214735739118,
      "objective/train/tokens_used": 326839776,
      "theoretical_loss": 4.136214735739118,
      "tokens_seen": 306379776
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0004581644934804413,
      "loss": 3.0853,
      "theoretical_loss": 4.136188872977938,
      "tokens_seen": 306396160
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00045815446339017053,
      "loss": 3.0507,
      "theoretical_loss": 4.1360854396324145,
      "tokens_seen": 306461696
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0004581444332998997,
      "loss": 3.0544,
      "theoretical_loss": 4.13598203459526,
      "tokens_seen": 306527232
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0004581344032096289,
      "loss": 3.0592,
      "theoretical_loss": 4.135878657852677,
      "tokens_seen": 306592768
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0004581243731193581,
      "loss": 3.1452,
      "theoretical_loss": 4.135775309390876,
      "tokens_seen": 306658304
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00045811434302908726,
      "loss": 2.9714,
      "theoretical_loss": 4.135671989196081,
      "tokens_seen": 306723840
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00045810431293881644,
      "loss": 3.279,
      "theoretical_loss": 4.135568697254522,
      "tokens_seen": 306789376
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0004580942828485457,
      "loss": 3.2136,
      "theoretical_loss": 4.135465433552441,
      "tokens_seen": 306854912
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0004580842527582748,
      "loss": 3.1741,
      "theoretical_loss": 4.135362198076088,
      "tokens_seen": 306920448
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00045807422266800404,
      "loss": 3.2101,
      "theoretical_loss": 4.135258990811723,
      "tokens_seen": 306985984
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00045806419257773316,
      "loss": 3.0828,
      "theoretical_loss": 4.1351558117456175,
      "tokens_seen": 307051520
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0004580541624874624,
      "loss": 3.0378,
      "theoretical_loss": 4.13505266086405,
      "tokens_seen": 307117056
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0004580441323971916,
      "loss": 3.0833,
      "theoretical_loss": 4.134949538153309,
      "tokens_seen": 307182592
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00045803410230692076,
      "loss": 3.0974,
      "theoretical_loss": 4.134846443599694,
      "tokens_seen": 307248128
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00045802407221664994,
      "loss": 3.1697,
      "theoretical_loss": 4.1347433771895155,
      "tokens_seen": 307313664
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0004580140421263791,
      "loss": 3.1566,
      "theoretical_loss": 4.134640338909088,
      "tokens_seen": 307379200
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0004580040120361083,
      "loss": 2.9549,
      "theoretical_loss": 4.134537328744742,
      "tokens_seen": 307444736
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00045799398194583754,
      "loss": 3.0834,
      "theoretical_loss": 4.134434346682812,
      "tokens_seen": 307510272
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00045798395185556667,
      "loss": 2.92,
      "theoretical_loss": 4.134331392709647,
      "tokens_seen": 307575808
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0004579739217652959,
      "loss": 3.018,
      "theoretical_loss": 4.134228466811601,
      "tokens_seen": 307641344
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0004579638916750251,
      "loss": 3.0011,
      "theoretical_loss": 4.1341255689750405,
      "tokens_seen": 307706880
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00045795386158475426,
      "loss": 3.1451,
      "theoretical_loss": 4.134022699186342,
      "tokens_seen": 307772416
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0004579438314944835,
      "loss": 3.0998,
      "theoretical_loss": 4.133919857431889,
      "tokens_seen": 307837952
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0004579338014042126,
      "loss": 3.0204,
      "theoretical_loss": 4.133817043698075,
      "tokens_seen": 307903488
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00045792377131394186,
      "loss": 3.0224,
      "theoretical_loss": 4.133714257971306,
      "tokens_seen": 307969024
    },
    {
      "epoch": 1.0,
      "objective/train/docs_used": 418392,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.903918981552124,
      "objective/train/theoretical_loss": 4.133637187047677,
      "objective/train/tokens_used": 328478176,
      "theoretical_loss": 4.133637187047677,
      "tokens_seen": 308018176
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00045791374122367104,
      "loss": 2.9841,
      "theoretical_loss": 4.133611500237993,
      "tokens_seen": 308034560
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0004579037111334002,
      "loss": 2.9786,
      "theoretical_loss": 4.133508770484561,
      "tokens_seen": 308100096
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0004578936810431294,
      "loss": 3.0327,
      "theoretical_loss": 4.13340606869744,
      "tokens_seen": 308165632
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0004578836509528586,
      "loss": 3.1817,
      "theoretical_loss": 4.133303394863072,
      "tokens_seen": 308231168
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00045787362086258777,
      "loss": 2.8875,
      "theoretical_loss": 4.133200748967909,
      "tokens_seen": 308296704
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.000457863590772317,
      "loss": 3.3509,
      "theoretical_loss": 4.133098130998412,
      "tokens_seen": 308362240
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00045785356068204613,
      "loss": 3.14,
      "theoretical_loss": 4.1329955409410495,
      "tokens_seen": 308427776
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00045784353059177536,
      "loss": 3.1157,
      "theoretical_loss": 4.132892978782301,
      "tokens_seen": 308493312
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00045783350050150455,
      "loss": 3.2767,
      "theoretical_loss": 4.1327904445086565,
      "tokens_seen": 308558848
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0004578234704112337,
      "loss": 3.1308,
      "theoretical_loss": 4.132687938106614,
      "tokens_seen": 308624384
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0004578134403209629,
      "loss": 2.9828,
      "theoretical_loss": 4.132585459562681,
      "tokens_seen": 308689920
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0004578034102306921,
      "loss": 3.2487,
      "theoretical_loss": 4.132483008863374,
      "tokens_seen": 308755456
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00045779338014042127,
      "loss": 2.948,
      "theoretical_loss": 4.13238058599522,
      "tokens_seen": 308820992
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0004577833500501505,
      "loss": 3.2227,
      "theoretical_loss": 4.132278190944755,
      "tokens_seen": 308886528
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00045777331995987963,
      "loss": 3.115,
      "theoretical_loss": 4.132175823698523,
      "tokens_seen": 308952064
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00045776328986960887,
      "loss": 3.1668,
      "theoretical_loss": 4.1320734842430795,
      "tokens_seen": 309017600
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.000457753259779338,
      "loss": 3.0767,
      "theoretical_loss": 4.131971172564989,
      "tokens_seen": 309083136
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00045774322968906723,
      "loss": 2.9926,
      "theoretical_loss": 4.131868888650824,
      "tokens_seen": 309148672
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0004577331995987964,
      "loss": 3.0552,
      "theoretical_loss": 4.1317666324871665,
      "tokens_seen": 309214208
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0004577231695085256,
      "loss": 3.0081,
      "theoretical_loss": 4.13166440406061,
      "tokens_seen": 309279744
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0004577131394182548,
      "loss": 2.9077,
      "theoretical_loss": 4.131562203357753,
      "tokens_seen": 309345280
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00045770310932798395,
      "loss": 3.0159,
      "theoretical_loss": 4.131460030365209,
      "tokens_seen": 309410816
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00045769307923771314,
      "loss": 3.2991,
      "theoretical_loss": 4.131357885069596,
      "tokens_seen": 309476352
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00045768304914744237,
      "loss": 3.2686,
      "theoretical_loss": 4.131255767457543,
      "tokens_seen": 309541888
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0004576730190571715,
      "loss": 2.9812,
      "theoretical_loss": 4.13115367751569,
      "tokens_seen": 309607424
    },
    {
      "epoch": 1.0,
      "objective/train/docs_used": 423122,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.105621099472046,
      "objective/train/theoretical_loss": 4.131077128209836,
      "objective/train/tokens_used": 330116576,
      "theoretical_loss": 4.131077128209836,
      "tokens_seen": 309656576
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00045766298896690073,
      "loss": 3.2017,
      "theoretical_loss": 4.1310516152306835,
      "tokens_seen": 309672960
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0004576529588766299,
      "loss": 3.1174,
      "theoretical_loss": 4.130949580589181,
      "tokens_seen": 309738496
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0004576429287863591,
      "loss": 3.0827,
      "theoretical_loss": 4.130847573577848,
      "tokens_seen": 309804032
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0004576328986960883,
      "loss": 3.1589,
      "theoretical_loss": 4.13074559418336,
      "tokens_seen": 309869568
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00045762286860581746,
      "loss": 3.2419,
      "theoretical_loss": 4.1306436423924024,
      "tokens_seen": 309935104
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00045761283851554664,
      "loss": 3.0622,
      "theoretical_loss": 4.130541718191669,
      "tokens_seen": 310000640
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0004576028084252759,
      "loss": 3.0627,
      "theoretical_loss": 4.130439821567863,
      "tokens_seen": 310066176
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.000457592778335005,
      "loss": 2.9596,
      "theoretical_loss": 4.130337952507695,
      "tokens_seen": 310131712
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00045758274824473424,
      "loss": 3.1407,
      "theoretical_loss": 4.130236110997889,
      "tokens_seen": 310197248
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00045757271815446336,
      "loss": 3.0654,
      "theoretical_loss": 4.130134297025175,
      "tokens_seen": 310262784
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0004575626880641926,
      "loss": 3.1397,
      "theoretical_loss": 4.130032510576292,
      "tokens_seen": 310328320
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0004575526579739218,
      "loss": 3.171,
      "theoretical_loss": 4.129930751637991,
      "tokens_seen": 310393856
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00045754262788365096,
      "loss": 3.2332,
      "theoretical_loss": 4.129829020197029,
      "tokens_seen": 310459392
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00045753259779338014,
      "loss": 3.0021,
      "theoretical_loss": 4.1297273162401735,
      "tokens_seen": 310524928
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0004575225677031093,
      "loss": 2.9555,
      "theoretical_loss": 4.1296256397542015,
      "tokens_seen": 310590464
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0004575125376128385,
      "loss": 3.1127,
      "theoretical_loss": 4.1295239907258985,
      "tokens_seen": 310656000
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00045750250752256774,
      "loss": 3.0528,
      "theoretical_loss": 4.12942236914206,
      "tokens_seen": 310721536
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00045749247743229687,
      "loss": 3.0231,
      "theoretical_loss": 4.129320774989489,
      "tokens_seen": 310787072
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0004574824473420261,
      "loss": 3.1342,
      "theoretical_loss": 4.129219208255002,
      "tokens_seen": 310852608
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0004574724172517553,
      "loss": 3.113,
      "theoretical_loss": 4.129117668925415,
      "tokens_seen": 310918144
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00045746238716148446,
      "loss": 2.9051,
      "theoretical_loss": 4.129016156987566,
      "tokens_seen": 310983680
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00045745235707121364,
      "loss": 2.9875,
      "theoretical_loss": 4.128914672428293,
      "tokens_seen": 311049216
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0004574423269809428,
      "loss": 3.06,
      "theoretical_loss": 4.128813215234444,
      "tokens_seen": 311114752
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.000457432296890672,
      "loss": 3.0676,
      "theoretical_loss": 4.128711785392879,
      "tokens_seen": 311180288
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00045742226680040124,
      "loss": 3.1011,
      "theoretical_loss": 4.128610382890467,
      "tokens_seen": 311245824
    },
    {
      "epoch": 1.0,
      "objective/train/docs_used": 428239,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.340728282928467,
      "objective/train/theoretical_loss": 4.128534348947081,
      "objective/train/tokens_used": 331754976,
      "theoretical_loss": 4.128534348947081,
      "tokens_seen": 311294976
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00045741223671013037,
      "loss": 3.1619,
      "theoretical_loss": 4.128509007714083,
      "tokens_seen": 311311360
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0004574022066198596,
      "loss": 3.0496,
      "theoretical_loss": 4.128407659850614,
      "tokens_seen": 311376896
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00045739217652958873,
      "loss": 2.8642,
      "theoretical_loss": 4.128306339286954,
      "tokens_seen": 311442432
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00045738214643931797,
      "loss": 3.0832,
      "theoretical_loss": 4.1282050460100095,
      "tokens_seen": 311507968
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00045737211634904715,
      "loss": 3.1083,
      "theoretical_loss": 4.12810378000669,
      "tokens_seen": 311573504
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00045736208625877633,
      "loss": 2.8529,
      "theoretical_loss": 4.12800254126392,
      "tokens_seen": 311639040
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0004573520561685055,
      "loss": 3.1388,
      "theoretical_loss": 4.12790132976863,
      "tokens_seen": 311704576
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00045734202607823475,
      "loss": 2.9444,
      "theoretical_loss": 4.127800145507759,
      "tokens_seen": 311770112
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00045733199598796387,
      "loss": 2.8887,
      "theoretical_loss": 4.127698988468259,
      "tokens_seen": 311835648
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0004573219658976931,
      "loss": 3.1635,
      "theoretical_loss": 4.127597858637085,
      "tokens_seen": 311901184
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00045731193580742223,
      "loss": 3.0093,
      "theoretical_loss": 4.1274967560012055,
      "tokens_seen": 311966720
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00045730190571715147,
      "loss": 3.1035,
      "theoretical_loss": 4.127395680547597,
      "tokens_seen": 312032256
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00045729187562688065,
      "loss": 3.207,
      "theoretical_loss": 4.127294632263243,
      "tokens_seen": 312097792
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00045728184553660983,
      "loss": 3.0375,
      "theoretical_loss": 4.127193611135139,
      "tokens_seen": 312163328
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.000457271815446339,
      "loss": 2.8885,
      "theoretical_loss": 4.1270926171502875,
      "tokens_seen": 312228864
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0004572617853560682,
      "loss": 3.0487,
      "theoretical_loss": 4.1269916502957,
      "tokens_seen": 312294400
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0004572517552657974,
      "loss": 3.0653,
      "theoretical_loss": 4.126890710558398,
      "tokens_seen": 312359936
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0004572417251755266,
      "loss": 2.9576,
      "theoretical_loss": 4.126789797925411,
      "tokens_seen": 312425472
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00045723169508525574,
      "loss": 3.2364,
      "theoretical_loss": 4.126688912383777,
      "tokens_seen": 312491008
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.000457221664994985,
      "loss": 3.2365,
      "theoretical_loss": 4.126588053920545,
      "tokens_seen": 312556544
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0004572116349047141,
      "loss": 3.1703,
      "theoretical_loss": 4.126487222522771,
      "tokens_seen": 312622080
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00045720160481444334,
      "loss": 3.1546,
      "theoretical_loss": 4.12638641817752,
      "tokens_seen": 312687616
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00045719157472417257,
      "loss": 3.2756,
      "theoretical_loss": 4.126285640871867,
      "tokens_seen": 312753152
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0004571815446339017,
      "loss": 3.1369,
      "theoretical_loss": 4.126184890592894,
      "tokens_seen": 312818688
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00045717151454363093,
      "loss": 2.9366,
      "theoretical_loss": 4.126084167327694,
      "tokens_seen": 312884224
    },
    {
      "epoch": 1.0,
      "objective/train/docs_used": 433098,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.1568994522094727,
      "objective/train/theoretical_loss": 4.126008642598823,
      "objective/train/tokens_used": 333393376,
      "theoretical_loss": 4.126008642598823,
      "tokens_seen": 312933376
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0004571614844533601,
      "loss": 3.0267,
      "theoretical_loss": 4.125983471063368,
      "tokens_seen": 312949760
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0004571514543630893,
      "loss": 3.0007,
      "theoretical_loss": 4.125882801787025,
      "tokens_seen": 313015296
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0004571414242728185,
      "loss": 3.1487,
      "theoretical_loss": 4.125782159485785,
      "tokens_seen": 313080832
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00045713139418254766,
      "loss": 3.1603,
      "theoretical_loss": 4.125681544146774,
      "tokens_seen": 313146368
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00045712136409227684,
      "loss": 3.1322,
      "theoretical_loss": 4.12558095575713,
      "tokens_seen": 313211904
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004571113340020061,
      "loss": 3.1128,
      "theoretical_loss": 4.125480394303996,
      "tokens_seen": 313277440
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004571013039117352,
      "loss": 3.0045,
      "theoretical_loss": 4.125379859774528,
      "tokens_seen": 313342976
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045709127382146444,
      "loss": 3.0059,
      "theoretical_loss": 4.1252793521558875,
      "tokens_seen": 313408512
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045708124373119356,
      "loss": 2.9702,
      "theoretical_loss": 4.125178871435247,
      "tokens_seen": 313474048
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004570712136409228,
      "loss": 3.2466,
      "theoretical_loss": 4.125078417599785,
      "tokens_seen": 313539584
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.000457061183550652,
      "loss": 2.8958,
      "theoretical_loss": 4.124977990636694,
      "tokens_seen": 313605120
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045705115346038116,
      "loss": 2.9426,
      "theoretical_loss": 4.124877590533169,
      "tokens_seen": 313670656
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045704112337011034,
      "loss": 3.1383,
      "theoretical_loss": 4.124777217276417,
      "tokens_seen": 313736192
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004570310932798395,
      "loss": 2.9676,
      "theoretical_loss": 4.124676870853656,
      "tokens_seen": 313801728
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004570210631895687,
      "loss": 2.6934,
      "theoretical_loss": 4.124576551252107,
      "tokens_seen": 313867264
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045701103309929794,
      "loss": 2.9895,
      "theoretical_loss": 4.124476258459005,
      "tokens_seen": 313932800
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045700100300902707,
      "loss": 3.3011,
      "theoretical_loss": 4.124375992461592,
      "tokens_seen": 313998336
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004569909729187563,
      "loss": 2.9931,
      "theoretical_loss": 4.124275753247116,
      "tokens_seen": 314063872
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004569809428284855,
      "loss": 3.0106,
      "theoretical_loss": 4.124175540802839,
      "tokens_seen": 314129408
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045697091273821466,
      "loss": 3.0805,
      "theoretical_loss": 4.124075355116028,
      "tokens_seen": 314194944
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045696088264794384,
      "loss": 3.1193,
      "theoretical_loss": 4.123975196173959,
      "tokens_seen": 314260480
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.000456950852557673,
      "loss": 2.9561,
      "theoretical_loss": 4.123875063963917,
      "tokens_seen": 314326016
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004569408224674022,
      "loss": 3.0758,
      "theoretical_loss": 4.123774958473197,
      "tokens_seen": 314391552
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045693079237713144,
      "loss": 3.1406,
      "theoretical_loss": 4.123674879689101,
      "tokens_seen": 314457088
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045692076228686057,
      "loss": 3.1404,
      "theoretical_loss": 4.123574827598942,
      "tokens_seen": 314522624
    },
    {
      "epoch": 1.01,
      "objective/train/docs_used": 438319,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.2052862644195557,
      "objective/train/theoretical_loss": 4.12349980604159,
      "objective/train/tokens_used": 335031776,
      "theoretical_loss": 4.12349980604159,
      "tokens_seen": 314571776
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004569107321965898,
      "loss": 3.1546,
      "theoretical_loss": 4.123474802190039,
      "tokens_seen": 314588160
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045690070210631893,
      "loss": 3.1286,
      "theoretical_loss": 4.123374803449719,
      "tokens_seen": 314653696
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045689067201604817,
      "loss": 3.022,
      "theoretical_loss": 4.123274831365322,
      "tokens_seen": 314719232
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045688064192577735,
      "loss": 3.0278,
      "theoretical_loss": 4.123174885924192,
      "tokens_seen": 314784768
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045687061183550653,
      "loss": 3.1057,
      "theoretical_loss": 4.123074967113686,
      "tokens_seen": 314850304
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004568605817452357,
      "loss": 3.2642,
      "theoretical_loss": 4.122975074921165,
      "tokens_seen": 314915840
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045685055165496495,
      "loss": 2.8919,
      "theoretical_loss": 4.122875209334002,
      "tokens_seen": 314981376
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045684052156469407,
      "loss": 3.119,
      "theoretical_loss": 4.122775370339577,
      "tokens_seen": 315046912
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004568304914744233,
      "loss": 3.188,
      "theoretical_loss": 4.12267555792528,
      "tokens_seen": 315112448
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045682046138415243,
      "loss": 3.0429,
      "theoretical_loss": 4.122575772078507,
      "tokens_seen": 315177984
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045681043129388167,
      "loss": 3.2101,
      "theoretical_loss": 4.122476012786667,
      "tokens_seen": 315243520
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045680040120361085,
      "loss": 3.2976,
      "theoretical_loss": 4.122376280037173,
      "tokens_seen": 315309056
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045679037111334003,
      "loss": 3.0171,
      "theoretical_loss": 4.122276573817448,
      "tokens_seen": 315374592
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004567803410230692,
      "loss": 3.0178,
      "theoretical_loss": 4.122176894114926,
      "tokens_seen": 315440128
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004567703109327984,
      "loss": 3.2494,
      "theoretical_loss": 4.122077240917046,
      "tokens_seen": 315505664
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004567602808425276,
      "loss": 2.9009,
      "theoretical_loss": 4.121977614211259,
      "tokens_seen": 315571200
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004567502507522568,
      "loss": 3.2204,
      "theoretical_loss": 4.12187801398502,
      "tokens_seen": 315636736
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045674022066198594,
      "loss": 3.1532,
      "theoretical_loss": 4.121778440225798,
      "tokens_seen": 315702272
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004567301905717152,
      "loss": 3.0529,
      "theoretical_loss": 4.121678892921066,
      "tokens_seen": 315767808
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004567201604814443,
      "loss": 2.8736,
      "theoretical_loss": 4.121579372058308,
      "tokens_seen": 315833344
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045671013039117354,
      "loss": 3.103,
      "theoretical_loss": 4.121479877625016,
      "tokens_seen": 315898880
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004567001003009027,
      "loss": 3.0901,
      "theoretical_loss": 4.12138040960869,
      "tokens_seen": 315964416
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004566900702106319,
      "loss": 2.9924,
      "theoretical_loss": 4.12128096799684,
      "tokens_seen": 316029952
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004566800401203611,
      "loss": 3.1131,
      "theoretical_loss": 4.1211815527769815,
      "tokens_seen": 316095488
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004566700100300903,
      "loss": 3.2011,
      "theoretical_loss": 4.1210821639366415,
      "tokens_seen": 316161024
    },
    {
      "epoch": 1.01,
      "objective/train/docs_used": 440664,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.3091273307800293,
      "objective/train/theoretical_loss": 4.121007639610445,
      "objective/train/tokens_used": 336670176,
      "theoretical_loss": 4.121007639610445,
      "tokens_seen": 316210176
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045665997993981944,
      "loss": 3.1517,
      "theoretical_loss": 4.120982801463353,
      "tokens_seen": 316226560
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004566499498495487,
      "loss": 3.0728,
      "theoretical_loss": 4.1208834653446615,
      "tokens_seen": 316292096
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004566399197592778,
      "loss": 3.2325,
      "theoretical_loss": 4.120784155568115,
      "tokens_seen": 316357632
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045662988966900704,
      "loss": 3.1194,
      "theoretical_loss": 4.120684872121277,
      "tokens_seen": 316423168
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004566198595787362,
      "loss": 3.1201,
      "theoretical_loss": 4.120585614991712,
      "tokens_seen": 316488704
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004566098294884654,
      "loss": 3.1454,
      "theoretical_loss": 4.120486384166998,
      "tokens_seen": 316554240
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004565997993981946,
      "loss": 2.9571,
      "theoretical_loss": 4.120387179634721,
      "tokens_seen": 316619776
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045658976930792376,
      "loss": 3.1203,
      "theoretical_loss": 4.120288001382474,
      "tokens_seen": 316685312
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045657973921765294,
      "loss": 3.1862,
      "theoretical_loss": 4.1201888493978585,
      "tokens_seen": 316750848
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004565697091273822,
      "loss": 3.0073,
      "theoretical_loss": 4.120089723668484,
      "tokens_seen": 316816384
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004565596790371113,
      "loss": 3.1276,
      "theoretical_loss": 4.1199906241819715,
      "tokens_seen": 316881920
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045654964894684054,
      "loss": 3.1453,
      "theoretical_loss": 4.119891550925946,
      "tokens_seen": 316947456
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045653961885656967,
      "loss": 3.1468,
      "theoretical_loss": 4.119792503888043,
      "tokens_seen": 317012992
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004565295887662989,
      "loss": 3.0982,
      "theoretical_loss": 4.119693483055908,
      "tokens_seen": 317078528
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004565195586760281,
      "loss": 2.9135,
      "theoretical_loss": 4.119594488417192,
      "tokens_seen": 317144064
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045650952858575727,
      "loss": 3.1709,
      "theoretical_loss": 4.119495519959556,
      "tokens_seen": 317209600
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045649949849548645,
      "loss": 3.1193,
      "theoretical_loss": 4.119396577670669,
      "tokens_seen": 317275136
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004564894684052157,
      "loss": 3.1036,
      "theoretical_loss": 4.119297661538209,
      "tokens_seen": 317340672
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004564794383149448,
      "loss": 2.9798,
      "theoretical_loss": 4.11919877154986,
      "tokens_seen": 317406208
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045646940822467405,
      "loss": 3.0293,
      "theoretical_loss": 4.119099907693318,
      "tokens_seen": 317471744
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045645937813440317,
      "loss": 3.0721,
      "theoretical_loss": 4.119001069956284,
      "tokens_seen": 317537280
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004564493480441324,
      "loss": 3.3166,
      "theoretical_loss": 4.11890225832647,
      "tokens_seen": 317602816
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045643931795386164,
      "loss": 2.9832,
      "theoretical_loss": 4.118803472791592,
      "tokens_seen": 317668352
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045642928786359077,
      "loss": 3.1933,
      "theoretical_loss": 4.1187047133393815,
      "tokens_seen": 317733888
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045641925777332,
      "loss": 3.0072,
      "theoretical_loss": 4.118605979957572,
      "tokens_seen": 317799424
    },
    {
      "epoch": 1.01,
      "objective/train/docs_used": 445676,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.2730116844177246,
      "objective/train/theoretical_loss": 4.118531947022541,
      "objective/train/tokens_used": 338308576,
      "theoretical_loss": 4.118531947022541,
      "tokens_seen": 317848576
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045640922768304913,
      "loss": 3.2766,
      "theoretical_loss": 4.118507272633906,
      "tokens_seen": 317864960
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045639919759277837,
      "loss": 3.1235,
      "theoretical_loss": 4.118408591356138,
      "tokens_seen": 317930496
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045638916750250755,
      "loss": 2.7607,
      "theoretical_loss": 4.118309936112027,
      "tokens_seen": 317996032
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045637913741223673,
      "loss": 3.1125,
      "theoretical_loss": 4.118211306889343,
      "tokens_seen": 318061568
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004563691073219659,
      "loss": 3.1064,
      "theoretical_loss": 4.1181127036758625,
      "tokens_seen": 318127104
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045635907723169515,
      "loss": 3.0764,
      "theoretical_loss": 4.118014126459371,
      "tokens_seen": 318192640
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045634904714142427,
      "loss": 3.127,
      "theoretical_loss": 4.11791557522766,
      "tokens_seen": 318258176
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004563390170511535,
      "loss": 3.0333,
      "theoretical_loss": 4.117817049968535,
      "tokens_seen": 318323712
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045632898696088263,
      "loss": 3.21,
      "theoretical_loss": 4.117718550669802,
      "tokens_seen": 318389248
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045631895687061187,
      "loss": 3.0226,
      "theoretical_loss": 4.117620077319282,
      "tokens_seen": 318454784
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045630892678034105,
      "loss": 3.2592,
      "theoretical_loss": 4.117521629904801,
      "tokens_seen": 318520320
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045629889669007023,
      "loss": 2.9392,
      "theoretical_loss": 4.117423208414193,
      "tokens_seen": 318585856
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004562888665997994,
      "loss": 3.0316,
      "theoretical_loss": 4.117324812835301,
      "tokens_seen": 318651392
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004562788365095286,
      "loss": 3.1168,
      "theoretical_loss": 4.117226443155976,
      "tokens_seen": 318716928
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004562688064192578,
      "loss": 3.0642,
      "theoretical_loss": 4.117128099364078,
      "tokens_seen": 318782464
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.000456258776328987,
      "loss": 2.8883,
      "theoretical_loss": 4.1170297814474734,
      "tokens_seen": 318848000
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045624874623871614,
      "loss": 3.0011,
      "theoretical_loss": 4.116931489394039,
      "tokens_seen": 318913536
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004562387161484454,
      "loss": 2.9764,
      "theoretical_loss": 4.116833223191657,
      "tokens_seen": 318979072
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004562286860581745,
      "loss": 2.9898,
      "theoretical_loss": 4.116734982828222,
      "tokens_seen": 319044608
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045621865596790374,
      "loss": 3.1809,
      "theoretical_loss": 4.116636768291631,
      "tokens_seen": 319110144
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004562086258776329,
      "loss": 3.1512,
      "theoretical_loss": 4.116538579569793,
      "tokens_seen": 319175680
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004561985957873621,
      "loss": 3.0586,
      "theoretical_loss": 4.116440416650627,
      "tokens_seen": 319241216
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004561885656970913,
      "loss": 3.0317,
      "theoretical_loss": 4.116342279522055,
      "tokens_seen": 319306752
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004561785356068205,
      "loss": 3.145,
      "theoretical_loss": 4.11624416817201,
      "tokens_seen": 319372288
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045616850551654964,
      "loss": 2.9133,
      "theoretical_loss": 4.1161460825884335,
      "tokens_seen": 319437824
    },
    {
      "epoch": 1.01,
      "objective/train/docs_used": 450671,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.054537773132324,
      "objective/train/theoretical_loss": 4.116072535302745,
      "objective/train/tokens_used": 339946976,
      "theoretical_loss": 4.116072535302745,
      "tokens_seen": 319486976
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004561584754262789,
      "loss": 2.8832,
      "theoretical_loss": 4.116048022759274,
      "tokens_seen": 319503360
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.000456148445336008,
      "loss": 3.0356,
      "theoretical_loss": 4.115949988672488,
      "tokens_seen": 319568896
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045613841524573724,
      "loss": 3.0226,
      "theoretical_loss": 4.115851980316042,
      "tokens_seen": 319634432
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004561283851554664,
      "loss": 2.753,
      "theoretical_loss": 4.115753997677907,
      "tokens_seen": 319699968
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004561183550651956,
      "loss": 3.0136,
      "theoretical_loss": 4.115656040746067,
      "tokens_seen": 319765504
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004561083249749248,
      "loss": 3.0264,
      "theoretical_loss": 4.11555810950851,
      "tokens_seen": 319831040
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045609829488465396,
      "loss": 3.0399,
      "theoretical_loss": 4.1154602039532335,
      "tokens_seen": 319896576
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045608826479438314,
      "loss": 2.8733,
      "theoretical_loss": 4.115362324068242,
      "tokens_seen": 319962112
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004560782347041124,
      "loss": 3.123,
      "theoretical_loss": 4.115264469841551,
      "tokens_seen": 320027648
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004560682046138415,
      "loss": 3.013,
      "theoretical_loss": 4.115166641261181,
      "tokens_seen": 320093184
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045605817452357074,
      "loss": 3.0235,
      "theoretical_loss": 4.115068838315162,
      "tokens_seen": 320158720
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045604814443329987,
      "loss": 2.9962,
      "theoretical_loss": 4.1149710609915315,
      "tokens_seen": 320224256
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004560381143430291,
      "loss": 3.0287,
      "theoretical_loss": 4.114873309278335,
      "tokens_seen": 320289792
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004560280842527583,
      "loss": 2.9697,
      "theoretical_loss": 4.114775583163627,
      "tokens_seen": 320355328
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045601805416248747,
      "loss": 2.974,
      "theoretical_loss": 4.114677882635468,
      "tokens_seen": 320420864
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045600802407221665,
      "loss": 3.1124,
      "theoretical_loss": 4.114580207681929,
      "tokens_seen": 320486400
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004559979939819459,
      "loss": 3.2231,
      "theoretical_loss": 4.114482558291087,
      "tokens_seen": 320551936
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.000455987963891675,
      "loss": 2.8844,
      "theoretical_loss": 4.114384934451029,
      "tokens_seen": 320617472
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045597793380140425,
      "loss": 3.105,
      "theoretical_loss": 4.114287336149847,
      "tokens_seen": 320683008
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045596790371113337,
      "loss": 2.9922,
      "theoretical_loss": 4.114189763375643,
      "tokens_seen": 320748544
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004559578736208626,
      "loss": 3.102,
      "theoretical_loss": 4.1140922161165285,
      "tokens_seen": 320814080
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004559478435305918,
      "loss": 3.1323,
      "theoretical_loss": 4.11399469436062,
      "tokens_seen": 320879616
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045593781344032097,
      "loss": 2.9403,
      "theoretical_loss": 4.113897198096042,
      "tokens_seen": 320945152
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045592778335005015,
      "loss": 3.1139,
      "theoretical_loss": 4.11379972731093,
      "tokens_seen": 321010688
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045591775325977933,
      "loss": 2.9772,
      "theoretical_loss": 4.113702281993424,
      "tokens_seen": 321076224
    },
    {
      "epoch": 1.01,
      "objective/train/docs_used": 452128,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.911926031112671,
      "objective/train/theoretical_loss": 4.113629214711283,
      "objective/train/tokens_used": 341585376,
      "theoretical_loss": 4.113629214711283,
      "tokens_seen": 321125376
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004559077231695085,
      "loss": 2.9501,
      "theoretical_loss": 4.113604862131675,
      "tokens_seen": 321141760
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045589769307923775,
      "loss": 2.9992,
      "theoretical_loss": 4.113507467713839,
      "tokens_seen": 321207296
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004558876629889669,
      "loss": 3.1123,
      "theoretical_loss": 4.113410098728081,
      "tokens_seen": 321272832
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004558776328986961,
      "loss": 3.1029,
      "theoretical_loss": 4.1133127551625766,
      "tokens_seen": 321338368
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045586760280842524,
      "loss": 3.1395,
      "theoretical_loss": 4.113215437005504,
      "tokens_seen": 321403904
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004558575727181545,
      "loss": 2.9044,
      "theoretical_loss": 4.113118144245054,
      "tokens_seen": 321469440
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045584754262788365,
      "loss": 3.1156,
      "theoretical_loss": 4.113020876869424,
      "tokens_seen": 321534976
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045583751253761284,
      "loss": 3.1783,
      "theoretical_loss": 4.112923634866817,
      "tokens_seen": 321600512
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.000455827482447342,
      "loss": 3.0768,
      "theoretical_loss": 4.112826418225447,
      "tokens_seen": 321666048
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045581745235707125,
      "loss": 2.8923,
      "theoretical_loss": 4.112729226933536,
      "tokens_seen": 321731584
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004558074222668004,
      "loss": 2.8801,
      "theoretical_loss": 4.112632060979309,
      "tokens_seen": 321797120
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004557973921765296,
      "loss": 2.9205,
      "theoretical_loss": 4.112534920351004,
      "tokens_seen": 321862656
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045578736208625874,
      "loss": 2.9779,
      "theoretical_loss": 4.112437805036866,
      "tokens_seen": 321928192
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.000455777331995988,
      "loss": 3.0556,
      "theoretical_loss": 4.112340715025147,
      "tokens_seen": 321993728
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045576730190571716,
      "loss": 2.9769,
      "theoretical_loss": 4.112243650304106,
      "tokens_seen": 322059264
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045575727181544634,
      "loss": 3.1575,
      "theoretical_loss": 4.11214661086201,
      "tokens_seen": 322124800
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004557472417251755,
      "loss": 3.2281,
      "theoretical_loss": 4.112049596687136,
      "tokens_seen": 322190336
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004557372116349047,
      "loss": 3.2923,
      "theoretical_loss": 4.111952607767767,
      "tokens_seen": 322255872
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004557271815446339,
      "loss": 3.1374,
      "theoretical_loss": 4.111855644092194,
      "tokens_seen": 322321408
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004557171514543631,
      "loss": 2.9752,
      "theoretical_loss": 4.111758705648716,
      "tokens_seen": 322386944
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045570712136409224,
      "loss": 2.9913,
      "theoretical_loss": 4.11166179242564,
      "tokens_seen": 322452480
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004556970912738215,
      "loss": 2.955,
      "theoretical_loss": 4.11156490441128,
      "tokens_seen": 322518016
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004556870611835507,
      "loss": 3.1211,
      "theoretical_loss": 4.111468041593958,
      "tokens_seen": 322583552
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045567703109327984,
      "loss": 2.7598,
      "theoretical_loss": 4.111371203962006,
      "tokens_seen": 322649088
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004556670010030091,
      "loss": 3.1663,
      "theoretical_loss": 4.11127439150376,
      "tokens_seen": 322714624
    },
    {
      "epoch": 1.01,
      "objective/train/docs_used": 452464,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7435483932495117,
      "objective/train/theoretical_loss": 4.111201798673309,
      "objective/train/tokens_used": 343223776,
      "theoretical_loss": 4.111201798673309,
      "tokens_seen": 322763776
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004556569709127382,
      "loss": 3.0986,
      "theoretical_loss": 4.111177604207566,
      "tokens_seen": 322780160
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045564694082246744,
      "loss": 3.123,
      "theoretical_loss": 4.111080842061779,
      "tokens_seen": 322845696
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004556369107321966,
      "loss": 2.9448,
      "theoretical_loss": 4.110984105054758,
      "tokens_seen": 322911232
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004556268806419258,
      "loss": 2.6903,
      "theoretical_loss": 4.110887393174874,
      "tokens_seen": 322976768
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.000455616850551655,
      "loss": 3.3049,
      "theoretical_loss": 4.110790706410502,
      "tokens_seen": 323042304
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045560682046138416,
      "loss": 3.0815,
      "theoretical_loss": 4.110694044750028,
      "tokens_seen": 323107840
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045559679037111334,
      "loss": 3.0381,
      "theoretical_loss": 4.110597408181842,
      "tokens_seen": 323173376
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004555867602808426,
      "loss": 3.1317,
      "theoretical_loss": 4.110500796694346,
      "tokens_seen": 323238912
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004555767301905717,
      "loss": 2.9622,
      "theoretical_loss": 4.110404210275947,
      "tokens_seen": 323304448
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045556670010030094,
      "loss": 3.0428,
      "theoretical_loss": 4.110307648915059,
      "tokens_seen": 323369984
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045555667001003007,
      "loss": 3.0631,
      "theoretical_loss": 4.110211112600107,
      "tokens_seen": 323435520
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004555466399197593,
      "loss": 3.0437,
      "theoretical_loss": 4.11011460131952,
      "tokens_seen": 323501056
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004555366098294885,
      "loss": 3.08,
      "theoretical_loss": 4.110018115061737,
      "tokens_seen": 323566592
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045552657973921767,
      "loss": 2.8587,
      "theoretical_loss": 4.109921653815205,
      "tokens_seen": 323632128
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045551654964894685,
      "loss": 2.8589,
      "theoretical_loss": 4.109825217568375,
      "tokens_seen": 323697664
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004555065195586761,
      "loss": 2.9758,
      "theoretical_loss": 4.109728806309711,
      "tokens_seen": 323763200
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004554964894684052,
      "loss": 3.0621,
      "theoretical_loss": 4.1096324200276815,
      "tokens_seen": 323828736
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045548645937813445,
      "loss": 3.2006,
      "theoretical_loss": 4.109536058710763,
      "tokens_seen": 323894272
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045547642928786357,
      "loss": 3.2291,
      "theoretical_loss": 4.109439722347439,
      "tokens_seen": 323959808
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004554663991975928,
      "loss": 3.0339,
      "theoretical_loss": 4.109343410926202,
      "tokens_seen": 324025344
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.000455456369107322,
      "loss": 3.1038,
      "theoretical_loss": 4.109247124435552,
      "tokens_seen": 324090880
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045544633901705117,
      "loss": 3.2273,
      "theoretical_loss": 4.109150862863997,
      "tokens_seen": 324156416
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045543630892678035,
      "loss": 3.0505,
      "theoretical_loss": 4.1090546262000505,
      "tokens_seen": 324221952
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045542627883650953,
      "loss": 3.101,
      "theoretical_loss": 4.108958414432235,
      "tokens_seen": 324287488
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004554162487462387,
      "loss": 3.1216,
      "theoretical_loss": 4.10886222754908,
      "tokens_seen": 324353024
    },
    {
      "epoch": 1.01,
      "objective/train/docs_used": 453833,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.195937395095825,
      "objective/train/theoretical_loss": 4.108790103710379,
      "objective/train/tokens_used": 344862176,
      "theoretical_loss": 4.108790103710379,
      "tokens_seen": 324402176
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045540621865596795,
      "loss": 3.2008,
      "theoretical_loss": 4.108766065539125,
      "tokens_seen": 324418560
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004553961885656971,
      "loss": 3.0328,
      "theoretical_loss": 4.108669928390915,
      "tokens_seen": 324484096
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004553861584754263,
      "loss": 3.0984,
      "theoretical_loss": 4.108573816093003,
      "tokens_seen": 324549632
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045537612838515544,
      "loss": 2.8293,
      "theoretical_loss": 4.108477728633948,
      "tokens_seen": 324615168
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004553660982948847,
      "loss": 3.2014,
      "theoretical_loss": 4.108381666002318,
      "tokens_seen": 324680704
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045535606820461385,
      "loss": 3.0183,
      "theoretical_loss": 4.10828562818669,
      "tokens_seen": 324746240
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045534603811434304,
      "loss": 3.1596,
      "theoretical_loss": 4.108189615175646,
      "tokens_seen": 324811776
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004553360080240722,
      "loss": 2.7589,
      "theoretical_loss": 4.1080936269577775,
      "tokens_seen": 324877312
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045532597793380145,
      "loss": 2.8328,
      "theoretical_loss": 4.107997663521683,
      "tokens_seen": 324942848
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004553159478435306,
      "loss": 2.7061,
      "theoretical_loss": 4.107901724855966,
      "tokens_seen": 325008384
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004553059177532598,
      "loss": 2.8632,
      "theoretical_loss": 4.107805810949244,
      "tokens_seen": 325073920
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045529588766298894,
      "loss": 2.9472,
      "theoretical_loss": 4.107709921790134,
      "tokens_seen": 325139456
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004552858575727182,
      "loss": 2.9933,
      "theoretical_loss": 4.1076140573672655,
      "tokens_seen": 325204992
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045527582748244736,
      "loss": 3.0292,
      "theoretical_loss": 4.107518217669275,
      "tokens_seen": 325270528
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045526579739217654,
      "loss": 2.9263,
      "theoretical_loss": 4.107422402684806,
      "tokens_seen": 325336064
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004552557673019057,
      "loss": 3.1346,
      "theoretical_loss": 4.107326612402509,
      "tokens_seen": 325401600
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004552457372116349,
      "loss": 3.1047,
      "theoretical_loss": 4.107230846811042,
      "tokens_seen": 325467136
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004552357071213641,
      "loss": 3.0838,
      "theoretical_loss": 4.107135105899072,
      "tokens_seen": 325532672
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004552256770310933,
      "loss": 3.2631,
      "theoretical_loss": 4.107039389655271,
      "tokens_seen": 325598208
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045521564694082244,
      "loss": 2.9412,
      "theoretical_loss": 4.106943698068321,
      "tokens_seen": 325663744
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004552056168505517,
      "loss": 3.1445,
      "theoretical_loss": 4.106848031126909,
      "tokens_seen": 325729280
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004551955867602808,
      "loss": 2.9177,
      "theoretical_loss": 4.106752388819732,
      "tokens_seen": 325794816
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045518555667001004,
      "loss": 2.8788,
      "theoretical_loss": 4.106656771135494,
      "tokens_seen": 325860352
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004551755265797392,
      "loss": 3.0533,
      "theoretical_loss": 4.106561178062903,
      "tokens_seen": 325925888
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004551654964894684,
      "loss": 3.2877,
      "theoretical_loss": 4.10646560959068,
      "tokens_seen": 325991424
    },
    {
      "epoch": 1.01,
      "objective/train/docs_used": 455028,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.201298236846924,
      "objective/train/theoretical_loss": 4.10639394937372,
      "objective/train/tokens_used": 346500576,
      "theoretical_loss": 4.10639394937372,
      "tokens_seen": 326040576
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004551554663991976,
      "loss": 3.1642,
      "theoretical_loss": 4.106370065707549,
      "tokens_seen": 326056960
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004551454363089268,
      "loss": 2.8599,
      "theoretical_loss": 4.1062745464022425,
      "tokens_seen": 326122496
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045513540621865595,
      "loss": 2.9966,
      "theoretical_loss": 4.106179051663502,
      "tokens_seen": 326188032
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004551253761283852,
      "loss": 3.0048,
      "theoretical_loss": 4.106083581480076,
      "tokens_seen": 326253568
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004551153460381143,
      "loss": 3.1238,
      "theoretical_loss": 4.105988135840717,
      "tokens_seen": 326319104
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045510531594784354,
      "loss": 2.9494,
      "theoretical_loss": 4.105892714734191,
      "tokens_seen": 326384640
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004550952858575727,
      "loss": 2.8344,
      "theoretical_loss": 4.105797318149266,
      "tokens_seen": 326450176
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004550852557673019,
      "loss": 3.2039,
      "theoretical_loss": 4.105701946074721,
      "tokens_seen": 326515712
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004550752256770311,
      "loss": 3.1371,
      "theoretical_loss": 4.105606598499339,
      "tokens_seen": 326581248
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045506519558676027,
      "loss": 3.0273,
      "theoretical_loss": 4.105511275411914,
      "tokens_seen": 326646784
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045505516549648945,
      "loss": 2.9177,
      "theoretical_loss": 4.105415976801243,
      "tokens_seen": 326712320
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004550451354062187,
      "loss": 3.1699,
      "theoretical_loss": 4.105320702656137,
      "tokens_seen": 326777856
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004550351053159478,
      "loss": 3.0427,
      "theoretical_loss": 4.105225452965406,
      "tokens_seen": 326843392
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045502507522567705,
      "loss": 3.0283,
      "theoretical_loss": 4.105130227717876,
      "tokens_seen": 326908928
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045501504513540623,
      "loss": 3.0798,
      "theoretical_loss": 4.105035026902372,
      "tokens_seen": 326974464
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004550050150451354,
      "loss": 3.1388,
      "theoretical_loss": 4.104939850507733,
      "tokens_seen": 327040000
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004549949849548646,
      "loss": 2.9548,
      "theoretical_loss": 4.104844698522801,
      "tokens_seen": 327105536
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045498495486459377,
      "loss": 3.1927,
      "theoretical_loss": 4.104749570936429,
      "tokens_seen": 327171072
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045497492477432295,
      "loss": 2.939,
      "theoretical_loss": 4.104654467737474,
      "tokens_seen": 327236608
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004549648946840522,
      "loss": 3.1495,
      "theoretical_loss": 4.104559388914802,
      "tokens_seen": 327302144
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045495486459378137,
      "loss": 3.1292,
      "theoretical_loss": 4.104464334457286,
      "tokens_seen": 327367680
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045494483450351055,
      "loss": 3.1137,
      "theoretical_loss": 4.104369304353806,
      "tokens_seen": 327433216
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045493480441323973,
      "loss": 3.1885,
      "theoretical_loss": 4.104274298593252,
      "tokens_seen": 327498752
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004549247743229689,
      "loss": 3.1253,
      "theoretical_loss": 4.1041793171645145,
      "tokens_seen": 327564288
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045491474423269815,
      "loss": 2.9915,
      "theoretical_loss": 4.104084360056499,
      "tokens_seen": 327629824
    },
    {
      "debugging/Self-BLEU-5": 0.6136948618909069,
      "debugging/distinct-1-grams": 0.7435805096055491,
      "debugging/distinct-2-grams": 0.8810311841604012,
      "debugging/entropy-1-grams": 6.325230867052365,
      "debugging/entropy-2-grams": 7.6356135460409735,
      "debugging/length": 511.3030303030303,
      "debugging/num_segments": 33,
      "debugging/score": 0.0005877098914164371,
      "debugging/score_std": 0.0013283427361280467,
      "epoch": 1.01,
      "objective/train/docs_used": 455830,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.0165796279907227,
      "objective/train/theoretical_loss": 4.1040131581792885,
      "objective/train/tokens_used": 348138976,
      "theoretical_loss": 4.1040131581792885,
      "tokens_seen": 327678976
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004549047141424273,
      "loss": 2.8767,
      "theoretical_loss": 4.103989427258114,
      "tokens_seen": 327695360
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004548946840521565,
      "loss": 3.0342,
      "theoretical_loss": 4.103894518758277,
      "tokens_seen": 327760896
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045488465396188564,
      "loss": 3.0731,
      "theoretical_loss": 4.1037996345459105,
      "tokens_seen": 327826432
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004548746238716149,
      "loss": 3.1697,
      "theoretical_loss": 4.103704774609946,
      "tokens_seen": 327891968
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045486459378134405,
      "loss": 3.1106,
      "theoretical_loss": 4.103609938939324,
      "tokens_seen": 327957504
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045485456369107324,
      "loss": 3.1378,
      "theoretical_loss": 4.103515127522988,
      "tokens_seen": 328023040
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004548445336008024,
      "loss": 2.9621,
      "theoretical_loss": 4.103420340349893,
      "tokens_seen": 328088576
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045483450351053165,
      "loss": 2.9669,
      "theoretical_loss": 4.103325577408997,
      "tokens_seen": 328154112
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004548244734202608,
      "loss": 3.0197,
      "theoretical_loss": 4.103230838689269,
      "tokens_seen": 328219648
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045481444332999,
      "loss": 3.0518,
      "theoretical_loss": 4.103136124179684,
      "tokens_seen": 328285184
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045480441323971914,
      "loss": 3.2399,
      "theoretical_loss": 4.103041433869223,
      "tokens_seen": 328350720
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004547943831494484,
      "loss": 3.0063,
      "theoretical_loss": 4.102946767746875,
      "tokens_seen": 328416256
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045478435305917756,
      "loss": 2.907,
      "theoretical_loss": 4.102852125801638,
      "tokens_seen": 328481792
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045477432296890674,
      "loss": 3.197,
      "theoretical_loss": 4.102757508022513,
      "tokens_seen": 328547328
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004547642928786359,
      "loss": 3.2354,
      "theoretical_loss": 4.102662914398513,
      "tokens_seen": 328612864
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004547542627883651,
      "loss": 2.9079,
      "theoretical_loss": 4.102568344918655,
      "tokens_seen": 328678400
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004547442326980943,
      "loss": 2.8512,
      "theoretical_loss": 4.102473799571964,
      "tokens_seen": 328743936
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004547342026078235,
      "loss": 3.0689,
      "theoretical_loss": 4.102379278347472,
      "tokens_seen": 328809472
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045472417251755264,
      "loss": 2.9659,
      "theoretical_loss": 4.10228478123422,
      "tokens_seen": 328875008
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004547141424272819,
      "loss": 3.2956,
      "theoretical_loss": 4.102190308221253,
      "tokens_seen": 328940544
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.000454704112337011,
      "loss": 3.0001,
      "theoretical_loss": 4.102095859297625,
      "tokens_seen": 329006080
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045469408224674024,
      "loss": 3.0739,
      "theoretical_loss": 4.102001434452398,
      "tokens_seen": 329071616
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004546840521564694,
      "loss": 3.0976,
      "theoretical_loss": 4.101907033674639,
      "tokens_seen": 329137152
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004546740220661986,
      "loss": 3.0629,
      "theoretical_loss": 4.101812656953424,
      "tokens_seen": 329202688
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004546639919759278,
      "loss": 2.9582,
      "theoretical_loss": 4.101718304277834,
      "tokens_seen": 329268224
    },
    {
      "epoch": 1.01,
      "objective/train/docs_used": 456654,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.36710524559021,
      "objective/train/theoretical_loss": 4.10164755554452,
      "objective/train/tokens_used": 349777376,
      "theoretical_loss": 4.10164755554452,
      "tokens_seen": 329317376
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.000454653961885657,
      "loss": 2.9395,
      "theoretical_loss": 4.10162397563696,
      "tokens_seen": 329333760
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045464393179538615,
      "loss": 3.1366,
      "theoretical_loss": 4.101529671019898,
      "tokens_seen": 329399296
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004546339017051154,
      "loss": 2.7547,
      "theoretical_loss": 4.101435390415752,
      "tokens_seen": 329464832
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004546238716148445,
      "loss": 3.222,
      "theoretical_loss": 4.101341133813632,
      "tokens_seen": 329530368
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045461384152457374,
      "loss": 2.9597,
      "theoretical_loss": 4.101246901202655,
      "tokens_seen": 329595904
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004546038114343029,
      "loss": 3.0324,
      "theoretical_loss": 4.101152692571949,
      "tokens_seen": 329661440
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004545937813440321,
      "loss": 3.1991,
      "theoretical_loss": 4.101058507910645,
      "tokens_seen": 329726976
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004545837512537613,
      "loss": 3.0136,
      "theoretical_loss": 4.100964347207881,
      "tokens_seen": 329792512
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045457372116349047,
      "loss": 3.1253,
      "theoretical_loss": 4.100870210452804,
      "tokens_seen": 329858048
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045456369107321965,
      "loss": 3.1968,
      "theoretical_loss": 4.100776097634567,
      "tokens_seen": 329923584
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004545536609829489,
      "loss": 3.0509,
      "theoretical_loss": 4.100682008742331,
      "tokens_seen": 329989120
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.000454543630892678,
      "loss": 2.961,
      "theoretical_loss": 4.100587943765264,
      "tokens_seen": 330054656
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045453360080240725,
      "loss": 2.9026,
      "theoretical_loss": 4.100493902692539,
      "tokens_seen": 330120192
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045452357071213643,
      "loss": 3.0556,
      "theoretical_loss": 4.100399885513339,
      "tokens_seen": 330185728
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004545135406218656,
      "loss": 3.03,
      "theoretical_loss": 4.10030589221685,
      "tokens_seen": 330251264
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004545035105315948,
      "loss": 2.9057,
      "theoretical_loss": 4.100211922792271,
      "tokens_seen": 330316800
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045449348044132397,
      "loss": 3.1829,
      "theoretical_loss": 4.100117977228804,
      "tokens_seen": 330382336
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045448345035105315,
      "loss": 2.9077,
      "theoretical_loss": 4.1000240555156555,
      "tokens_seen": 330447872
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004544734202607824,
      "loss": 3.0517,
      "theoretical_loss": 4.099930157642047,
      "tokens_seen": 330513408
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004544633901705115,
      "loss": 2.9146,
      "theoretical_loss": 4.099836283597199,
      "tokens_seen": 330578944
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045445336008024075,
      "loss": 2.989,
      "theoretical_loss": 4.099742433370342,
      "tokens_seen": 330644480
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004544433299899699,
      "loss": 2.8763,
      "theoretical_loss": 4.0996486069507165,
      "tokens_seen": 330710016
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004544332998996991,
      "loss": 2.9498,
      "theoretical_loss": 4.099554804327565,
      "tokens_seen": 330775552
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004544232698094283,
      "loss": 2.8131,
      "theoretical_loss": 4.09946102549014,
      "tokens_seen": 330841088
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004544132397191575,
      "loss": 2.9849,
      "theoretical_loss": 4.099367270427699,
      "tokens_seen": 330906624
    },
    {
      "epoch": 1.01,
      "objective/train/docs_used": 458161,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8100082874298096,
      "objective/train/theoretical_loss": 4.099296969726746,
      "objective/train/tokens_used": 351415776,
      "theoretical_loss": 4.099296969726746,
      "tokens_seen": 330955776
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045440320962888666,
      "loss": 2.9439,
      "theoretical_loss": 4.099273539129509,
      "tokens_seen": 330972160
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045439317953861584,
      "loss": 3.0252,
      "theoretical_loss": 4.099179831584843,
      "tokens_seen": 331037696
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.000454383149448345,
      "loss": 3.0193,
      "theoretical_loss": 4.09908614778298,
      "tokens_seen": 331103232
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045437311935807425,
      "loss": 2.9958,
      "theoretical_loss": 4.098992487713207,
      "tokens_seen": 331168768
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004543630892678034,
      "loss": 3.0734,
      "theoretical_loss": 4.098898851364815,
      "tokens_seen": 331234304
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004543530591775326,
      "loss": 2.9824,
      "theoretical_loss": 4.098805238727109,
      "tokens_seen": 331299840
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004543430290872618,
      "loss": 3.2124,
      "theoretical_loss": 4.098711649789392,
      "tokens_seen": 331365376
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.000454332998996991,
      "loss": 2.8041,
      "theoretical_loss": 4.098618084540981,
      "tokens_seen": 331430912
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045432296890672016,
      "loss": 2.8408,
      "theoretical_loss": 4.098524542971197,
      "tokens_seen": 331496448
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045431293881644934,
      "loss": 3.1551,
      "theoretical_loss": 4.098431025069368,
      "tokens_seen": 331561984
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004543029087261785,
      "loss": 3.0807,
      "theoretical_loss": 4.098337530824828,
      "tokens_seen": 331627520
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045429287863590776,
      "loss": 2.9272,
      "theoretical_loss": 4.09824406022692,
      "tokens_seen": 331693056
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004542828485456369,
      "loss": 3.0841,
      "theoretical_loss": 4.098150613264993,
      "tokens_seen": 331758592
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004542728184553661,
      "loss": 3.2841,
      "theoretical_loss": 4.0980571899284035,
      "tokens_seen": 331824128
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045426278836509525,
      "loss": 2.9232,
      "theoretical_loss": 4.097963790206513,
      "tokens_seen": 331889664
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004542527582748245,
      "loss": 2.9302,
      "theoretical_loss": 4.097870414088691,
      "tokens_seen": 331955200
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045424272818455366,
      "loss": 3.2363,
      "theoretical_loss": 4.097777061564315,
      "tokens_seen": 332020736
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045423269809428284,
      "loss": 3.1058,
      "theoretical_loss": 4.097683732622768,
      "tokens_seen": 332086272
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.000454222668004012,
      "loss": 2.9857,
      "theoretical_loss": 4.09759042725344,
      "tokens_seen": 332151808
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004542126379137412,
      "loss": 3.1798,
      "theoretical_loss": 4.097497145445729,
      "tokens_seen": 332217344
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045420260782347044,
      "loss": 2.9651,
      "theoretical_loss": 4.097403887189038,
      "tokens_seen": 332282880
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004541925777331996,
      "loss": 3.1198,
      "theoretical_loss": 4.097310652472778,
      "tokens_seen": 332348416
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004541825476429288,
      "loss": 2.9496,
      "theoretical_loss": 4.097217441286367,
      "tokens_seen": 332413952
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.000454172517552658,
      "loss": 3.0268,
      "theoretical_loss": 4.09712425361923,
      "tokens_seen": 332479488
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004541624874623872,
      "loss": 2.9316,
      "theoretical_loss": 4.097031089460796,
      "tokens_seen": 332545024
    },
    {
      "epoch": 1.01,
      "objective/train/docs_used": 459492,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.3007261753082275,
      "objective/train/theoretical_loss": 4.096961231763205,
      "objective/train/tokens_used": 353054176,
      "theoretical_loss": 4.096961231763205,
      "tokens_seen": 332594176
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045415245737211635,
      "loss": 3.137,
      "theoretical_loss": 4.096937948800506,
      "tokens_seen": 332610560
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004541424272818456,
      "loss": 3.1584,
      "theoretical_loss": 4.096844831627804,
      "tokens_seen": 332676096
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004541323971915747,
      "loss": 3.3038,
      "theoretical_loss": 4.096751737932141,
      "tokens_seen": 332741632
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045412236710130395,
      "loss": 2.9974,
      "theoretical_loss": 4.096658667702978,
      "tokens_seen": 332807168
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004541123370110331,
      "loss": 2.7968,
      "theoretical_loss": 4.096565620929778,
      "tokens_seen": 332872704
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004541023069207623,
      "loss": 2.9963,
      "theoretical_loss": 4.096472597602014,
      "tokens_seen": 332938240
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004540922768304915,
      "loss": 3.0869,
      "theoretical_loss": 4.096379597709166,
      "tokens_seen": 333003776
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045408224674022067,
      "loss": 3.1367,
      "theoretical_loss": 4.0962866212407185,
      "tokens_seen": 333069312
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045407221664994985,
      "loss": 2.9563,
      "theoretical_loss": 4.096193668186165,
      "tokens_seen": 333134848
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004540621865596791,
      "loss": 2.7802,
      "theoretical_loss": 4.096100738535004,
      "tokens_seen": 333200384
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004540521564694082,
      "loss": 3.09,
      "theoretical_loss": 4.096007832276742,
      "tokens_seen": 333265920
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045404212637913745,
      "loss": 2.9668,
      "theoretical_loss": 4.0959149494008935,
      "tokens_seen": 333331456
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045403209628886663,
      "loss": 3.0414,
      "theoretical_loss": 4.095822089896976,
      "tokens_seen": 333396992
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004540220661985958,
      "loss": 2.8439,
      "theoretical_loss": 4.0957292537545165,
      "tokens_seen": 333462528
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.000454012036108325,
      "loss": 3.1205,
      "theoretical_loss": 4.09563644096305,
      "tokens_seen": 333528064
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045400200601805417,
      "loss": 3.0406,
      "theoretical_loss": 4.095543651512115,
      "tokens_seen": 333593600
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045399197592778335,
      "loss": 3.1992,
      "theoretical_loss": 4.095450885391257,
      "tokens_seen": 333659136
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004539819458375126,
      "loss": 2.9049,
      "theoretical_loss": 4.095358142590031,
      "tokens_seen": 333724672
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004539719157472417,
      "loss": 2.8475,
      "theoretical_loss": 4.095265423097998,
      "tokens_seen": 333790208
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045396188565697095,
      "loss": 3.0034,
      "theoretical_loss": 4.095172726904723,
      "tokens_seen": 333855744
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004539518555667001,
      "loss": 3.0068,
      "theoretical_loss": 4.09508005399978,
      "tokens_seen": 333921280
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004539418254764293,
      "loss": 2.9264,
      "theoretical_loss": 4.094987404372751,
      "tokens_seen": 333986816
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004539317953861585,
      "loss": 2.9861,
      "theoretical_loss": 4.094894778013221,
      "tokens_seen": 334052352
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004539217652958877,
      "loss": 2.9869,
      "theoretical_loss": 4.094802174910784,
      "tokens_seen": 334117888
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045391173520561686,
      "loss": 3.2001,
      "theoretical_loss": 4.094709595055042,
      "tokens_seen": 334183424
    },
    {
      "epoch": 1.01,
      "objective/train/docs_used": 460159,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9135870933532715,
      "objective/train/theoretical_loss": 4.094640175412626,
      "objective/train/tokens_used": 354692576,
      "theoretical_loss": 4.094640175412626,
      "tokens_seen": 334232576
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045390170511534604,
      "loss": 3.0121,
      "theoretical_loss": 4.0946170384356,
      "tokens_seen": 334248960
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004538916750250752,
      "loss": 2.7814,
      "theoretical_loss": 4.094524505042074,
      "tokens_seen": 334314496
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045388164493480445,
      "loss": 2.9696,
      "theoretical_loss": 4.094431994864082,
      "tokens_seen": 334380032
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004538716148445336,
      "loss": 2.9823,
      "theoretical_loss": 4.094339507891252,
      "tokens_seen": 334445568
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004538615847542628,
      "loss": 3.209,
      "theoretical_loss": 4.094247044113219,
      "tokens_seen": 334511104
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.000453851554663992,
      "loss": 3.0497,
      "theoretical_loss": 4.094154603519623,
      "tokens_seen": 334576640
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004538415245737212,
      "loss": 3.0694,
      "theoretical_loss": 4.0940621861001105,
      "tokens_seen": 334642176
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045383149448345036,
      "loss": 2.9195,
      "theoretical_loss": 4.093969791844335,
      "tokens_seen": 334707712
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045382146439317954,
      "loss": 2.8378,
      "theoretical_loss": 4.093877420741958,
      "tokens_seen": 334773248
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004538114343029087,
      "loss": 3.0146,
      "theoretical_loss": 4.093785072782646,
      "tokens_seen": 334838784
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045380140421263796,
      "loss": 3.0526,
      "theoretical_loss": 4.093692747956072,
      "tokens_seen": 334904320
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004537913741223671,
      "loss": 3.1662,
      "theoretical_loss": 4.0936004462519175,
      "tokens_seen": 334969856
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004537813440320963,
      "loss": 2.9986,
      "theoretical_loss": 4.093508167659869,
      "tokens_seen": 335035392
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045377131394182545,
      "loss": 3.1044,
      "theoretical_loss": 4.09341591216962,
      "tokens_seen": 335100928
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004537612838515547,
      "loss": 2.893,
      "theoretical_loss": 4.0933236797708705,
      "tokens_seen": 335166464
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045375125376128386,
      "loss": 2.9955,
      "theoretical_loss": 4.093231470453327,
      "tokens_seen": 335232000
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045374122367101304,
      "loss": 3.1863,
      "theoretical_loss": 4.093139284206703,
      "tokens_seen": 335297536
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004537311935807422,
      "loss": 3.2446,
      "theoretical_loss": 4.09304712102072,
      "tokens_seen": 335363072
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004537211634904714,
      "loss": 3.3968,
      "theoretical_loss": 4.092954980885102,
      "tokens_seen": 335428608
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004537111334002006,
      "loss": 3.0159,
      "theoretical_loss": 4.092862863789584,
      "tokens_seen": 335494144
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004537011033099298,
      "loss": 3.1544,
      "theoretical_loss": 4.092770769723905,
      "tokens_seen": 335559680
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045369107321965895,
      "loss": 3.0962,
      "theoretical_loss": 4.092678698677811,
      "tokens_seen": 335625216
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004536810431293882,
      "loss": 3.0672,
      "theoretical_loss": 4.0925866506410555,
      "tokens_seen": 335690752
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045367101303911737,
      "loss": 3.2044,
      "theoretical_loss": 4.092494625603397,
      "tokens_seen": 335756288
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045366098294884655,
      "loss": 2.981,
      "theoretical_loss": 4.0924026235546025,
      "tokens_seen": 335821824
    },
    {
      "epoch": 1.01,
      "objective/train/docs_used": 460945,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6775364875793457,
      "objective/train/theoretical_loss": 4.092333637098296,
      "objective/train/tokens_used": 356330976,
      "theoretical_loss": 4.092333637098296,
      "tokens_seen": 335870976
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045365095285857573,
      "loss": 2.8384,
      "theoretical_loss": 4.092310644484444,
      "tokens_seen": 335887360
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004536409227683049,
      "loss": 2.7994,
      "theoretical_loss": 4.092218688382701,
      "tokens_seen": 335952896
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004536308926780341,
      "loss": 2.7465,
      "theoretical_loss": 4.092126755239159,
      "tokens_seen": 336018432
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004536208625877633,
      "loss": 3.0426,
      "theoretical_loss": 4.092034845043608,
      "tokens_seen": 336083968
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045361083249749245,
      "loss": 3.0685,
      "theoretical_loss": 4.09194295778585,
      "tokens_seen": 336149504
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004536008024072217,
      "loss": 2.9471,
      "theoretical_loss": 4.091851093455689,
      "tokens_seen": 336215040
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004535907723169508,
      "loss": 2.9211,
      "theoretical_loss": 4.091759252042936,
      "tokens_seen": 336280576
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045358074222668005,
      "loss": 2.8346,
      "theoretical_loss": 4.091667433537408,
      "tokens_seen": 336346112
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045357071213640923,
      "loss": 2.8476,
      "theoretical_loss": 4.0915756379289325,
      "tokens_seen": 336411648
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004535606820461384,
      "loss": 2.9507,
      "theoretical_loss": 4.09148386520734,
      "tokens_seen": 336477184
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004535506519558676,
      "loss": 3.0561,
      "theoretical_loss": 4.091392115362467,
      "tokens_seen": 336542720
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045354062186559683,
      "loss": 2.7804,
      "theoretical_loss": 4.091300388384158,
      "tokens_seen": 336608256
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045353059177532596,
      "loss": 3.1013,
      "theoretical_loss": 4.091208684262265,
      "tokens_seen": 336673792
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004535205616850552,
      "loss": 3.2401,
      "theoretical_loss": 4.0911170029866435,
      "tokens_seen": 336739328
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004535105315947843,
      "loss": 3.0821,
      "theoretical_loss": 4.091025344547158,
      "tokens_seen": 336804864
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045350050150451355,
      "loss": 3.2724,
      "theoretical_loss": 4.090933708933679,
      "tokens_seen": 336870400
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045349047141424274,
      "loss": 3.2233,
      "theoretical_loss": 4.0908420961360825,
      "tokens_seen": 336935936
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004534804413239719,
      "loss": 2.8302,
      "theoretical_loss": 4.090750506144251,
      "tokens_seen": 337001472
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004534704112337011,
      "loss": 2.9115,
      "theoretical_loss": 4.090658938948074,
      "tokens_seen": 337067008
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004534603811434303,
      "loss": 3.08,
      "theoretical_loss": 4.090567394537449,
      "tokens_seen": 337132544
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004534503510531595,
      "loss": 2.9024,
      "theoretical_loss": 4.090475872902277,
      "tokens_seen": 337198080
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004534403209628887,
      "loss": 3.1429,
      "theoretical_loss": 4.0903843740324675,
      "tokens_seen": 337263616
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004534302908726179,
      "loss": 2.8771,
      "theoretical_loss": 4.090292897917935,
      "tokens_seen": 337329152
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045342026078234706,
      "loss": 3.1249,
      "theoretical_loss": 4.090201444548601,
      "tokens_seen": 337394688
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045341023069207624,
      "loss": 3.1044,
      "theoretical_loss": 4.090110013914395,
      "tokens_seen": 337460224
    },
    {
      "epoch": 1.01,
      "objective/train/docs_used": 462430,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.748074531555176,
      "objective/train/theoretical_loss": 4.090041455852612,
      "objective/train/tokens_used": 357969376,
      "theoretical_loss": 4.090041455852612,
      "tokens_seen": 337509376
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004534002006018054,
      "loss": 2.9316,
      "theoretical_loss": 4.09001860600525,
      "tokens_seen": 337525760
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045339017051153465,
      "loss": 2.849,
      "theoretical_loss": 4.0899272208111075,
      "tokens_seen": 337591296
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004533801404212638,
      "loss": 3.1926,
      "theoretical_loss": 4.089835858321916,
      "tokens_seen": 337656832
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.000453370110330993,
      "loss": 3.0211,
      "theoretical_loss": 4.089744518527627,
      "tokens_seen": 337722368
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004533600802407222,
      "loss": 2.9728,
      "theoretical_loss": 4.089653201418201,
      "tokens_seen": 337787904
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004533500501504514,
      "loss": 3.1214,
      "theoretical_loss": 4.089561906983606,
      "tokens_seen": 337853440
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045334002006018056,
      "loss": 3.0502,
      "theoretical_loss": 4.089470635213814,
      "tokens_seen": 337918976
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045332998996990974,
      "loss": 3.1409,
      "theoretical_loss": 4.089379386098804,
      "tokens_seen": 337984512
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004533199598796389,
      "loss": 2.8961,
      "theoretical_loss": 4.089288159628562,
      "tokens_seen": 338050048
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045330992978936816,
      "loss": 2.851,
      "theoretical_loss": 4.08919695579308,
      "tokens_seen": 338115584
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004532998996990973,
      "loss": 2.8503,
      "theoretical_loss": 4.089105774582356,
      "tokens_seen": 338181120
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004532898696088265,
      "loss": 3.0414,
      "theoretical_loss": 4.089014615986394,
      "tokens_seen": 338246656
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045327983951855565,
      "loss": 2.8973,
      "theoretical_loss": 4.0889234799952066,
      "tokens_seen": 338312192
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004532698094282849,
      "loss": 2.8654,
      "theoretical_loss": 4.088832366598811,
      "tokens_seen": 338377728
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045325977933801406,
      "loss": 2.9145,
      "theoretical_loss": 4.08874127578723,
      "tokens_seen": 338443264
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045324974924774324,
      "loss": 2.9358,
      "theoretical_loss": 4.0886502075504945,
      "tokens_seen": 338508800
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004532397191574724,
      "loss": 3.1151,
      "theoretical_loss": 4.08855916187864,
      "tokens_seen": 338574336
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004532296890672016,
      "loss": 3.2944,
      "theoretical_loss": 4.08846813876171,
      "tokens_seen": 338639872
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004532196589769308,
      "loss": 2.8268,
      "theoretical_loss": 4.088377138189754,
      "tokens_seen": 338705408
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045320962888666,
      "loss": 3.0523,
      "theoretical_loss": 4.088286160152827,
      "tokens_seen": 338770944
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045319959879638915,
      "loss": 2.8429,
      "theoretical_loss": 4.088195204640989,
      "tokens_seen": 338836480
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004531895687061184,
      "loss": 2.8917,
      "theoretical_loss": 4.088104271644311,
      "tokens_seen": 338902016
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045317953861584757,
      "loss": 3.2186,
      "theoretical_loss": 4.088013361152865,
      "tokens_seen": 338967552
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045316950852557675,
      "loss": 2.8986,
      "theoretical_loss": 4.087922473156732,
      "tokens_seen": 339033088
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045315947843530593,
      "loss": 3.0517,
      "theoretical_loss": 4.087831607646,
      "tokens_seen": 339098624
    },
    {
      "epoch": 1.01,
      "objective/train/docs_used": 463251,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.839505434036255,
      "objective/train/theoretical_loss": 4.087763473263035,
      "objective/train/tokens_used": 359607776,
      "theoretical_loss": 4.087763473263035,
      "tokens_seen": 339147776
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004531494483450351,
      "loss": 2.941,
      "theoretical_loss": 4.087740764610761,
      "tokens_seen": 339164160
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004531394182547643,
      "loss": 2.9092,
      "theoretical_loss": 4.0876499440411145,
      "tokens_seen": 339229696
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004531293881644935,
      "loss": 2.8542,
      "theoretical_loss": 4.087559145927166,
      "tokens_seen": 339295232
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045311935807422265,
      "loss": 3.1179,
      "theoretical_loss": 4.087468370259028,
      "tokens_seen": 339360768
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004531093279839519,
      "loss": 2.9362,
      "theoretical_loss": 4.087377617026819,
      "tokens_seen": 339426304
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.000453099297893681,
      "loss": 2.8019,
      "theoretical_loss": 4.087286886220663,
      "tokens_seen": 339491840
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045308926780341025,
      "loss": 3.1268,
      "theoretical_loss": 4.087196177830691,
      "tokens_seen": 339557376
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045307923771313943,
      "loss": 2.732,
      "theoretical_loss": 4.08710549184704,
      "tokens_seen": 339622912
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004530692076228686,
      "loss": 3.1866,
      "theoretical_loss": 4.087014828259853,
      "tokens_seen": 339688448
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004530591775325978,
      "loss": 2.9075,
      "theoretical_loss": 4.086924187059279,
      "tokens_seen": 339753984
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045304914744232703,
      "loss": 3.0424,
      "theoretical_loss": 4.086833568235474,
      "tokens_seen": 339819520
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045303911735205616,
      "loss": 3.0533,
      "theoretical_loss": 4.086742971778601,
      "tokens_seen": 339885056
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004530290872617854,
      "loss": 3.0583,
      "theoretical_loss": 4.086652397678827,
      "tokens_seen": 339950592
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004530190571715145,
      "loss": 2.9185,
      "theoretical_loss": 4.086561845926326,
      "tokens_seen": 340016128
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045300902708124375,
      "loss": 2.9725,
      "theoretical_loss": 4.086471316511281,
      "tokens_seen": 340081664
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045299899699097294,
      "loss": 2.917,
      "theoretical_loss": 4.086380809423876,
      "tokens_seen": 340147200
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004529889669007021,
      "loss": 3.1826,
      "theoretical_loss": 4.086290324654303,
      "tokens_seen": 340212736
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004529789368104313,
      "loss": 2.8471,
      "theoretical_loss": 4.086199862192766,
      "tokens_seen": 340278272
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004529689067201605,
      "loss": 3.1347,
      "theoretical_loss": 4.086109422029466,
      "tokens_seen": 340343808
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045295887662988966,
      "loss": 2.7183,
      "theoretical_loss": 4.0860190041546165,
      "tokens_seen": 340409344
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004529488465396189,
      "loss": 2.9848,
      "theoretical_loss": 4.085928608558435,
      "tokens_seen": 340474880
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.000452938816449348,
      "loss": 3.2565,
      "theoretical_loss": 4.085838235231145,
      "tokens_seen": 340540416
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045292878635907726,
      "loss": 3.0739,
      "theoretical_loss": 4.085747884162976,
      "tokens_seen": 340605952
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004529187562688064,
      "loss": 2.9197,
      "theoretical_loss": 4.0856575553441665,
      "tokens_seen": 340671488
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004529087261785356,
      "loss": 3.154,
      "theoretical_loss": 4.085567248764956,
      "tokens_seen": 340737024
    },
    {
      "epoch": 1.01,
      "objective/train/docs_used": 464753,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.337740182876587,
      "objective/train/theoretical_loss": 4.08549953341942,
      "objective/train/tokens_used": 361246176,
      "theoretical_loss": 4.08549953341942,
      "tokens_seen": 340786176
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004528986960882648,
      "loss": 3.2774,
      "theoretical_loss": 4.085476964415595,
      "tokens_seen": 340802560
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.000452888665997994,
      "loss": 3.0845,
      "theoretical_loss": 4.085386702286338,
      "tokens_seen": 340868096
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045287863590772316,
      "loss": 2.9111,
      "theoretical_loss": 4.085296462367445,
      "tokens_seen": 340933632
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004528686058174524,
      "loss": 3.3461,
      "theoretical_loss": 4.085206244649184,
      "tokens_seen": 340999168
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004528585757271815,
      "loss": 2.8499,
      "theoretical_loss": 4.085116049121828,
      "tokens_seen": 341064704
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045284854563691076,
      "loss": 3.4213,
      "theoretical_loss": 4.085025875775655,
      "tokens_seen": 341130240
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004528385155466399,
      "loss": 2.9534,
      "theoretical_loss": 4.0849357246009514,
      "tokens_seen": 341195776
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004528284854563691,
      "loss": 3.0351,
      "theoretical_loss": 4.084845595588009,
      "tokens_seen": 341261312
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004528184553660983,
      "loss": 3.1023,
      "theoretical_loss": 4.084755488727124,
      "tokens_seen": 341326848
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004528084252758275,
      "loss": 2.9196,
      "theoretical_loss": 4.084665404008602,
      "tokens_seen": 341392384
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045279839518555667,
      "loss": 2.8665,
      "theoretical_loss": 4.084575341422752,
      "tokens_seen": 341457920
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045278836509528585,
      "loss": 3.187,
      "theoretical_loss": 4.084485300959889,
      "tokens_seen": 341523456
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045277833500501503,
      "loss": 3.1001,
      "theoretical_loss": 4.084395282610337,
      "tokens_seen": 341588992
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045276830491474426,
      "loss": 3.0675,
      "theoretical_loss": 4.0843052863644225,
      "tokens_seen": 341654528
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004527582748244734,
      "loss": 3.1473,
      "theoretical_loss": 4.08421531221248,
      "tokens_seen": 341720064
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004527482447342026,
      "loss": 3.0618,
      "theoretical_loss": 4.084125360144849,
      "tokens_seen": 341785600
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045273821464393175,
      "loss": 3.083,
      "theoretical_loss": 4.084035430151879,
      "tokens_seen": 341851136
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.000452728184553661,
      "loss": 2.9518,
      "theoretical_loss": 4.083945522223919,
      "tokens_seen": 341916672
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045271815446339017,
      "loss": 3.136,
      "theoretical_loss": 4.083855636351329,
      "tokens_seen": 341982208
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045270812437311935,
      "loss": 2.9707,
      "theoretical_loss": 4.0837657725244725,
      "tokens_seen": 342047744
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004526980942828486,
      "loss": 2.9953,
      "theoretical_loss": 4.083675930733721,
      "tokens_seen": 342113280
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045268806419257777,
      "loss": 2.6718,
      "theoretical_loss": 4.083586110969451,
      "tokens_seen": 342178816
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045267803410230695,
      "loss": 3.0025,
      "theoretical_loss": 4.083496313222046,
      "tokens_seen": 342244352
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045266800401203613,
      "loss": 3.0224,
      "theoretical_loss": 4.083406537481893,
      "tokens_seen": 342309888
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004526579739217653,
      "loss": 2.7411,
      "theoretical_loss": 4.083316783739388,
      "tokens_seen": 342375424
    },
    {
      "epoch": 1.01,
      "objective/train/docs_used": 465448,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.2629733085632324,
      "objective/train/theoretical_loss": 4.083249482862691,
      "objective/train/tokens_used": 362884576,
      "theoretical_loss": 4.083249482862691,
      "tokens_seen": 342424576
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004526479438314945,
      "loss": 2.895,
      "theoretical_loss": 4.083227051984932,
      "tokens_seen": 342440960
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004526379137412237,
      "loss": 2.8273,
      "theoretical_loss": 4.08313734220893,
      "tokens_seen": 342506496
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045262788365095285,
      "loss": 2.9571,
      "theoretical_loss": 4.083047654401797,
      "tokens_seen": 342572032
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004526178535606821,
      "loss": 2.6863,
      "theoretical_loss": 4.082957988553951,
      "tokens_seen": 342637568
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004526078234704112,
      "loss": 2.9279,
      "theoretical_loss": 4.082868344655816,
      "tokens_seen": 342703104
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045259779338014045,
      "loss": 3.321,
      "theoretical_loss": 4.082778722697825,
      "tokens_seen": 342768640
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045258776328986963,
      "loss": 3.1226,
      "theoretical_loss": 4.082689122670413,
      "tokens_seen": 342834176
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004525777331995988,
      "loss": 3.1412,
      "theoretical_loss": 4.082599544564024,
      "tokens_seen": 342899712
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.000452567703109328,
      "loss": 2.9624,
      "theoretical_loss": 4.082509988369106,
      "tokens_seen": 342965248
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045255767301905723,
      "loss": 3.0756,
      "theoretical_loss": 4.0824204540761135,
      "tokens_seen": 343030784
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045254764292878636,
      "loss": 2.9775,
      "theoretical_loss": 4.082330941675508,
      "tokens_seen": 343096320
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004525376128385156,
      "loss": 2.9741,
      "theoretical_loss": 4.082241451157757,
      "tokens_seen": 343161856
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004525275827482447,
      "loss": 3.0864,
      "theoretical_loss": 4.0821519825133326,
      "tokens_seen": 343227392
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045251755265797395,
      "loss": 3.3031,
      "theoretical_loss": 4.082062535732713,
      "tokens_seen": 343292928
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045250752256770314,
      "loss": 2.8638,
      "theoretical_loss": 4.081973110806383,
      "tokens_seen": 343358464
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004524974924774323,
      "loss": 3.0286,
      "theoretical_loss": 4.081883707724835,
      "tokens_seen": 343424000
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004524874623871615,
      "loss": 2.8336,
      "theoretical_loss": 4.081794326478563,
      "tokens_seen": 343489536
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004524774322968907,
      "loss": 3.2582,
      "theoretical_loss": 4.081704967058071,
      "tokens_seen": 343555072
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045246740220661986,
      "loss": 3.1308,
      "theoretical_loss": 4.081615629453868,
      "tokens_seen": 343620608
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004524573721163491,
      "loss": 3.0641,
      "theoretical_loss": 4.081526313656466,
      "tokens_seen": 343686144
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004524473420260782,
      "loss": 3.0581,
      "theoretical_loss": 4.081437019656389,
      "tokens_seen": 343751680
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045243731193580746,
      "loss": 3.0062,
      "theoretical_loss": 4.081347747444161,
      "tokens_seen": 343817216
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004524272818455366,
      "loss": 2.7659,
      "theoretical_loss": 4.081258497010314,
      "tokens_seen": 343882752
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004524172517552658,
      "loss": 3.0713,
      "theoretical_loss": 4.081169268345387,
      "tokens_seen": 343948288
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.000452407221664995,
      "loss": 3.0026,
      "theoretical_loss": 4.081080061439923,
      "tokens_seen": 344013824
    },
    {
      "epoch": 1.01,
      "objective/train/docs_used": 466745,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.469646692276001,
      "objective/train/theoretical_loss": 4.0810131705347885,
      "objective/train/tokens_used": 364522976,
      "theoretical_loss": 4.0810131705347885,
      "tokens_seen": 344062976
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004523971915747242,
      "loss": 3.0142,
      "theoretical_loss": 4.080990876284473,
      "tokens_seen": 344079360
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045238716148445336,
      "loss": 3.1162,
      "theoretical_loss": 4.0809017128695935,
      "tokens_seen": 344144896
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004523771313941826,
      "loss": 3.3286,
      "theoretical_loss": 4.080812571185845,
      "tokens_seen": 344210432
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004523671013039117,
      "loss": 2.7948,
      "theoretical_loss": 4.080723451223795,
      "tokens_seen": 344275968
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045235707121364096,
      "loss": 3.0005,
      "theoretical_loss": 4.080634352974018,
      "tokens_seen": 344341504
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004523470411233701,
      "loss": 3.055,
      "theoretical_loss": 4.080545276427092,
      "tokens_seen": 344407040
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004523370110330993,
      "loss": 3.0788,
      "theoretical_loss": 4.080456221573604,
      "tokens_seen": 344472576
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004523269809428285,
      "loss": 2.9626,
      "theoretical_loss": 4.080367188404144,
      "tokens_seen": 344538112
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004523169508525577,
      "loss": 2.8527,
      "theoretical_loss": 4.080278176909309,
      "tokens_seen": 344603648
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045230692076228687,
      "loss": 3.0898,
      "theoretical_loss": 4.080189187079703,
      "tokens_seen": 344669184
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045229689067201605,
      "loss": 3.1361,
      "theoretical_loss": 4.080100218905933,
      "tokens_seen": 344734720
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045228686058174523,
      "loss": 3.1205,
      "theoretical_loss": 4.080011272378616,
      "tokens_seen": 344800256
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045227683049147446,
      "loss": 3.0475,
      "theoretical_loss": 4.07992234748837,
      "tokens_seen": 344865792
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004522668004012036,
      "loss": 2.9471,
      "theoretical_loss": 4.079833444225824,
      "tokens_seen": 344931328
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004522567703109328,
      "loss": 3.1468,
      "theoretical_loss": 4.079744562581608,
      "tokens_seen": 344996864
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045224674022066195,
      "loss": 3.1542,
      "theoretical_loss": 4.0796557025463605,
      "tokens_seen": 345062400
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004522367101303912,
      "loss": 2.932,
      "theoretical_loss": 4.079566864110725,
      "tokens_seen": 345127936
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045222668004012037,
      "loss": 2.7915,
      "theoretical_loss": 4.0794780472653525,
      "tokens_seen": 345193472
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045221664994984955,
      "loss": 3.0188,
      "theoretical_loss": 4.0793892520008965,
      "tokens_seen": 345259008
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045220661985957873,
      "loss": 3.2555,
      "theoretical_loss": 4.07930047830802,
      "tokens_seen": 345324544
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045219658976930797,
      "loss": 2.9997,
      "theoretical_loss": 4.079211726177389,
      "tokens_seen": 345390080
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004521865596790371,
      "loss": 2.8722,
      "theoretical_loss": 4.079122995599677,
      "tokens_seen": 345455616
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045217652958876633,
      "loss": 2.816,
      "theoretical_loss": 4.079034286565563,
      "tokens_seen": 345521152
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045216649949849546,
      "loss": 3.0846,
      "theoretical_loss": 4.078945599065731,
      "tokens_seen": 345586688
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004521564694082247,
      "loss": 3.057,
      "theoretical_loss": 4.078856933090871,
      "tokens_seen": 345652224
    },
    {
      "epoch": 1.01,
      "objective/train/docs_used": 467437,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.132908344268799,
      "objective/train/theoretical_loss": 4.078790447729892,
      "objective/train/tokens_used": 366161376,
      "theoretical_loss": 4.078790447729892,
      "tokens_seen": 345701376
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045214643931795387,
      "loss": 3.0499,
      "theoretical_loss": 4.07876828863168,
      "tokens_seen": 345717760
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045213640922768305,
      "loss": 2.9937,
      "theoretical_loss": 4.078679665678859,
      "tokens_seen": 345783296
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045212637913741223,
      "loss": 3.0931,
      "theoretical_loss": 4.078591064223116,
      "tokens_seen": 345848832
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004521163490471414,
      "loss": 2.9227,
      "theoretical_loss": 4.078502484255164,
      "tokens_seen": 345914368
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004521063189568706,
      "loss": 2.871,
      "theoretical_loss": 4.078413925765724,
      "tokens_seen": 345979904
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045209628886659983,
      "loss": 3.1182,
      "theoretical_loss": 4.078325388745519,
      "tokens_seen": 346045440
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.00045208625877632896,
      "loss": 3.0976,
      "theoretical_loss": 4.07823687318528,
      "tokens_seen": 346110976
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004520762286860582,
      "loss": 2.9415,
      "theoretical_loss": 4.078148379075744,
      "tokens_seen": 346176512
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0004520661985957873,
      "loss": 2.9781,
      "theoretical_loss": 4.078059906407653,
      "tokens_seen": 346242048
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00045205616850551656,
      "loss": 2.825,
      "theoretical_loss": 4.077971455171755,
      "tokens_seen": 346307584
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00045204613841524574,
      "loss": 2.8616,
      "theoretical_loss": 4.077883025358804,
      "tokens_seen": 346373120
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004520361083249749,
      "loss": 2.8784,
      "theoretical_loss": 4.07779461695956,
      "tokens_seen": 346438656
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004520260782347041,
      "loss": 2.9527,
      "theoretical_loss": 4.077706229964786,
      "tokens_seen": 346504192
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00045201604814443334,
      "loss": 3.3518,
      "theoretical_loss": 4.077617864365255,
      "tokens_seen": 346569728
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00045200601805416246,
      "loss": 2.7224,
      "theoretical_loss": 4.077529520151743,
      "tokens_seen": 346635264
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004519959879638917,
      "loss": 3.0247,
      "theoretical_loss": 4.077441197315032,
      "tokens_seen": 346700800
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004519859578736208,
      "loss": 3.023,
      "theoretical_loss": 4.0773528958459115,
      "tokens_seen": 346766336
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00045197592778335006,
      "loss": 2.9417,
      "theoretical_loss": 4.0772646157351735,
      "tokens_seen": 346831872
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00045196589769307924,
      "loss": 2.6767,
      "theoretical_loss": 4.077176356973618,
      "tokens_seen": 346897408
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004519558676028084,
      "loss": 3.0087,
      "theoretical_loss": 4.077088119552052,
      "tokens_seen": 346962944
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00045194583751253766,
      "loss": 3.1533,
      "theoretical_loss": 4.076999903461283,
      "tokens_seen": 347028480
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004519358074222668,
      "loss": 2.7099,
      "theoretical_loss": 4.07691170869213,
      "tokens_seen": 347094016
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.000451925777331996,
      "loss": 2.8064,
      "theoretical_loss": 4.076823535235415,
      "tokens_seen": 347159552
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004519157472417252,
      "loss": 3.2354,
      "theoretical_loss": 4.0767353830819655,
      "tokens_seen": 347225088
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004519057171514544,
      "loss": 2.6779,
      "theoretical_loss": 4.076647252222616,
      "tokens_seen": 347290624
    },
    {
      "epoch": 1.02,
      "objective/train/docs_used": 469057,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.269904375076294,
      "objective/train/theoretical_loss": 4.076581168046844,
      "objective/train/tokens_used": 367799776,
      "theoretical_loss": 4.076581168046844,
      "tokens_seen": 347339776
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00045189568706118356,
      "loss": 3.1138,
      "theoretical_loss": 4.076559142648204,
      "tokens_seen": 347356160
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004518856569709128,
      "loss": 2.8085,
      "theoretical_loss": 4.076471054349575,
      "tokens_seen": 347421696
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004518756268806419,
      "loss": 3.0591,
      "theoretical_loss": 4.076382987317581,
      "tokens_seen": 347487232
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00045186559679037116,
      "loss": 3.0139,
      "theoretical_loss": 4.076294941543078,
      "tokens_seen": 347552768
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004518555667001003,
      "loss": 3.0779,
      "theoretical_loss": 4.076206917016927,
      "tokens_seen": 347618304
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004518455366098295,
      "loss": 2.8424,
      "theoretical_loss": 4.076118913729996,
      "tokens_seen": 347683840
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004518355065195587,
      "loss": 2.9054,
      "theoretical_loss": 4.0760309316731576,
      "tokens_seen": 347749376
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004518254764292879,
      "loss": 3.2079,
      "theoretical_loss": 4.075942970837292,
      "tokens_seen": 347814912
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00045181544633901707,
      "loss": 2.9092,
      "theoretical_loss": 4.075855031213283,
      "tokens_seen": 347880448
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00045180541624874625,
      "loss": 3.1271,
      "theoretical_loss": 4.075767112792021,
      "tokens_seen": 347945984
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00045179538615847543,
      "loss": 2.9086,
      "theoretical_loss": 4.075679215564401,
      "tokens_seen": 348011520
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00045178535606820466,
      "loss": 2.8916,
      "theoretical_loss": 4.075591339521326,
      "tokens_seen": 348077056
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004517753259779338,
      "loss": 3.1885,
      "theoretical_loss": 4.0755034846537015,
      "tokens_seen": 348142592
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.000451765295887663,
      "loss": 2.7859,
      "theoretical_loss": 4.07541565095244,
      "tokens_seen": 348208128
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00045175526579739215,
      "loss": 3.1342,
      "theoretical_loss": 4.0753278384084615,
      "tokens_seen": 348273664
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004517452357071214,
      "loss": 3.0243,
      "theoretical_loss": 4.075240047012688,
      "tokens_seen": 348339200
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00045173520561685057,
      "loss": 2.9966,
      "theoretical_loss": 4.07515227675605,
      "tokens_seen": 348404736
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00045172517552657975,
      "loss": 3.2115,
      "theoretical_loss": 4.075064527629483,
      "tokens_seen": 348470272
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00045171514543630893,
      "loss": 3.2221,
      "theoretical_loss": 4.074976799623926,
      "tokens_seen": 348535808
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00045170511534603817,
      "loss": 2.8883,
      "theoretical_loss": 4.074889092730325,
      "tokens_seen": 348601344
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004516950852557673,
      "loss": 3.1443,
      "theoretical_loss": 4.074801406939635,
      "tokens_seen": 348666880
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00045168505516549653,
      "loss": 3.1144,
      "theoretical_loss": 4.07471374224281,
      "tokens_seen": 348732416
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00045167502507522566,
      "loss": 2.9944,
      "theoretical_loss": 4.074626098630815,
      "tokens_seen": 348797952
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004516649949849549,
      "loss": 2.8234,
      "theoretical_loss": 4.074538476094617,
      "tokens_seen": 348863488
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00045165496489468407,
      "loss": 2.834,
      "theoretical_loss": 4.0744508746251915,
      "tokens_seen": 348929024
    },
    {
      "epoch": 1.02,
      "objective/train/docs_used": 469778,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.0074195861816406,
      "objective/train/theoretical_loss": 4.074385187342765,
      "objective/train/tokens_used": 369438176,
      "theoretical_loss": 4.074385187342765,
      "tokens_seen": 348978176
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00045164493480441325,
      "loss": 3.1902,
      "theoretical_loss": 4.0743632942135175,
      "tokens_seen": 348994560
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00045163490471414244,
      "loss": 3.0386,
      "theoretical_loss": 4.07427573485058,
      "tokens_seen": 349060096
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004516248746238716,
      "loss": 2.8443,
      "theoretical_loss": 4.07418819652737,
      "tokens_seen": 349125632
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004516148445336008,
      "loss": 2.8372,
      "theoretical_loss": 4.074100679234883,
      "tokens_seen": 349191168
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00045160481444333003,
      "loss": 3.1183,
      "theoretical_loss": 4.074013182964123,
      "tokens_seen": 349256704
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00045159478435305916,
      "loss": 3.104,
      "theoretical_loss": 4.0739257077060955,
      "tokens_seen": 349322240
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004515847542627884,
      "loss": 3.0652,
      "theoretical_loss": 4.073838253451814,
      "tokens_seen": 349387776
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004515747241725175,
      "loss": 3.0714,
      "theoretical_loss": 4.073750820192296,
      "tokens_seen": 349453312
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00045156469408224676,
      "loss": 2.9575,
      "theoretical_loss": 4.073663407918566,
      "tokens_seen": 349518848
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00045155466399197594,
      "loss": 3.0506,
      "theoretical_loss": 4.073576016621656,
      "tokens_seen": 349584384
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004515446339017051,
      "loss": 2.8519,
      "theoretical_loss": 4.073488646292597,
      "tokens_seen": 349649920
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004515346038114343,
      "loss": 2.8341,
      "theoretical_loss": 4.073401296922432,
      "tokens_seen": 349715456
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00045152457372116354,
      "loss": 3.1848,
      "theoretical_loss": 4.073313968502206,
      "tokens_seen": 349780992
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00045151454363089266,
      "loss": 2.9554,
      "theoretical_loss": 4.0732266610229715,
      "tokens_seen": 349846528
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004515045135406219,
      "loss": 2.7127,
      "theoretical_loss": 4.073139374475784,
      "tokens_seen": 349912064
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.000451494483450351,
      "loss": 3.185,
      "theoretical_loss": 4.073052108851707,
      "tokens_seen": 349977600
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00045148445336008026,
      "loss": 3.3413,
      "theoretical_loss": 4.072964864141809,
      "tokens_seen": 350043136
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00045147442326980944,
      "loss": 3.0749,
      "theoretical_loss": 4.072877640337162,
      "tokens_seen": 350108672
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004514643931795386,
      "loss": 3.016,
      "theoretical_loss": 4.072790437428846,
      "tokens_seen": 350174208
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004514543630892678,
      "loss": 3.1049,
      "theoretical_loss": 4.072703255407946,
      "tokens_seen": 350239744
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.000451444332998997,
      "loss": 3.0196,
      "theoretical_loss": 4.07261609426555,
      "tokens_seen": 350305280
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00045143430290872617,
      "loss": 2.9661,
      "theoretical_loss": 4.072528953992756,
      "tokens_seen": 350370816
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004514242728184554,
      "loss": 2.7829,
      "theoretical_loss": 4.072441834580663,
      "tokens_seen": 350436352
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00045141424272818453,
      "loss": 3.1238,
      "theoretical_loss": 4.072354736020377,
      "tokens_seen": 350501888
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00045140421263791376,
      "loss": 3.0012,
      "theoretical_loss": 4.07226765830301,
      "tokens_seen": 350567424
    },
    {
      "epoch": 1.02,
      "objective/train/docs_used": 471047,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.1343419551849365,
      "objective/train/theoretical_loss": 4.072202363687808,
      "objective/train/tokens_used": 371076576,
      "theoretical_loss": 4.072202363687808,
      "tokens_seen": 350616576
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00045139418254764294,
      "loss": 3.0134,
      "theoretical_loss": 4.072180601419681,
      "tokens_seen": 350632960
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004513841524573721,
      "loss": 2.9707,
      "theoretical_loss": 4.072093565361511,
      "tokens_seen": 350698496
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004513741223671013,
      "loss": 2.9995,
      "theoretical_loss": 4.072006550119628,
      "tokens_seen": 350764032
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004513640922768305,
      "loss": 2.9866,
      "theoretical_loss": 4.071919555685166,
      "tokens_seen": 350829568
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00045135406218655967,
      "loss": 2.9988,
      "theoretical_loss": 4.071832582049264,
      "tokens_seen": 350895104
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004513440320962889,
      "loss": 2.6705,
      "theoretical_loss": 4.071745629203066,
      "tokens_seen": 350960640
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00045133400200601803,
      "loss": 3.159,
      "theoretical_loss": 4.071658697137722,
      "tokens_seen": 351026176
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00045132397191574727,
      "loss": 2.6714,
      "theoretical_loss": 4.071571785844387,
      "tokens_seen": 351091712
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004513139418254764,
      "loss": 3.0984,
      "theoretical_loss": 4.071484895314223,
      "tokens_seen": 351157248
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00045130391173520563,
      "loss": 3.1536,
      "theoretical_loss": 4.071398025538394,
      "tokens_seen": 351222784
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004512938816449348,
      "loss": 2.8315,
      "theoretical_loss": 4.071311176508073,
      "tokens_seen": 351288320
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.000451283851554664,
      "loss": 3.047,
      "theoretical_loss": 4.071224348214435,
      "tokens_seen": 351353856
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00045127382146439317,
      "loss": 3.2294,
      "theoretical_loss": 4.071137540648665,
      "tokens_seen": 351419392
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00045126379137412235,
      "loss": 3.1709,
      "theoretical_loss": 4.0710507538019485,
      "tokens_seen": 351484928
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00045125376128385153,
      "loss": 3.2218,
      "theoretical_loss": 4.070963987665479,
      "tokens_seen": 351550464
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00045124373119358077,
      "loss": 2.9715,
      "theoretical_loss": 4.0708772422304556,
      "tokens_seen": 351616000
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004512337011033099,
      "loss": 3.0497,
      "theoretical_loss": 4.070790517488081,
      "tokens_seen": 351681536
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00045122367101303913,
      "loss": 2.9379,
      "theoretical_loss": 4.070703813429566,
      "tokens_seen": 351747072
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004512136409227683,
      "loss": 2.9688,
      "theoretical_loss": 4.070617130046124,
      "tokens_seen": 351812608
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004512036108324975,
      "loss": 3.1147,
      "theoretical_loss": 4.070530467328975,
      "tokens_seen": 351878144
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00045119358074222673,
      "loss": 2.9876,
      "theoretical_loss": 4.070443825269344,
      "tokens_seen": 351943680
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00045118355065195586,
      "loss": 3.2065,
      "theoretical_loss": 4.070357203858462,
      "tokens_seen": 352009216
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004511735205616851,
      "loss": 3.0601,
      "theoretical_loss": 4.070270603087565,
      "tokens_seen": 352074752
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004511634904714143,
      "loss": 3.2063,
      "theoretical_loss": 4.0701840229478945,
      "tokens_seen": 352140288
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00045115346038114345,
      "loss": 2.9421,
      "theoretical_loss": 4.070097463430697,
      "tokens_seen": 352205824
    },
    {
      "epoch": 1.02,
      "objective/train/docs_used": 471658,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.1030523777008057,
      "objective/train/theoretical_loss": 4.070032557321034,
      "objective/train/tokens_used": 372714976,
      "theoretical_loss": 4.070032557321034,
      "tokens_seen": 352254976
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00045114343029087264,
      "loss": 2.8468,
      "theoretical_loss": 4.0700109245272245,
      "tokens_seen": 352271360
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004511334002006018,
      "loss": 3.0807,
      "theoretical_loss": 4.0699244062287345,
      "tokens_seen": 352336896
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.000451123370110331,
      "loss": 2.9094,
      "theoretical_loss": 4.069837908526489,
      "tokens_seen": 352402432
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00045111334002006023,
      "loss": 3.0076,
      "theoretical_loss": 4.069751431411758,
      "tokens_seen": 352467968
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00045110330992978936,
      "loss": 3.1799,
      "theoretical_loss": 4.0696649748758125,
      "tokens_seen": 352533504
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004510932798395186,
      "loss": 3.1436,
      "theoretical_loss": 4.069578538909933,
      "tokens_seen": 352599040
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004510832497492477,
      "loss": 2.9335,
      "theoretical_loss": 4.069492123505402,
      "tokens_seen": 352664576
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00045107321965897696,
      "loss": 3.0457,
      "theoretical_loss": 4.069405728653509,
      "tokens_seen": 352730112
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00045106318956870614,
      "loss": 3.0376,
      "theoretical_loss": 4.0693193543455495,
      "tokens_seen": 352795648
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004510531594784353,
      "loss": 3.1604,
      "theoretical_loss": 4.069233000572823,
      "tokens_seen": 352861184
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004510431293881645,
      "loss": 3.2364,
      "theoretical_loss": 4.069146667326635,
      "tokens_seen": 352926720
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00045103309929789374,
      "loss": 2.9682,
      "theoretical_loss": 4.069060354598296,
      "tokens_seen": 352992256
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00045102306920762286,
      "loss": 3.0206,
      "theoretical_loss": 4.06897406237912,
      "tokens_seen": 353057792
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004510130391173521,
      "loss": 2.8432,
      "theoretical_loss": 4.0688877906604315,
      "tokens_seen": 353123328
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004510030090270812,
      "loss": 2.9897,
      "theoretical_loss": 4.068801539433554,
      "tokens_seen": 353188864
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00045099297893681046,
      "loss": 3.0886,
      "theoretical_loss": 4.06871530868982,
      "tokens_seen": 353254400
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00045098294884653964,
      "loss": 3.1712,
      "theoretical_loss": 4.068629098420567,
      "tokens_seen": 353319936
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004509729187562688,
      "loss": 2.9892,
      "theoretical_loss": 4.068542908617136,
      "tokens_seen": 353385472
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.000450962888665998,
      "loss": 3.1108,
      "theoretical_loss": 4.068456739270876,
      "tokens_seen": 353451008
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004509528585757272,
      "loss": 3.0438,
      "theoretical_loss": 4.068370590373139,
      "tokens_seen": 353516544
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00045094282848545637,
      "loss": 2.8643,
      "theoretical_loss": 4.068284461915282,
      "tokens_seen": 353582080
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004509327983951856,
      "loss": 2.9281,
      "theoretical_loss": 4.06819835388867,
      "tokens_seen": 353647616
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00045092276830491473,
      "loss": 3.2311,
      "theoretical_loss": 4.06811226628467,
      "tokens_seen": 353713152
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00045091273821464396,
      "loss": 3.1787,
      "theoretical_loss": 4.068026199094657,
      "tokens_seen": 353778688
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00045090270812437314,
      "loss": 3.188,
      "theoretical_loss": 4.067940152310008,
      "tokens_seen": 353844224
    },
    {
      "epoch": 1.02,
      "objective/train/docs_used": 473221,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.438317060470581,
      "objective/train/theoretical_loss": 4.06787563060736,
      "objective/train/tokens_used": 374353376,
      "theoretical_loss": 4.06787563060736,
      "tokens_seen": 353893376
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004508926780341023,
      "loss": 3.0398,
      "theoretical_loss": 4.06785412592211,
      "tokens_seen": 353909760
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004508826479438315,
      "loss": 3.107,
      "theoretical_loss": 4.06776811992235,
      "tokens_seen": 353975296
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004508726178535607,
      "loss": 3.1091,
      "theoretical_loss": 4.067682134302124,
      "tokens_seen": 354040832
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00045086258776328987,
      "loss": 3.1053,
      "theoretical_loss": 4.06759616905283,
      "tokens_seen": 354106368
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004508525576730191,
      "loss": 3.3487,
      "theoretical_loss": 4.0675102241658765,
      "tokens_seen": 354171904
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00045084252758274823,
      "loss": 3.0064,
      "theoretical_loss": 4.0674242996326715,
      "tokens_seen": 354237440
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00045083249749247747,
      "loss": 2.8039,
      "theoretical_loss": 4.06733839544463,
      "tokens_seen": 354302976
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004508224674022066,
      "loss": 3.0391,
      "theoretical_loss": 4.067252511593175,
      "tokens_seen": 354368512
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00045081243731193583,
      "loss": 2.8456,
      "theoretical_loss": 4.067166648069731,
      "tokens_seen": 354434048
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.000450802407221665,
      "loss": 2.9981,
      "theoretical_loss": 4.067080804865728,
      "tokens_seen": 354499584
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004507923771313942,
      "loss": 2.7936,
      "theoretical_loss": 4.066994981972604,
      "tokens_seen": 354565120
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00045078234704112337,
      "loss": 3.2045,
      "theoretical_loss": 4.066909179381801,
      "tokens_seen": 354630656
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00045077231695085255,
      "loss": 2.9505,
      "theoretical_loss": 4.066823397084764,
      "tokens_seen": 354696192
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00045076228686058173,
      "loss": 2.6534,
      "theoretical_loss": 4.066737635072946,
      "tokens_seen": 354761728
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00045075225677031097,
      "loss": 2.7746,
      "theoretical_loss": 4.066651893337804,
      "tokens_seen": 354827264
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004507422266800401,
      "loss": 3.0798,
      "theoretical_loss": 4.066566171870799,
      "tokens_seen": 354892800
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00045073219658976933,
      "loss": 3.0763,
      "theoretical_loss": 4.066480470663401,
      "tokens_seen": 354958336
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004507221664994985,
      "loss": 3.2006,
      "theoretical_loss": 4.06639478970708,
      "tokens_seen": 355023872
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004507121364092277,
      "loss": 2.9487,
      "theoretical_loss": 4.066309128993316,
      "tokens_seen": 355089408
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004507021063189569,
      "loss": 3.2167,
      "theoretical_loss": 4.0662234885135895,
      "tokens_seen": 355154944
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00045069207622868606,
      "loss": 2.998,
      "theoretical_loss": 4.066137868259391,
      "tokens_seen": 355220480
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00045068204613841524,
      "loss": 3.0349,
      "theoretical_loss": 4.0660522682222116,
      "tokens_seen": 355286016
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004506720160481445,
      "loss": 3.0877,
      "theoretical_loss": 4.065966688393551,
      "tokens_seen": 355351552
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004506619859578736,
      "loss": 2.874,
      "theoretical_loss": 4.065881128764912,
      "tokens_seen": 355417088
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00045065195586760284,
      "loss": 3.1169,
      "theoretical_loss": 4.065795589327804,
      "tokens_seen": 355482624
    },
    {
      "epoch": 1.02,
      "objective/train/docs_used": 473851,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.1546261310577393,
      "objective/train/theoretical_loss": 4.065731447995559,
      "objective/train/tokens_used": 375991776,
      "theoretical_loss": 4.065731447995559,
      "tokens_seen": 355531776
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00045064192577733196,
      "loss": 2.968,
      "theoretical_loss": 4.06571007007374,
      "tokens_seen": 355548160
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004506318956870612,
      "loss": 3.0017,
      "theoretical_loss": 4.065624570994239,
      "tokens_seen": 355613696
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004506218655967904,
      "loss": 3.1213,
      "theoretical_loss": 4.065539092080826,
      "tokens_seen": 355679232
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00045061183550651956,
      "loss": 3.0196,
      "theoretical_loss": 4.0654536333250295,
      "tokens_seen": 355744768
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00045060180541624874,
      "loss": 2.9044,
      "theoretical_loss": 4.065368194718383,
      "tokens_seen": 355810304
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004505917753259779,
      "loss": 3.0828,
      "theoretical_loss": 4.065282776252427,
      "tokens_seen": 355875840
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004505817452357071,
      "loss": 2.883,
      "theoretical_loss": 4.065197377918706,
      "tokens_seen": 355941376
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00045057171514543634,
      "loss": 2.8719,
      "theoretical_loss": 4.065111999708767,
      "tokens_seen": 356006912
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00045056168505516547,
      "loss": 2.7669,
      "theoretical_loss": 4.065026641614169,
      "tokens_seen": 356072448
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004505516549648947,
      "loss": 2.9525,
      "theoretical_loss": 4.064941303626469,
      "tokens_seen": 356137984
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004505416248746239,
      "loss": 2.9686,
      "theoretical_loss": 4.064855985737234,
      "tokens_seen": 356203520
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00045053159478435306,
      "loss": 3.0237,
      "theoretical_loss": 4.064770687938031,
      "tokens_seen": 356269056
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00045052156469408224,
      "loss": 3.1213,
      "theoretical_loss": 4.064685410220437,
      "tokens_seen": 356334592
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004505115346038114,
      "loss": 3.1203,
      "theoretical_loss": 4.064600152576032,
      "tokens_seen": 356400128
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004505015045135406,
      "loss": 3.2214,
      "theoretical_loss": 4.0645149149964,
      "tokens_seen": 356465664
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00045049147442326984,
      "loss": 2.9673,
      "theoretical_loss": 4.064429697473134,
      "tokens_seen": 356531200
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00045048144433299897,
      "loss": 2.9293,
      "theoretical_loss": 4.064344499997826,
      "tokens_seen": 356596736
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004504714142427282,
      "loss": 2.7086,
      "theoretical_loss": 4.06425932256208,
      "tokens_seen": 356662272
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00045046138415245733,
      "loss": 3.1569,
      "theoretical_loss": 4.064174165157499,
      "tokens_seen": 356727808
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00045045135406218657,
      "loss": 2.777,
      "theoretical_loss": 4.064089027775694,
      "tokens_seen": 356793344
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004504413239719158,
      "loss": 2.8872,
      "theoretical_loss": 4.064003910408281,
      "tokens_seen": 356858880
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00045043129388164493,
      "loss": 3.1839,
      "theoretical_loss": 4.06391881304688,
      "tokens_seen": 356924416
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00045042126379137416,
      "loss": 3.1245,
      "theoretical_loss": 4.063833735683118,
      "tokens_seen": 356989952
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00045041123370110334,
      "loss": 2.964,
      "theoretical_loss": 4.063748678308624,
      "tokens_seen": 357055488
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004504012036108325,
      "loss": 3.1374,
      "theoretical_loss": 4.063663640915035,
      "tokens_seen": 357121024
    },
    {
      "epoch": 1.02,
      "objective/train/docs_used": 475208,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.0372159481048584,
      "objective/train/theoretical_loss": 4.063599875977284,
      "objective/train/tokens_used": 377630176,
      "theoretical_loss": 4.063599875977284,
      "tokens_seen": 357170176
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004503911735205617,
      "loss": 3.083,
      "theoretical_loss": 4.063578623493992,
      "tokens_seen": 357186560
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004503811434302909,
      "loss": 3.0497,
      "theoretical_loss": 4.06349362603714,
      "tokens_seen": 357252096
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00045037111334002007,
      "loss": 3.3375,
      "theoretical_loss": 4.0634086485361305,
      "tokens_seen": 357317632
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004503610832497493,
      "loss": 2.9054,
      "theoretical_loss": 4.063323690982619,
      "tokens_seen": 357383168
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00045035105315947843,
      "loss": 2.9316,
      "theoretical_loss": 4.0632387533682675,
      "tokens_seen": 357448704
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00045034102306920767,
      "loss": 3.0126,
      "theoretical_loss": 4.06315383568474,
      "tokens_seen": 357514240
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004503309929789368,
      "loss": 3.0289,
      "theoretical_loss": 4.063068937923709,
      "tokens_seen": 357579776
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00045032096288866603,
      "loss": 2.7174,
      "theoretical_loss": 4.062984060076849,
      "tokens_seen": 357645312
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004503109327983952,
      "loss": 3.0905,
      "theoretical_loss": 4.062899202135844,
      "tokens_seen": 357710848
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004503009027081244,
      "loss": 2.8817,
      "theoretical_loss": 4.062814364092376,
      "tokens_seen": 357776384
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00045029087261785357,
      "loss": 2.7171,
      "theoretical_loss": 4.062729545938138,
      "tokens_seen": 357841920
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00045028084252758275,
      "loss": 2.963,
      "theoretical_loss": 4.062644747664827,
      "tokens_seen": 357907456
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00045027081243731193,
      "loss": 3.0307,
      "theoretical_loss": 4.062559969264141,
      "tokens_seen": 357972992
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00045026078234704117,
      "loss": 3.2462,
      "theoretical_loss": 4.062475210727789,
      "tokens_seen": 358038528
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004502507522567703,
      "loss": 2.9787,
      "theoretical_loss": 4.06239047204748,
      "tokens_seen": 358104064
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00045024072216649953,
      "loss": 2.927,
      "theoretical_loss": 4.06230575321493,
      "tokens_seen": 358169600
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004502306920762287,
      "loss": 3.0136,
      "theoretical_loss": 4.06222105422186,
      "tokens_seen": 358235136
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004502206619859579,
      "loss": 3.1244,
      "theoretical_loss": 4.062136375059996,
      "tokens_seen": 358300672
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004502106318956871,
      "loss": 2.8825,
      "theoretical_loss": 4.062051715721069,
      "tokens_seen": 358366208
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00045020060180541626,
      "loss": 3.108,
      "theoretical_loss": 4.061967076196815,
      "tokens_seen": 358431744
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00045019057171514544,
      "loss": 3.1374,
      "theoretical_loss": 4.061882456478973,
      "tokens_seen": 358497280
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004501805416248747,
      "loss": 3.17,
      "theoretical_loss": 4.0617978565592905,
      "tokens_seen": 358562816
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004501705115346038,
      "loss": 3.2081,
      "theoretical_loss": 4.061713276429517,
      "tokens_seen": 358628352
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00045016048144433304,
      "loss": 3.1158,
      "theoretical_loss": 4.061628716081408,
      "tokens_seen": 358693888
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00045015045135406216,
      "loss": 3.176,
      "theoretical_loss": 4.061544175506725,
      "tokens_seen": 358759424
    },
    {
      "epoch": 1.02,
      "objective/train/docs_used": 475771,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.862164258956909,
      "objective/train/theoretical_loss": 4.061480783047069,
      "objective/train/tokens_used": 379268576,
      "theoretical_loss": 4.061480783047069,
      "tokens_seen": 358808576
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004501404212637914,
      "loss": 3.0135,
      "theoretical_loss": 4.061459654697233,
      "tokens_seen": 358824960
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004501303911735206,
      "loss": 3.2575,
      "theoretical_loss": 4.061375153644701,
      "tokens_seen": 358890496
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00045012036108324976,
      "loss": 3.0728,
      "theoretical_loss": 4.061290672340906,
      "tokens_seen": 358956032
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00045011033099297894,
      "loss": 3.1692,
      "theoretical_loss": 4.061206210777627,
      "tokens_seen": 359021568
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004501003009027081,
      "loss": 3.0361,
      "theoretical_loss": 4.0611217689466494,
      "tokens_seen": 359087104
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004500902708124373,
      "loss": 3.0446,
      "theoretical_loss": 4.061037346839764,
      "tokens_seen": 359152640
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00045008024072216654,
      "loss": 3.0256,
      "theoretical_loss": 4.060952944448765,
      "tokens_seen": 359218176
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00045007021063189567,
      "loss": 3.0506,
      "theoretical_loss": 4.060868561765452,
      "tokens_seen": 359283712
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004500601805416249,
      "loss": 3.0259,
      "theoretical_loss": 4.060784198781631,
      "tokens_seen": 359349248
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004500501504513541,
      "loss": 3.1077,
      "theoretical_loss": 4.06069985548911,
      "tokens_seen": 359414784
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00045004012036108326,
      "loss": 3.1462,
      "theoretical_loss": 4.060615531879705,
      "tokens_seen": 359480320
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00045003009027081244,
      "loss": 3.0044,
      "theoretical_loss": 4.0605312279452335,
      "tokens_seen": 359545856
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004500200601805416,
      "loss": 2.947,
      "theoretical_loss": 4.060446943677523,
      "tokens_seen": 359611392
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004500100300902708,
      "loss": 3.0459,
      "theoretical_loss": 4.060362679068399,
      "tokens_seen": 359676928
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00045000000000000004,
      "loss": 3.2938,
      "theoretical_loss": 4.060278434109699,
      "tokens_seen": 359742464
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044998996990972917,
      "loss": 2.9767,
      "theoretical_loss": 4.060194208793259,
      "tokens_seen": 359808000
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004499799398194584,
      "loss": 2.9244,
      "theoretical_loss": 4.060110003110925,
      "tokens_seen": 359873536
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044996990972918753,
      "loss": 3.1669,
      "theoretical_loss": 4.0600258170545445,
      "tokens_seen": 359939072
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044995987963891677,
      "loss": 3.0841,
      "theoretical_loss": 4.059941650615972,
      "tokens_seen": 360004608
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044994984954864595,
      "loss": 3.0194,
      "theoretical_loss": 4.059857503787066,
      "tokens_seen": 360070144
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044993981945837513,
      "loss": 2.9446,
      "theoretical_loss": 4.059773376559689,
      "tokens_seen": 360135680
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004499297893681043,
      "loss": 2.8132,
      "theoretical_loss": 4.05968926892571,
      "tokens_seen": 360201216
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044991975927783355,
      "loss": 3.0256,
      "theoretical_loss": 4.059605180877001,
      "tokens_seen": 360266752
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044990972918756267,
      "loss": 3.0306,
      "theoretical_loss": 4.05952111240544,
      "tokens_seen": 360332288
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004498996990972919,
      "loss": 2.9097,
      "theoretical_loss": 4.0594370635029104,
      "tokens_seen": 360397824
    },
    {
      "epoch": 1.02,
      "objective/train/docs_used": 477253,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.3396573066711426,
      "objective/train/theoretical_loss": 4.059374039663311,
      "objective/train/tokens_used": 380906976,
      "theoretical_loss": 4.059374039663311,
      "tokens_seen": 360446976
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044988966900702103,
      "loss": 3.1348,
      "theoretical_loss": 4.0593530341613,
      "tokens_seen": 360463360
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044987963891675027,
      "loss": 2.9424,
      "theoretical_loss": 4.059269024372501,
      "tokens_seen": 360528896
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044986960882647945,
      "loss": 3.1762,
      "theoretical_loss": 4.05918503412841,
      "tokens_seen": 360594432
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044985957873620863,
      "loss": 3.2145,
      "theoretical_loss": 4.059101063420929,
      "tokens_seen": 360659968
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004498495486459378,
      "loss": 2.899,
      "theoretical_loss": 4.059017112241966,
      "tokens_seen": 360725504
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.000449839518555667,
      "loss": 3.1533,
      "theoretical_loss": 4.058933180583431,
      "tokens_seen": 360791040
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004498294884653962,
      "loss": 3.015,
      "theoretical_loss": 4.0588492684372435,
      "tokens_seen": 360856576
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004498194583751254,
      "loss": 3.1045,
      "theoretical_loss": 4.058765375795321,
      "tokens_seen": 360922112
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044980942828485454,
      "loss": 3.291,
      "theoretical_loss": 4.058681502649593,
      "tokens_seen": 360987648
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044979939819458377,
      "loss": 2.8875,
      "theoretical_loss": 4.058597648991988,
      "tokens_seen": 361053184
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004497893681043129,
      "loss": 2.8632,
      "theoretical_loss": 4.058513814814443,
      "tokens_seen": 361118720
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044977933801404213,
      "loss": 2.8654,
      "theoretical_loss": 4.058430000108898,
      "tokens_seen": 361184256
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004497693079237713,
      "loss": 3.1835,
      "theoretical_loss": 4.058346204867299,
      "tokens_seen": 361249792
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004497592778335005,
      "loss": 3.1647,
      "theoretical_loss": 4.058262429081596,
      "tokens_seen": 361315328
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004497492477432297,
      "loss": 3.0466,
      "theoretical_loss": 4.058178672743744,
      "tokens_seen": 361380864
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004497392176529589,
      "loss": 3.0466,
      "theoretical_loss": 4.058094935845703,
      "tokens_seen": 361446400
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044972918756268804,
      "loss": 3.0096,
      "theoretical_loss": 4.058011218379436,
      "tokens_seen": 361511936
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004497191574724173,
      "loss": 3.2179,
      "theoretical_loss": 4.057927520336913,
      "tokens_seen": 361577472
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004497091273821464,
      "loss": 3.4683,
      "theoretical_loss": 4.05784384171011,
      "tokens_seen": 361643008
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044969909729187564,
      "loss": 3.0278,
      "theoretical_loss": 4.057760182491003,
      "tokens_seen": 361708544
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004496890672016049,
      "loss": 2.9469,
      "theoretical_loss": 4.057676542671577,
      "tokens_seen": 361774080
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.000449679037111334,
      "loss": 3.0088,
      "theoretical_loss": 4.0575929222438205,
      "tokens_seen": 361839616
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044966900702106324,
      "loss": 2.9364,
      "theoretical_loss": 4.057509321199726,
      "tokens_seen": 361905152
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044965897693079236,
      "loss": 2.808,
      "theoretical_loss": 4.057425739531292,
      "tokens_seen": 361970688
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004496489468405216,
      "loss": 3.017,
      "theoretical_loss": 4.057342177230519,
      "tokens_seen": 362036224
    },
    {
      "epoch": 1.02,
      "objective/train/docs_used": 478095,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.1937191486358643,
      "objective/train/theoretical_loss": 4.057279518210161,
      "objective/train/tokens_used": 382545376,
      "theoretical_loss": 4.057279518210161,
      "tokens_seen": 362085376
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004496389167502508,
      "loss": 3.2181,
      "theoretical_loss": 4.057258634289418,
      "tokens_seen": 362101760
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044962888665997996,
      "loss": 2.8883,
      "theoretical_loss": 4.057175110699999,
      "tokens_seen": 362167296
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044961885656970914,
      "loss": 2.9905,
      "theoretical_loss": 4.057091606454279,
      "tokens_seen": 362232832
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004496088264794383,
      "loss": 2.9349,
      "theoretical_loss": 4.057008121544279,
      "tokens_seen": 362298368
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004495987963891675,
      "loss": 3.0119,
      "theoretical_loss": 4.056924655962027,
      "tokens_seen": 362363904
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044958876629889674,
      "loss": 2.8529,
      "theoretical_loss": 4.056841209699553,
      "tokens_seen": 362429440
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044957873620862587,
      "loss": 3.0478,
      "theoretical_loss": 4.0567577827488925,
      "tokens_seen": 362494976
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004495687061183551,
      "loss": 3.0898,
      "theoretical_loss": 4.056674375102086,
      "tokens_seen": 362560512
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004495586760280843,
      "loss": 2.9169,
      "theoretical_loss": 4.05659098675118,
      "tokens_seen": 362626048
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044954864593781346,
      "loss": 2.6805,
      "theoretical_loss": 4.056507617688223,
      "tokens_seen": 362691584
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044953861584754264,
      "loss": 3.002,
      "theoretical_loss": 4.05642426790527,
      "tokens_seen": 362757120
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004495285857572718,
      "loss": 3.2354,
      "theoretical_loss": 4.056340937394381,
      "tokens_seen": 362822656
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.000449518555667001,
      "loss": 2.8529,
      "theoretical_loss": 4.0562576261476195,
      "tokens_seen": 362888192
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044950852557673024,
      "loss": 3.0567,
      "theoretical_loss": 4.056174334157054,
      "tokens_seen": 362953728
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044949849548645937,
      "loss": 3.1938,
      "theoretical_loss": 4.056091061414759,
      "tokens_seen": 363019264
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004494884653961886,
      "loss": 2.914,
      "theoretical_loss": 4.0560078079128115,
      "tokens_seen": 363084800
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044947843530591773,
      "loss": 2.9784,
      "theoretical_loss": 4.055924573643295,
      "tokens_seen": 363150336
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044946840521564697,
      "loss": 2.9387,
      "theoretical_loss": 4.055841358598297,
      "tokens_seen": 363215872
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044945837512537615,
      "loss": 3.182,
      "theoretical_loss": 4.055758162769909,
      "tokens_seen": 363281408
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044944834503510533,
      "loss": 3.0795,
      "theoretical_loss": 4.055674986150228,
      "tokens_seen": 363346944
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004494383149448345,
      "loss": 3.0805,
      "theoretical_loss": 4.055591828731356,
      "tokens_seen": 363412480
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044942828485456375,
      "loss": 3.2346,
      "theoretical_loss": 4.0555086905054,
      "tokens_seen": 363478016
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044941825476429287,
      "loss": 3.0509,
      "theoretical_loss": 4.05542557146447,
      "tokens_seen": 363543552
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004494082246740221,
      "loss": 2.9644,
      "theoretical_loss": 4.055342471600682,
      "tokens_seen": 363609088
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044939819458375123,
      "loss": 3.0934,
      "theoretical_loss": 4.055259390906155,
      "tokens_seen": 363674624
    },
    {
      "epoch": 1.02,
      "objective/train/docs_used": 479451,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.949183702468872,
      "objective/train/theoretical_loss": 4.05519709296035,
      "objective/train/tokens_used": 384183776,
      "theoretical_loss": 4.05519709296035,
      "tokens_seen": 363723776
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044938816449348047,
      "loss": 2.8857,
      "theoretical_loss": 4.055176329373015,
      "tokens_seen": 363740160
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044937813440320965,
      "loss": 2.8673,
      "theoretical_loss": 4.055093286993392,
      "tokens_seen": 363805696
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044936810431293883,
      "loss": 2.9403,
      "theoretical_loss": 4.055010263759419,
      "tokens_seen": 363871232
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.000449358074222668,
      "loss": 2.9892,
      "theoretical_loss": 4.054927259663235,
      "tokens_seen": 363936768
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004493480441323972,
      "loss": 3.1854,
      "theoretical_loss": 4.054844274696984,
      "tokens_seen": 364002304
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004493380140421264,
      "loss": 3.1805,
      "theoretical_loss": 4.0547613088528145,
      "tokens_seen": 364067840
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004493279839518556,
      "loss": 2.9269,
      "theoretical_loss": 4.054678362122878,
      "tokens_seen": 364133376
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044931795386158474,
      "loss": 3.0968,
      "theoretical_loss": 4.0545954344993325,
      "tokens_seen": 364198912
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044930792377131397,
      "loss": 3.1839,
      "theoretical_loss": 4.05451252597434,
      "tokens_seen": 364264448
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004492978936810431,
      "loss": 3.0138,
      "theoretical_loss": 4.054429636540068,
      "tokens_seen": 364329984
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044928786359077234,
      "loss": 2.8682,
      "theoretical_loss": 4.0543467661886865,
      "tokens_seen": 364395520
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004492778335005015,
      "loss": 2.8514,
      "theoretical_loss": 4.054263914912372,
      "tokens_seen": 364461056
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004492678034102307,
      "loss": 2.8017,
      "theoretical_loss": 4.0541810827033045,
      "tokens_seen": 364526592
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004492577733199599,
      "loss": 3.0815,
      "theoretical_loss": 4.054098269553669,
      "tokens_seen": 364592128
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004492477432296891,
      "loss": 2.8874,
      "theoretical_loss": 4.054015475455656,
      "tokens_seen": 364657664
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044923771313941824,
      "loss": 3.159,
      "theoretical_loss": 4.053932700401459,
      "tokens_seen": 364723200
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004492276830491475,
      "loss": 3.033,
      "theoretical_loss": 4.053849944383279,
      "tokens_seen": 364788736
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004492176529588766,
      "loss": 3.0449,
      "theoretical_loss": 4.0537672073933155,
      "tokens_seen": 364854272
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044920762286860584,
      "loss": 3.2518,
      "theoretical_loss": 4.05368448942378,
      "tokens_seen": 364919808
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.000449197592778335,
      "loss": 3.0096,
      "theoretical_loss": 4.053601790466884,
      "tokens_seen": 364985344
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004491875626880642,
      "loss": 3.032,
      "theoretical_loss": 4.053519110514845,
      "tokens_seen": 365050880
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004491775325977934,
      "loss": 3.2143,
      "theoretical_loss": 4.053436449559886,
      "tokens_seen": 365116416
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044916750250752256,
      "loss": 2.92,
      "theoretical_loss": 4.05335380759423,
      "tokens_seen": 365181952
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044915747241725174,
      "loss": 3.1182,
      "theoretical_loss": 4.053271184610111,
      "tokens_seen": 365247488
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.000449147442326981,
      "loss": 2.9261,
      "theoretical_loss": 4.053188580599764,
      "tokens_seen": 365313024
    },
    {
      "epoch": 1.02,
      "objective/train/docs_used": 480072,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.2834551334381104,
      "objective/train/theoretical_loss": 4.053126640038872,
      "objective/train/tokens_used": 385822176,
      "theoretical_loss": 4.053126640038872,
      "tokens_seen": 365362176
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004491374122367101,
      "loss": 2.9938,
      "theoretical_loss": 4.053105995555429,
      "tokens_seen": 365378560
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044912738214643934,
      "loss": 2.9087,
      "theoretical_loss": 4.0530234294693495,
      "tokens_seen": 365444096
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044911735205616847,
      "loss": 2.8535,
      "theoretical_loss": 4.052940882333777,
      "tokens_seen": 365509632
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004491073219658977,
      "loss": 3.0883,
      "theoretical_loss": 4.052858354140964,
      "tokens_seen": 365575168
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004490972918756269,
      "loss": 2.8664,
      "theoretical_loss": 4.052775844883168,
      "tokens_seen": 365640704
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044908726178535607,
      "loss": 3.0148,
      "theoretical_loss": 4.0526933545526544,
      "tokens_seen": 365706240
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044907723169508525,
      "loss": 3.1162,
      "theoretical_loss": 4.052610883141689,
      "tokens_seen": 365771776
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004490672016048145,
      "loss": 3.2736,
      "theoretical_loss": 4.052528430642544,
      "tokens_seen": 365837312
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004490571715145436,
      "loss": 2.9012,
      "theoretical_loss": 4.052445997047497,
      "tokens_seen": 365902848
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044904714142427284,
      "loss": 3.0513,
      "theoretical_loss": 4.052363582348827,
      "tokens_seen": 365968384
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044903711133400197,
      "loss": 3.3379,
      "theoretical_loss": 4.052281186538821,
      "tokens_seen": 366033920
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004490270812437312,
      "loss": 3.041,
      "theoretical_loss": 4.052198809609769,
      "tokens_seen": 366099456
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004490170511534604,
      "loss": 2.924,
      "theoretical_loss": 4.0521164515539665,
      "tokens_seen": 366164992
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044900702106318957,
      "loss": 2.9846,
      "theoretical_loss": 4.052034112363711,
      "tokens_seen": 366230528
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044899699097291875,
      "loss": 3.087,
      "theoretical_loss": 4.051951792031308,
      "tokens_seen": 366296064
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044898696088264793,
      "loss": 3.0914,
      "theoretical_loss": 4.051869490549064,
      "tokens_seen": 366361600
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004489769307923771,
      "loss": 2.9215,
      "theoretical_loss": 4.051787207909294,
      "tokens_seen": 366427136
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044896690070210635,
      "loss": 2.9358,
      "theoretical_loss": 4.051704944104313,
      "tokens_seen": 366492672
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004489568706118355,
      "loss": 3.0429,
      "theoretical_loss": 4.051622699126444,
      "tokens_seen": 366558208
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004489468405215647,
      "loss": 3.0332,
      "theoretical_loss": 4.0515404729680125,
      "tokens_seen": 366623744
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044893681043129395,
      "loss": 3.057,
      "theoretical_loss": 4.051458265621351,
      "tokens_seen": 366689280
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044892678034102307,
      "loss": 3.0415,
      "theoretical_loss": 4.0513760770787925,
      "tokens_seen": 366754816
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004489167502507523,
      "loss": 3.3661,
      "theoretical_loss": 4.051293907332678,
      "tokens_seen": 366820352
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044890672016048143,
      "loss": 3.2608,
      "theoretical_loss": 4.051211756375352,
      "tokens_seen": 366885888
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044889669007021067,
      "loss": 3.042,
      "theoretical_loss": 4.051129624199163,
      "tokens_seen": 366951424
    },
    {
      "epoch": 1.02,
      "objective/train/docs_used": 481451,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.963195323944092,
      "objective/train/theoretical_loss": 4.051068037387542,
      "objective/train/tokens_used": 387460576,
      "theoretical_loss": 4.051068037387542,
      "tokens_seen": 367000576
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044888665997993985,
      "loss": 3.0687,
      "theoretical_loss": 4.051047510796463,
      "tokens_seen": 367016960
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044887662988966903,
      "loss": 2.9519,
      "theoretical_loss": 4.050965416159612,
      "tokens_seen": 367082496
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004488665997993982,
      "loss": 2.8079,
      "theoretical_loss": 4.05088334028097,
      "tokens_seen": 367148032
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004488565697091274,
      "loss": 3.114,
      "theoretical_loss": 4.050801283152905,
      "tokens_seen": 367213568
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004488465396188566,
      "loss": 2.8896,
      "theoretical_loss": 4.050719244767788,
      "tokens_seen": 367279104
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004488365095285858,
      "loss": 2.8136,
      "theoretical_loss": 4.0506372251179945,
      "tokens_seen": 367344640
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044882647943831494,
      "loss": 2.8483,
      "theoretical_loss": 4.050555224195904,
      "tokens_seen": 367410176
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004488164493480442,
      "loss": 3.2117,
      "theoretical_loss": 4.050473241993901,
      "tokens_seen": 367475712
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004488064192577733,
      "loss": 3.0052,
      "theoretical_loss": 4.0503912785043745,
      "tokens_seen": 367541248
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044879638916750254,
      "loss": 2.9371,
      "theoretical_loss": 4.050309333719719,
      "tokens_seen": 367606784
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004487863590772317,
      "loss": 3.207,
      "theoretical_loss": 4.050227407632331,
      "tokens_seen": 367672320
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004487763289869609,
      "loss": 3.1296,
      "theoretical_loss": 4.050145500234613,
      "tokens_seen": 367737856
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004487662988966901,
      "loss": 2.7027,
      "theoretical_loss": 4.050063611518974,
      "tokens_seen": 367803392
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004487562688064193,
      "loss": 3.1566,
      "theoretical_loss": 4.049981741477821,
      "tokens_seen": 367868928
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044874623871614844,
      "loss": 2.958,
      "theoretical_loss": 4.049899890103573,
      "tokens_seen": 367934464
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004487362086258777,
      "loss": 3.1707,
      "theoretical_loss": 4.049818057388649,
      "tokens_seen": 368000000
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004487261785356068,
      "loss": 3.0476,
      "theoretical_loss": 4.049736243325473,
      "tokens_seen": 368065536
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044871614844533604,
      "loss": 2.7808,
      "theoretical_loss": 4.049654447906473,
      "tokens_seen": 368131072
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004487061183550652,
      "loss": 3.0115,
      "theoretical_loss": 4.049572671124086,
      "tokens_seen": 368196608
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004486960882647944,
      "loss": 3.1322,
      "theoretical_loss": 4.049490912970745,
      "tokens_seen": 368262144
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004486860581745236,
      "loss": 2.9311,
      "theoretical_loss": 4.049409173438896,
      "tokens_seen": 368327680
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044867602808425276,
      "loss": 3.1052,
      "theoretical_loss": 4.049327452520982,
      "tokens_seen": 368393216
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044866599799398194,
      "loss": 2.8961,
      "theoretical_loss": 4.049245750209458,
      "tokens_seen": 368458752
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004486559679037112,
      "loss": 2.943,
      "theoretical_loss": 4.049164066496776,
      "tokens_seen": 368524288
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004486459378134403,
      "loss": 2.9977,
      "theoretical_loss": 4.049082401375397,
      "tokens_seen": 368589824
    },
    {
      "epoch": 1.02,
      "objective/train/docs_used": 482108,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.12469744682312,
      "objective/train/theoretical_loss": 4.049021164730373,
      "objective/train/tokens_used": 389098976,
      "theoretical_loss": 4.049021164730373,
      "tokens_seen": 368638976
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044863590772316954,
      "loss": 3.2333,
      "theoretical_loss": 4.049000754837786,
      "tokens_seen": 368655360
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044862587763289867,
      "loss": 2.8482,
      "theoretical_loss": 4.04891912687641,
      "tokens_seen": 368720896
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004486158475426279,
      "loss": 2.9425,
      "theoretical_loss": 4.048837517483744,
      "tokens_seen": 368786432
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004486058174523571,
      "loss": 3.0541,
      "theoretical_loss": 4.048755926652262,
      "tokens_seen": 368851968
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044859578736208627,
      "loss": 2.9036,
      "theoretical_loss": 4.04867435437445,
      "tokens_seen": 368917504
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044858575727181545,
      "loss": 3.1034,
      "theoretical_loss": 4.0485928006427905,
      "tokens_seen": 368983040
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004485757271815447,
      "loss": 3.0932,
      "theoretical_loss": 4.0485112654497755,
      "tokens_seen": 369048576
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004485656970912738,
      "loss": 3.0221,
      "theoretical_loss": 4.048429748787899,
      "tokens_seen": 369114112
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044855566700100304,
      "loss": 2.7535,
      "theoretical_loss": 4.048348250649662,
      "tokens_seen": 369179648
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044854563691073217,
      "loss": 3.1027,
      "theoretical_loss": 4.048266771027565,
      "tokens_seen": 369245184
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004485356068204614,
      "loss": 3.3042,
      "theoretical_loss": 4.048185309914119,
      "tokens_seen": 369310720
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004485255767301906,
      "loss": 3.091,
      "theoretical_loss": 4.048103867301836,
      "tokens_seen": 369376256
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044851554663991977,
      "loss": 2.9852,
      "theoretical_loss": 4.04802244318323,
      "tokens_seen": 369441792
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044850551654964895,
      "loss": 3.1036,
      "theoretical_loss": 4.047941037550825,
      "tokens_seen": 369507328
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044849548645937813,
      "loss": 2.9854,
      "theoretical_loss": 4.047859650397145,
      "tokens_seen": 369572864
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004484854563691073,
      "loss": 3.1245,
      "theoretical_loss": 4.04777828171472,
      "tokens_seen": 369638400
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044847542627883655,
      "loss": 2.9639,
      "theoretical_loss": 4.047696931496084,
      "tokens_seen": 369703936
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004484653961885657,
      "loss": 3.0576,
      "theoretical_loss": 4.047615599733775,
      "tokens_seen": 369769472
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004484553660982949,
      "loss": 3.1132,
      "theoretical_loss": 4.047534286420337,
      "tokens_seen": 369835008
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044844533600802404,
      "loss": 3.0274,
      "theoretical_loss": 4.047452991548315,
      "tokens_seen": 369900544
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044843530591775327,
      "loss": 3.2649,
      "theoretical_loss": 4.047371715110262,
      "tokens_seen": 369966080
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044842527582748245,
      "loss": 2.865,
      "theoretical_loss": 4.047290457098733,
      "tokens_seen": 370031616
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044841524573721163,
      "loss": 3.3263,
      "theoretical_loss": 4.047209217506289,
      "tokens_seen": 370097152
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004484052156469408,
      "loss": 2.9046,
      "theoretical_loss": 4.047127996325493,
      "tokens_seen": 370162688
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044839518555667005,
      "loss": 3.1456,
      "theoretical_loss": 4.0470467935489145,
      "tokens_seen": 370228224
    },
    {
      "epoch": 1.02,
      "objective/train/docs_used": 483097,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9543488025665283,
      "objective/train/theoretical_loss": 4.04698590353978,
      "objective/train/tokens_used": 390737376,
      "theoretical_loss": 4.04698590353978,
      "tokens_seen": 370277376
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004483851554663992,
      "loss": 2.8715,
      "theoretical_loss": 4.046965609169126,
      "tokens_seen": 370293760
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004483751253761284,
      "loss": 2.9675,
      "theoretical_loss": 4.046884443178706,
      "tokens_seen": 370359296
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044836509528585754,
      "loss": 3.2167,
      "theoretical_loss": 4.046803295570235,
      "tokens_seen": 370424832
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004483550651955868,
      "loss": 3.1818,
      "theoretical_loss": 4.046722166336298,
      "tokens_seen": 370490368
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044834503510531596,
      "loss": 3.1305,
      "theoretical_loss": 4.046641055469488,
      "tokens_seen": 370555904
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044833500501504514,
      "loss": 2.8626,
      "theoretical_loss": 4.046559962962398,
      "tokens_seen": 370621440
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004483249749247743,
      "loss": 2.9436,
      "theoretical_loss": 4.046478888807626,
      "tokens_seen": 370686976
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004483149448345035,
      "loss": 3.2137,
      "theoretical_loss": 4.046397832997776,
      "tokens_seen": 370752512
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004483049147442327,
      "loss": 3.1725,
      "theoretical_loss": 4.046316795525456,
      "tokens_seen": 370818048
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004482948846539619,
      "loss": 3.1907,
      "theoretical_loss": 4.046235776383276,
      "tokens_seen": 370883584
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044828485456369104,
      "loss": 2.9513,
      "theoretical_loss": 4.046154775563854,
      "tokens_seen": 370949120
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004482748244734203,
      "loss": 3.2516,
      "theoretical_loss": 4.046073793059808,
      "tokens_seen": 371014656
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004482647943831494,
      "loss": 3.1746,
      "theoretical_loss": 4.045992828863764,
      "tokens_seen": 371080192
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044825476429287864,
      "loss": 3.0957,
      "theoretical_loss": 4.045911882968351,
      "tokens_seen": 371145728
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004482447342026078,
      "loss": 3.1395,
      "theoretical_loss": 4.045830955366201,
      "tokens_seen": 371211264
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.000448234704112337,
      "loss": 3.0624,
      "theoretical_loss": 4.045750046049952,
      "tokens_seen": 371276800
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004482246740220662,
      "loss": 2.9705,
      "theoretical_loss": 4.045669155012246,
      "tokens_seen": 371342336
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004482146439317954,
      "loss": 2.8102,
      "theoretical_loss": 4.045588282245728,
      "tokens_seen": 371407872
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044820461384152455,
      "loss": 3.1322,
      "theoretical_loss": 4.045507427743048,
      "tokens_seen": 371473408
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004481945837512538,
      "loss": 3.0764,
      "theoretical_loss": 4.045426591496861,
      "tokens_seen": 371538944
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044818455366098296,
      "loss": 3.0931,
      "theoretical_loss": 4.045345773499825,
      "tokens_seen": 371604480
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044817452357071214,
      "loss": 3.1696,
      "theoretical_loss": 4.045264973744604,
      "tokens_seen": 371670016
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004481644934804414,
      "loss": 2.9068,
      "theoretical_loss": 4.045184192223863,
      "tokens_seen": 371735552
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004481544633901705,
      "loss": 3.0335,
      "theoretical_loss": 4.045103428930275,
      "tokens_seen": 371801088
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044814443329989974,
      "loss": 3.166,
      "theoretical_loss": 4.045022683856517,
      "tokens_seen": 371866624
    },
    {
      "epoch": 1.02,
      "objective/train/docs_used": 483681,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6399648189544678,
      "objective/train/theoretical_loss": 4.044962137003555,
      "objective/train/tokens_used": 392375776,
      "theoretical_loss": 4.044962137003555,
      "tokens_seen": 371915776
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044813440320962887,
      "loss": 2.9421,
      "theoretical_loss": 4.044941956995265,
      "tokens_seen": 371932160
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004481243731193581,
      "loss": 3.0246,
      "theoretical_loss": 4.044861248339206,
      "tokens_seen": 371997696
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004481143430290873,
      "loss": 3.1591,
      "theoretical_loss": 4.044780557881026,
      "tokens_seen": 372063232
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044810431293881647,
      "loss": 2.8982,
      "theoretical_loss": 4.04469988561342,
      "tokens_seen": 372128768
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044809428284854565,
      "loss": 2.7799,
      "theoretical_loss": 4.044619231529083,
      "tokens_seen": 372194304
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004480842527582749,
      "loss": 3.1344,
      "theoretical_loss": 4.0445385956207165,
      "tokens_seen": 372259840
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.000448074222668004,
      "loss": 2.8566,
      "theoretical_loss": 4.0444579778810255,
      "tokens_seen": 372325376
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044806419257773324,
      "loss": 2.9545,
      "theoretical_loss": 4.04437737830272,
      "tokens_seen": 372390912
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044805416248746237,
      "loss": 3.1218,
      "theoretical_loss": 4.044296796878512,
      "tokens_seen": 372456448
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004480441323971916,
      "loss": 3.2154,
      "theoretical_loss": 4.044216233601119,
      "tokens_seen": 372521984
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004480341023069208,
      "loss": 3.1728,
      "theoretical_loss": 4.044135688463267,
      "tokens_seen": 372587520
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044802407221664997,
      "loss": 2.8922,
      "theoretical_loss": 4.044055161457678,
      "tokens_seen": 372653056
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044801404212637915,
      "loss": 3.1435,
      "theoretical_loss": 4.0439746525770826,
      "tokens_seen": 372718592
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044800401203610833,
      "loss": 2.858,
      "theoretical_loss": 4.043894161814217,
      "tokens_seen": 372784128
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004479939819458375,
      "loss": 2.9771,
      "theoretical_loss": 4.043813689161819,
      "tokens_seen": 372849664
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044798395185556675,
      "loss": 3.1192,
      "theoretical_loss": 4.043733234612633,
      "tokens_seen": 372915200
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004479739217652959,
      "loss": 3.0501,
      "theoretical_loss": 4.043652798159405,
      "tokens_seen": 372980736
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004479638916750251,
      "loss": 3.1939,
      "theoretical_loss": 4.0435723797948855,
      "tokens_seen": 373046272
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044795386158475424,
      "loss": 3.0302,
      "theoretical_loss": 4.043491979511831,
      "tokens_seen": 373111808
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044794383149448347,
      "loss": 2.9471,
      "theoretical_loss": 4.043411597303,
      "tokens_seen": 373177344
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044793380140421265,
      "loss": 2.924,
      "theoretical_loss": 4.043331233161158,
      "tokens_seen": 373242880
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044792377131394183,
      "loss": 2.9933,
      "theoretical_loss": 4.043250887079072,
      "tokens_seen": 373308416
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.000447913741223671,
      "loss": 3.0149,
      "theoretical_loss": 4.043170559049514,
      "tokens_seen": 373373952
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044790371113340025,
      "loss": 2.4948,
      "theoretical_loss": 4.04309024906526,
      "tokens_seen": 373439488
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004478936810431294,
      "loss": 3.101,
      "theoretical_loss": 4.043009957119092,
      "tokens_seen": 373505024
    },
    {
      "epoch": 1.02,
      "objective/train/docs_used": 485048,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.050812244415283,
      "objective/train/theoretical_loss": 4.042949749992618,
      "objective/train/tokens_used": 394014176,
      "theoretical_loss": 4.042949749992618,
      "tokens_seen": 373554176
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004478836509528586,
      "loss": 3.1085,
      "theoretical_loss": 4.042929683203793,
      "tokens_seen": 373570560
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044787362086258774,
      "loss": 3.0532,
      "theoretical_loss": 4.042849427312152,
      "tokens_seen": 373636096
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.000447863590772317,
      "loss": 3.0551,
      "theoretical_loss": 4.042769189436964,
      "tokens_seen": 373701632
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044785356068204616,
      "loss": 3.2651,
      "theoretical_loss": 4.042688969571023,
      "tokens_seen": 373767168
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044784353059177534,
      "loss": 3.2799,
      "theoretical_loss": 4.042608767707131,
      "tokens_seen": 373832704
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004478335005015045,
      "loss": 2.9216,
      "theoretical_loss": 4.042528583838095,
      "tokens_seen": 373898240
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004478234704112337,
      "loss": 2.8679,
      "theoretical_loss": 4.042448417956722,
      "tokens_seen": 373963776
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004478134403209629,
      "loss": 3.1769,
      "theoretical_loss": 4.042368270055828,
      "tokens_seen": 374029312
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004478034102306921,
      "loss": 3.1326,
      "theoretical_loss": 4.042288140128229,
      "tokens_seen": 374094848
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044779338014042124,
      "loss": 3.1083,
      "theoretical_loss": 4.042208028166748,
      "tokens_seen": 374160384
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004477833500501505,
      "loss": 3.0138,
      "theoretical_loss": 4.04212793416421,
      "tokens_seen": 374225920
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004477733199598796,
      "loss": 2.8839,
      "theoretical_loss": 4.042047858113446,
      "tokens_seen": 374291456
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044776328986960884,
      "loss": 3.1341,
      "theoretical_loss": 4.041967800007289,
      "tokens_seen": 374356992
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.000447753259779338,
      "loss": 2.8194,
      "theoretical_loss": 4.041887759838579,
      "tokens_seen": 374422528
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004477432296890672,
      "loss": 3.1354,
      "theoretical_loss": 4.041807737600157,
      "tokens_seen": 374488064
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004477331995987964,
      "loss": 2.9551,
      "theoretical_loss": 4.041727733284871,
      "tokens_seen": 374553600
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004477231695085256,
      "loss": 3.1074,
      "theoretical_loss": 4.04164774688557,
      "tokens_seen": 374619136
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044771313941825475,
      "loss": 3.0928,
      "theoretical_loss": 4.04156777839511,
      "tokens_seen": 374684672
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.000447703109327984,
      "loss": 2.9912,
      "theoretical_loss": 4.041487827806349,
      "tokens_seen": 374750208
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004476930792377131,
      "loss": 3.1209,
      "theoretical_loss": 4.04140789511215,
      "tokens_seen": 374815744
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044768304914744234,
      "loss": 3.1041,
      "theoretical_loss": 4.041327980305381,
      "tokens_seen": 374881280
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004476730190571715,
      "loss": 3.141,
      "theoretical_loss": 4.041248083378912,
      "tokens_seen": 374946816
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004476629889669007,
      "loss": 3.1131,
      "theoretical_loss": 4.041168204325619,
      "tokens_seen": 375012352
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004476529588766299,
      "loss": 3.1638,
      "theoretical_loss": 4.041088343138382,
      "tokens_seen": 375077888
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044764292878635907,
      "loss": 3.3661,
      "theoretical_loss": 4.041008499810082,
      "tokens_seen": 375143424
    },
    {
      "epoch": 1.02,
      "objective/train/docs_used": 485815,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.0911853313446045,
      "objective/train/theoretical_loss": 4.040948629029507,
      "objective/train/tokens_used": 395652576,
      "theoretical_loss": 4.040948629029507,
      "tokens_seen": 375192576
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044763289869608825,
      "loss": 3.0192,
      "theoretical_loss": 4.040928674333609,
      "tokens_seen": 375208960
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004476228686058175,
      "loss": 3.1861,
      "theoretical_loss": 4.040848866701853,
      "tokens_seen": 375274496
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004476128385155466,
      "loss": 3.2503,
      "theoretical_loss": 4.04076907690771,
      "tokens_seen": 375340032
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044760280842527585,
      "loss": 3.1357,
      "theoretical_loss": 4.040689304944079,
      "tokens_seen": 375405568
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044759277833500503,
      "loss": 3.0223,
      "theoretical_loss": 4.0406095508038655,
      "tokens_seen": 375471104
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004475827482447342,
      "loss": 2.9539,
      "theoretical_loss": 4.040529814479976,
      "tokens_seen": 375536640
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004475727181544634,
      "loss": 3.148,
      "theoretical_loss": 4.040450095965323,
      "tokens_seen": 375602176
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044756268806419257,
      "loss": 2.8196,
      "theoretical_loss": 4.040370395252821,
      "tokens_seen": 375667712
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044755265797392175,
      "loss": 3.2455,
      "theoretical_loss": 4.040290712335391,
      "tokens_seen": 375733248
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.000447542627883651,
      "loss": 3.2465,
      "theoretical_loss": 4.0402110472059585,
      "tokens_seen": 375798784
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004475325977933801,
      "loss": 3.1096,
      "theoretical_loss": 4.040131399857449,
      "tokens_seen": 375864320
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044752256770310935,
      "loss": 3.0396,
      "theoretical_loss": 4.040051770282796,
      "tokens_seen": 375929856
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004475125376128385,
      "loss": 3.2297,
      "theoretical_loss": 4.039972158474936,
      "tokens_seen": 375995392
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004475025075225677,
      "loss": 3.0009,
      "theoretical_loss": 4.0398925644268076,
      "tokens_seen": 376060928
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004474924774322969,
      "loss": 2.9304,
      "theoretical_loss": 4.039812988131357,
      "tokens_seen": 376126464
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004474824473420261,
      "loss": 3.1235,
      "theoretical_loss": 4.03973342958153,
      "tokens_seen": 376192000
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044747241725175526,
      "loss": 2.9068,
      "theoretical_loss": 4.039653888770282,
      "tokens_seen": 376257536
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044746238716148444,
      "loss": 3.0073,
      "theoretical_loss": 4.039574365690567,
      "tokens_seen": 376323072
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004474523570712136,
      "loss": 2.8717,
      "theoretical_loss": 4.0394948603353455,
      "tokens_seen": 376388608
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044744232698094285,
      "loss": 2.926,
      "theoretical_loss": 4.039415372697583,
      "tokens_seen": 376454144
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044743229689067204,
      "loss": 2.7771,
      "theoretical_loss": 4.039335902770247,
      "tokens_seen": 376519680
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004474222668004012,
      "loss": 2.9918,
      "theoretical_loss": 4.03925645054631,
      "tokens_seen": 376585216
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044741223671013045,
      "loss": 2.9529,
      "theoretical_loss": 4.039177016018749,
      "tokens_seen": 376650752
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004474022066198596,
      "loss": 2.9578,
      "theoretical_loss": 4.039097599180543,
      "tokens_seen": 376716288
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004473921765295888,
      "loss": 2.9367,
      "theoretical_loss": 4.039018200024678,
      "tokens_seen": 376781824
    },
    {
      "epoch": 1.02,
      "objective/train/docs_used": 486601,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.832075834274292,
      "objective/train/theoretical_loss": 4.038958662257597,
      "objective/train/tokens_used": 397290976,
      "theoretical_loss": 4.038958662257597,
      "tokens_seen": 376830976
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044738214643931794,
      "loss": 2.9877,
      "theoretical_loss": 4.038938818544143,
      "tokens_seen": 376847360
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004473721163490472,
      "loss": 2.8239,
      "theoretical_loss": 4.038859454731927,
      "tokens_seen": 376912896
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044736208625877636,
      "loss": 3.158,
      "theoretical_loss": 4.0387801085810295,
      "tokens_seen": 376978432
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044735205616850554,
      "loss": 3.0955,
      "theoretical_loss": 4.038700780084451,
      "tokens_seen": 377043968
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004473420260782347,
      "loss": 3.0649,
      "theoretical_loss": 4.038621469235193,
      "tokens_seen": 377109504
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004473319959879639,
      "loss": 2.9452,
      "theoretical_loss": 4.038542176026267,
      "tokens_seen": 377175040
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004473219658976931,
      "loss": 3.0891,
      "theoretical_loss": 4.038462900450684,
      "tokens_seen": 377240576
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004473119358074223,
      "loss": 3.1298,
      "theoretical_loss": 4.03838364250146,
      "tokens_seen": 377306112
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044730190571715144,
      "loss": 3.1469,
      "theoretical_loss": 4.038304402171616,
      "tokens_seen": 377371648
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004472918756268807,
      "loss": 3.1611,
      "theoretical_loss": 4.038225179454177,
      "tokens_seen": 377437184
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004472818455366098,
      "loss": 3.165,
      "theoretical_loss": 4.038145974342169,
      "tokens_seen": 377502720
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044727181544633904,
      "loss": 2.916,
      "theoretical_loss": 4.038066786828626,
      "tokens_seen": 377568256
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004472617853560682,
      "loss": 2.9843,
      "theoretical_loss": 4.037987616906584,
      "tokens_seen": 377633792
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004472517552657974,
      "loss": 3.0845,
      "theoretical_loss": 4.037908464569083,
      "tokens_seen": 377699328
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004472417251755266,
      "loss": 3.1241,
      "theoretical_loss": 4.037829329809167,
      "tokens_seen": 377764864
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004472316950852558,
      "loss": 3.1601,
      "theoretical_loss": 4.037750212619883,
      "tokens_seen": 377830400
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044722166499498495,
      "loss": 3.0712,
      "theoretical_loss": 4.037671112994286,
      "tokens_seen": 377895936
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004472116349047142,
      "loss": 2.8286,
      "theoretical_loss": 4.037592030925429,
      "tokens_seen": 377961472
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004472016048144433,
      "loss": 2.8222,
      "theoretical_loss": 4.037512966406375,
      "tokens_seen": 378027008
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044719157472417254,
      "loss": 2.9329,
      "theoretical_loss": 4.0374339194301845,
      "tokens_seen": 378092544
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004471815446339017,
      "loss": 3.0938,
      "theoretical_loss": 4.037354889989928,
      "tokens_seen": 378158080
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004471715145436309,
      "loss": 3.2318,
      "theoretical_loss": 4.037275878078676,
      "tokens_seen": 378223616
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004471614844533601,
      "loss": 2.9692,
      "theoretical_loss": 4.037196883689504,
      "tokens_seen": 378289152
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044715145436308927,
      "loss": 3.0049,
      "theoretical_loss": 4.037117906815491,
      "tokens_seen": 378354688
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044714142427281845,
      "loss": 3.2511,
      "theoretical_loss": 4.037038947449723,
      "tokens_seen": 378420224
    },
    {
      "epoch": 1.02,
      "objective/train/docs_used": 487344,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.918959140777588,
      "objective/train/theoretical_loss": 4.036979739411024,
      "objective/train/tokens_used": 398929376,
      "theoretical_loss": 4.036979739411024,
      "tokens_seen": 378469376
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004471313941825477,
      "loss": 3.0742,
      "theoretical_loss": 4.036960005585286,
      "tokens_seen": 378485760
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004471213640922768,
      "loss": 3.0475,
      "theoretical_loss": 4.036881081215271,
      "tokens_seen": 378551296
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044711133400200605,
      "loss": 2.8166,
      "theoretical_loss": 4.036802174332774,
      "tokens_seen": 378616832
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044710130391173523,
      "loss": 3.1203,
      "theoretical_loss": 4.036723284930894,
      "tokens_seen": 378682368
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004470912738214644,
      "loss": 3.0322,
      "theoretical_loss": 4.036644413002735,
      "tokens_seen": 378747904
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004470812437311936,
      "loss": 3.1208,
      "theoretical_loss": 4.036565558541404,
      "tokens_seen": 378813440
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044707121364092277,
      "loss": 2.895,
      "theoretical_loss": 4.03648672154001,
      "tokens_seen": 378878976
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044706118355065195,
      "loss": 3.1827,
      "theoretical_loss": 4.03640790199167,
      "tokens_seen": 378944512
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004470511534603812,
      "loss": 3.1239,
      "theoretical_loss": 4.036329099889502,
      "tokens_seen": 379010048
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004470411233701103,
      "loss": 3.0161,
      "theoretical_loss": 4.03625031522663,
      "tokens_seen": 379075584
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00044703109327983955,
      "loss": 3.126,
      "theoretical_loss": 4.036171547996179,
      "tokens_seen": 379141120
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0004470210631895687,
      "loss": 3.2521,
      "theoretical_loss": 4.036092798191279,
      "tokens_seen": 379206656
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004470110330992979,
      "loss": 3.1731,
      "theoretical_loss": 4.036014065805067,
      "tokens_seen": 379272192
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004470010030090271,
      "loss": 3.1899,
      "theoretical_loss": 4.035935350830679,
      "tokens_seen": 379337728
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004469909729187563,
      "loss": 3.0432,
      "theoretical_loss": 4.035856653261258,
      "tokens_seen": 379403264
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044698094282848546,
      "loss": 3.1162,
      "theoretical_loss": 4.03577797308995,
      "tokens_seen": 379468800
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044697091273821464,
      "loss": 2.9898,
      "theoretical_loss": 4.035699310309906,
      "tokens_seen": 379534336
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004469608826479438,
      "loss": 3.2695,
      "theoretical_loss": 4.035620664914278,
      "tokens_seen": 379599872
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044695085255767305,
      "loss": 3.017,
      "theoretical_loss": 4.035542036896226,
      "tokens_seen": 379665408
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004469408224674022,
      "loss": 2.9715,
      "theoretical_loss": 4.035463426248908,
      "tokens_seen": 379730944
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004469307923771314,
      "loss": 2.9837,
      "theoretical_loss": 4.0353848329654936,
      "tokens_seen": 379796480
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004469207622868606,
      "loss": 3.0837,
      "theoretical_loss": 4.035306257039149,
      "tokens_seen": 379862016
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004469107321965898,
      "loss": 2.937,
      "theoretical_loss": 4.03522769846305,
      "tokens_seen": 379927552
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044690070210631896,
      "loss": 3.0365,
      "theoretical_loss": 4.0351491572303715,
      "tokens_seen": 379993088
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044689067201604814,
      "loss": 3.1037,
      "theoretical_loss": 4.035070633334295,
      "tokens_seen": 380058624
    },
    {
      "epoch": 1.03,
      "objective/train/docs_used": 488451,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9471192359924316,
      "objective/train/theoretical_loss": 4.035011751785283,
      "objective/train/tokens_used": 400567776,
      "theoretical_loss": 4.035011751785283,
      "tokens_seen": 380107776
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004468806419257773,
      "loss": 3.079,
      "theoretical_loss": 4.034992126768006,
      "tokens_seen": 380124160
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044687061183550656,
      "loss": 2.9451,
      "theoretical_loss": 4.034913637524692,
      "tokens_seen": 380189696
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004468605817452357,
      "loss": 3.2163,
      "theoretical_loss": 4.0348351655975465,
      "tokens_seen": 380255232
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004468505516549649,
      "loss": 2.9821,
      "theoretical_loss": 4.0347567109797655,
      "tokens_seen": 380320768
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044684052156469405,
      "loss": 3.075,
      "theoretical_loss": 4.034678273664548,
      "tokens_seen": 380386304
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004468304914744233,
      "loss": 2.9896,
      "theoretical_loss": 4.0345998536451,
      "tokens_seen": 380451840
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044682046138415246,
      "loss": 2.702,
      "theoretical_loss": 4.034521450914628,
      "tokens_seen": 380517376
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044681043129388164,
      "loss": 2.9938,
      "theoretical_loss": 4.034443065466345,
      "tokens_seen": 380582912
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004468004012036108,
      "loss": 3.0962,
      "theoretical_loss": 4.034364697293466,
      "tokens_seen": 380648448
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044679037111334,
      "loss": 3.1166,
      "theoretical_loss": 4.034286346389209,
      "tokens_seen": 380713984
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004467803410230692,
      "loss": 3.0233,
      "theoretical_loss": 4.034208012746798,
      "tokens_seen": 380779520
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004467703109327984,
      "loss": 3.026,
      "theoretical_loss": 4.03412969635946,
      "tokens_seen": 380845056
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044676028084252755,
      "loss": 2.9949,
      "theoretical_loss": 4.034051397220427,
      "tokens_seen": 380910592
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004467502507522568,
      "loss": 3.1725,
      "theoretical_loss": 4.033973115322932,
      "tokens_seen": 380976128
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044674022066198597,
      "loss": 3.2379,
      "theoretical_loss": 4.033894850660214,
      "tokens_seen": 381041664
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044673019057171515,
      "loss": 3.2136,
      "theoretical_loss": 4.0338166032255165,
      "tokens_seen": 381107200
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044672016048144433,
      "loss": 3.0469,
      "theoretical_loss": 4.033738373012083,
      "tokens_seen": 381172736
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004467101303911735,
      "loss": 3.164,
      "theoretical_loss": 4.0336601600131665,
      "tokens_seen": 381238272
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004467001003009027,
      "loss": 3.0613,
      "theoretical_loss": 4.033581964222018,
      "tokens_seen": 381303808
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004466900702106319,
      "loss": 3.239,
      "theoretical_loss": 4.033503785631897,
      "tokens_seen": 381369344
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004466800401203611,
      "loss": 2.5838,
      "theoretical_loss": 4.033425624236063,
      "tokens_seen": 381434880
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004466700100300903,
      "loss": 3.1374,
      "theoretical_loss": 4.033347480027782,
      "tokens_seen": 381500416
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044665997993981947,
      "loss": 2.9139,
      "theoretical_loss": 4.033269353000324,
      "tokens_seen": 381565952
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044664994984954865,
      "loss": 3.1137,
      "theoretical_loss": 4.03319124314696,
      "tokens_seen": 381631488
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004466399197592779,
      "loss": 2.9559,
      "theoretical_loss": 4.0331131504609665,
      "tokens_seen": 381697024
    },
    {
      "epoch": 1.03,
      "objective/train/docs_used": 489837,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9685873985290527,
      "objective/train/theoretical_loss": 4.033054592208517,
      "objective/train/tokens_used": 402206176,
      "theoretical_loss": 4.033054592208517,
      "tokens_seen": 381746176
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.000446629889669007,
      "loss": 3.1153,
      "theoretical_loss": 4.033035074935626,
      "tokens_seen": 381762560
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044661985957873625,
      "loss": 2.9333,
      "theoretical_loss": 4.03295701656422,
      "tokens_seen": 381828096
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044660982948846543,
      "loss": 2.8328,
      "theoretical_loss": 4.0328789753400365,
      "tokens_seen": 381893632
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004465997993981946,
      "loss": 2.9213,
      "theoretical_loss": 4.032800951256369,
      "tokens_seen": 381959168
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004465897693079238,
      "loss": 2.9061,
      "theoretical_loss": 4.032722944306511,
      "tokens_seen": 382024704
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044657973921765297,
      "loss": 3.0426,
      "theoretical_loss": 4.032644954483763,
      "tokens_seen": 382090240
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044656970912738215,
      "loss": 2.8363,
      "theoretical_loss": 4.032566981781427,
      "tokens_seen": 382155776
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004465596790371114,
      "loss": 3.1766,
      "theoretical_loss": 4.03248902619281,
      "tokens_seen": 382221312
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004465496489468405,
      "loss": 2.9881,
      "theoretical_loss": 4.032411087711223,
      "tokens_seen": 382286848
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044653961885656975,
      "loss": 3.2702,
      "theoretical_loss": 4.032333166329979,
      "tokens_seen": 382352384
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004465295887662989,
      "loss": 2.9304,
      "theoretical_loss": 4.032255262042397,
      "tokens_seen": 382417920
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004465195586760281,
      "loss": 2.8589,
      "theoretical_loss": 4.0321773748417975,
      "tokens_seen": 382483456
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004465095285857573,
      "loss": 2.8999,
      "theoretical_loss": 4.032099504721508,
      "tokens_seen": 382548992
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004464994984954865,
      "loss": 3.0983,
      "theoretical_loss": 4.032021651674857,
      "tokens_seen": 382614528
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044648946840521566,
      "loss": 3.0967,
      "theoretical_loss": 4.0319438156951755,
      "tokens_seen": 382680064
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044647943831494484,
      "loss": 3.0136,
      "theoretical_loss": 4.031865996775803,
      "tokens_seen": 382745600
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.000446469408224674,
      "loss": 3.0933,
      "theoretical_loss": 4.031788194910078,
      "tokens_seen": 382811136
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044645937813440325,
      "loss": 2.915,
      "theoretical_loss": 4.031710410091346,
      "tokens_seen": 382876672
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004464493480441324,
      "loss": 2.9882,
      "theoretical_loss": 4.031632642312955,
      "tokens_seen": 382942208
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004464393179538616,
      "loss": 3.0723,
      "theoretical_loss": 4.031554891568255,
      "tokens_seen": 383007744
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004464292878635908,
      "loss": 3.1333,
      "theoretical_loss": 4.031477157850603,
      "tokens_seen": 383073280
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044641925777332,
      "loss": 3.0266,
      "theoretical_loss": 4.031399441153358,
      "tokens_seen": 383138816
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044640922768304916,
      "loss": 3.0681,
      "theoretical_loss": 4.031321741469881,
      "tokens_seen": 383204352
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044639919759277834,
      "loss": 3.0067,
      "theoretical_loss": 4.031244058793542,
      "tokens_seen": 383269888
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004463891675025075,
      "loss": 2.9739,
      "theoretical_loss": 4.031166393117708,
      "tokens_seen": 383335424
    },
    {
      "epoch": 1.03,
      "objective/train/docs_used": 490505,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.2893757820129395,
      "objective/train/theoretical_loss": 4.031108155013429,
      "objective/train/tokens_used": 403844576,
      "theoretical_loss": 4.031108155013429,
      "tokens_seen": 383384576
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044637913741223676,
      "loss": 3.1077,
      "theoretical_loss": 4.031088744435754,
      "tokens_seen": 383400960
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004463691073219659,
      "loss": 3.155,
      "theoretical_loss": 4.0310111127410595,
      "tokens_seen": 383466496
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004463590772316951,
      "loss": 3.1105,
      "theoretical_loss": 4.030933498027005,
      "tokens_seen": 383532032
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044634904714142425,
      "loss": 3.181,
      "theoretical_loss": 4.030855900286974,
      "tokens_seen": 383597568
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004463390170511535,
      "loss": 2.9759,
      "theoretical_loss": 4.030778319514357,
      "tokens_seen": 383663104
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044632898696088266,
      "loss": 3.1096,
      "theoretical_loss": 4.030700755702545,
      "tokens_seen": 383728640
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044631895687061184,
      "loss": 3.0576,
      "theoretical_loss": 4.030623208844937,
      "tokens_seen": 383794176
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.000446308926780341,
      "loss": 3.1402,
      "theoretical_loss": 4.030545678934931,
      "tokens_seen": 383859712
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004462988966900702,
      "loss": 3.1986,
      "theoretical_loss": 4.030468165965931,
      "tokens_seen": 383925248
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004462888665997994,
      "loss": 3.0959,
      "theoretical_loss": 4.030390669931345,
      "tokens_seen": 383990784
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004462788365095286,
      "loss": 2.9131,
      "theoretical_loss": 4.030313190824582,
      "tokens_seen": 384056320
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044626880641925775,
      "loss": 3.1382,
      "theoretical_loss": 4.030235728639059,
      "tokens_seen": 384121856
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.000446258776328987,
      "loss": 3.0705,
      "theoretical_loss": 4.030158283368193,
      "tokens_seen": 384187392
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044624874623871617,
      "loss": 2.8131,
      "theoretical_loss": 4.030080855005408,
      "tokens_seen": 384252928
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044623871614844535,
      "loss": 3.0487,
      "theoretical_loss": 4.030003443544127,
      "tokens_seen": 384318464
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044622868605817453,
      "loss": 3.2123,
      "theoretical_loss": 4.0299260489777815,
      "tokens_seen": 384384000
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004462186559679037,
      "loss": 3.0109,
      "theoretical_loss": 4.0298486712998045,
      "tokens_seen": 384449536
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004462086258776329,
      "loss": 3.0026,
      "theoretical_loss": 4.029771310503632,
      "tokens_seen": 384515072
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004461985957873621,
      "loss": 3.0486,
      "theoretical_loss": 4.029693966582705,
      "tokens_seen": 384580608
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044618856569709125,
      "loss": 3.036,
      "theoretical_loss": 4.029616639530467,
      "tokens_seen": 384646144
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004461785356068205,
      "loss": 2.7565,
      "theoretical_loss": 4.029539329340366,
      "tokens_seen": 384711680
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004461685055165496,
      "loss": 3.1694,
      "theoretical_loss": 4.029462036005855,
      "tokens_seen": 384777216
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044615847542627885,
      "loss": 2.9726,
      "theoretical_loss": 4.029384759520386,
      "tokens_seen": 384842752
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044614844533600803,
      "loss": 2.9842,
      "theoretical_loss": 4.0293074998774205,
      "tokens_seen": 384908288
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004461384152457372,
      "loss": 2.8077,
      "theoretical_loss": 4.02923025707042,
      "tokens_seen": 384973824
    },
    {
      "epoch": 1.03,
      "objective/train/docs_used": 491392,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.1775639057159424,
      "objective/train/theoretical_loss": 4.029172336009841,
      "objective/train/tokens_used": 405482976,
      "theoretical_loss": 4.029172336009841,
      "tokens_seen": 385022976
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004461283851554664,
      "loss": 2.9078,
      "theoretical_loss": 4.029153031092851,
      "tokens_seen": 385039360
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044611835506519563,
      "loss": 2.9131,
      "theoretical_loss": 4.029075821938182,
      "tokens_seen": 385104896
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044610832497492476,
      "loss": 3.0688,
      "theoretical_loss": 4.028998629599888,
      "tokens_seen": 385170432
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.000446098294884654,
      "loss": 3.0794,
      "theoretical_loss": 4.028921454071445,
      "tokens_seen": 385235968
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004460882647943831,
      "loss": 3.0845,
      "theoretical_loss": 4.0288442953463335,
      "tokens_seen": 385301504
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044607823470411235,
      "loss": 3.1702,
      "theoretical_loss": 4.028767153418038,
      "tokens_seen": 385367040
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044606820461384153,
      "loss": 3.2518,
      "theoretical_loss": 4.028690028280046,
      "tokens_seen": 385432576
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004460581745235707,
      "loss": 2.9465,
      "theoretical_loss": 4.02861291992585,
      "tokens_seen": 385498112
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004460481444332999,
      "loss": 3.3349,
      "theoretical_loss": 4.028535828348945,
      "tokens_seen": 385563648
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004460381143430291,
      "loss": 2.9932,
      "theoretical_loss": 4.02845875354283,
      "tokens_seen": 385629184
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044602808425275826,
      "loss": 3.0814,
      "theoretical_loss": 4.0283816955010066,
      "tokens_seen": 385694720
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004460180541624875,
      "loss": 2.9856,
      "theoretical_loss": 4.0283046542169805,
      "tokens_seen": 385760256
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004460080240722166,
      "loss": 3.2441,
      "theoretical_loss": 4.028227629684263,
      "tokens_seen": 385825792
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044599799398194586,
      "loss": 3.0764,
      "theoretical_loss": 4.028150621896366,
      "tokens_seen": 385891328
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.000445987963891675,
      "loss": 3.0117,
      "theoretical_loss": 4.028073630846807,
      "tokens_seen": 385956864
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004459779338014042,
      "loss": 2.9714,
      "theoretical_loss": 4.0279966565291065,
      "tokens_seen": 386022400
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004459679037111334,
      "loss": 2.8873,
      "theoretical_loss": 4.027919698936788,
      "tokens_seen": 386087936
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004459578736208626,
      "loss": 3.2267,
      "theoretical_loss": 4.02784275806338,
      "tokens_seen": 386153472
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044594784353059176,
      "loss": 3.2652,
      "theoretical_loss": 4.027765833902413,
      "tokens_seen": 386219008
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.000445937813440321,
      "loss": 3.0323,
      "theoretical_loss": 4.027688926447423,
      "tokens_seen": 386284544
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004459277833500502,
      "loss": 3.0881,
      "theoretical_loss": 4.0276120356919485,
      "tokens_seen": 386350080
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044591775325977936,
      "loss": 3.2264,
      "theoretical_loss": 4.02753516162953,
      "tokens_seen": 386415616
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044590772316950854,
      "loss": 3.1626,
      "theoretical_loss": 4.027458304253715,
      "tokens_seen": 386481152
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004458976930792377,
      "loss": 3.2694,
      "theoretical_loss": 4.027381463558052,
      "tokens_seen": 386546688
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044588766298896696,
      "loss": 2.8776,
      "theoretical_loss": 4.027304639536093,
      "tokens_seen": 386612224
    },
    {
      "epoch": 1.03,
      "objective/train/docs_used": 492041,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7662007808685303,
      "objective/train/theoretical_loss": 4.027247032457867,
      "objective/train/tokens_used": 407121376,
      "theoretical_loss": 4.027247032457867,
      "tokens_seen": 386661376
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004458776328986961,
      "loss": 2.9938,
      "theoretical_loss": 4.027227832181397,
      "tokens_seen": 386677760
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004458676028084253,
      "loss": 2.8942,
      "theoretical_loss": 4.027151041487521,
      "tokens_seen": 386743296
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044585757271815445,
      "loss": 3.0691,
      "theoretical_loss": 4.027074267448031,
      "tokens_seen": 386808832
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004458475426278837,
      "loss": 3.1324,
      "theoretical_loss": 4.026997510056493,
      "tokens_seen": 386874368
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044583751253761286,
      "loss": 3.1412,
      "theoretical_loss": 4.026920769306478,
      "tokens_seen": 386939904
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044582748244734204,
      "loss": 2.9207,
      "theoretical_loss": 4.0268440451915595,
      "tokens_seen": 387005440
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004458174523570712,
      "loss": 2.9868,
      "theoretical_loss": 4.0267673377053175,
      "tokens_seen": 387070976
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004458074222668004,
      "loss": 3.0294,
      "theoretical_loss": 4.026690646841332,
      "tokens_seen": 387136512
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004457973921765296,
      "loss": 2.9402,
      "theoretical_loss": 4.026613972593188,
      "tokens_seen": 387202048
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004457873620862588,
      "loss": 3.1123,
      "theoretical_loss": 4.026537314954475,
      "tokens_seen": 387267584
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044577733199598795,
      "loss": 3.0472,
      "theoretical_loss": 4.026460673918785,
      "tokens_seen": 387333120
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004457673019057172,
      "loss": 2.9787,
      "theoretical_loss": 4.026384049479713,
      "tokens_seen": 387398656
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044575727181544637,
      "loss": 2.9231,
      "theoretical_loss": 4.026307441630858,
      "tokens_seen": 387464192
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044574724172517555,
      "loss": 2.9785,
      "theoretical_loss": 4.026230850365824,
      "tokens_seen": 387529728
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044573721163490473,
      "loss": 3.0627,
      "theoretical_loss": 4.026154275678216,
      "tokens_seen": 387595264
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004457271815446339,
      "loss": 3.1657,
      "theoretical_loss": 4.0260777175616465,
      "tokens_seen": 387660800
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004457171514543631,
      "loss": 2.9694,
      "theoretical_loss": 4.026001176009726,
      "tokens_seen": 387726336
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004457071213640923,
      "loss": 3.1933,
      "theoretical_loss": 4.025924651016073,
      "tokens_seen": 387791872
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044569709127382145,
      "loss": 2.8754,
      "theoretical_loss": 4.025848142574307,
      "tokens_seen": 387857408
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004456870611835507,
      "loss": 3.1072,
      "theoretical_loss": 4.025771650678053,
      "tokens_seen": 387922944
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004456770310932798,
      "loss": 3.3678,
      "theoretical_loss": 4.025695175320939,
      "tokens_seen": 387988480
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044566700100300905,
      "loss": 3.0147,
      "theoretical_loss": 4.025618716496595,
      "tokens_seen": 388054016
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044565697091273823,
      "loss": 3.3046,
      "theoretical_loss": 4.025542274198656,
      "tokens_seen": 388119552
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004456469408224674,
      "loss": 3.2766,
      "theoretical_loss": 4.025465848420762,
      "tokens_seen": 388185088
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004456369107321966,
      "loss": 3.0493,
      "theoretical_loss": 4.025389439156552,
      "tokens_seen": 388250624
    },
    {
      "epoch": 1.03,
      "objective/train/docs_used": 492346,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.0495290756225586,
      "objective/train/theoretical_loss": 4.025332143041677,
      "objective/train/tokens_used": 408759776,
      "theoretical_loss": 4.025332143041677,
      "tokens_seen": 388299776
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044562688064192583,
      "loss": 3.0685,
      "theoretical_loss": 4.025313046399672,
      "tokens_seen": 388316160
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044561685055165496,
      "loss": 3.2269,
      "theoretical_loss": 4.025236670143771,
      "tokens_seen": 388381696
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004456068204613842,
      "loss": 3.2541,
      "theoretical_loss": 4.025160310382502,
      "tokens_seen": 388447232
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004455967903711133,
      "loss": 2.9198,
      "theoretical_loss": 4.02508396710952,
      "tokens_seen": 388512768
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044558676028084255,
      "loss": 2.9715,
      "theoretical_loss": 4.025007640318485,
      "tokens_seen": 388578304
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044557673019057173,
      "loss": 3.109,
      "theoretical_loss": 4.024931330003058,
      "tokens_seen": 388643840
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004455667001003009,
      "loss": 3.1998,
      "theoretical_loss": 4.024855036156907,
      "tokens_seen": 388709376
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004455566700100301,
      "loss": 3.1695,
      "theoretical_loss": 4.024778758773702,
      "tokens_seen": 388774912
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004455466399197593,
      "loss": 2.9735,
      "theoretical_loss": 4.024702497847115,
      "tokens_seen": 388840448
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044553660982948846,
      "loss": 3.1217,
      "theoretical_loss": 4.0246262533708235,
      "tokens_seen": 388905984
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004455265797392177,
      "loss": 3.1701,
      "theoretical_loss": 4.024550025338508,
      "tokens_seen": 388971520
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004455165496489468,
      "loss": 3.2158,
      "theoretical_loss": 4.024473813743851,
      "tokens_seen": 389037056
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044550651955867606,
      "loss": 3.3209,
      "theoretical_loss": 4.024397618580542,
      "tokens_seen": 389102592
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004454964894684052,
      "loss": 3.1834,
      "theoretical_loss": 4.02432143984227,
      "tokens_seen": 389168128
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004454864593781344,
      "loss": 2.9146,
      "theoretical_loss": 4.024245277522729,
      "tokens_seen": 389233664
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004454764292878636,
      "loss": 3.3165,
      "theoretical_loss": 4.0241691316156185,
      "tokens_seen": 389299200
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004454663991975928,
      "loss": 3.4463,
      "theoretical_loss": 4.024093002114638,
      "tokens_seen": 389364736
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044545636910732196,
      "loss": 3.1539,
      "theoretical_loss": 4.0240168890134935,
      "tokens_seen": 389430272
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004454463390170512,
      "loss": 3.547,
      "theoretical_loss": 4.0239407923058925,
      "tokens_seen": 389495808
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004454363089267803,
      "loss": 3.1451,
      "theoretical_loss": 4.023864711985547,
      "tokens_seen": 389561344
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044542627883650956,
      "loss": 3.1436,
      "theoretical_loss": 4.023788648046171,
      "tokens_seen": 389626880
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004454162487462387,
      "loss": 2.9621,
      "theoretical_loss": 4.023712600481484,
      "tokens_seen": 389692416
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004454062186559679,
      "loss": 3.042,
      "theoretical_loss": 4.023636569285208,
      "tokens_seen": 389757952
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004453961885656971,
      "loss": 3.1537,
      "theoretical_loss": 4.023560554451068,
      "tokens_seen": 389823488
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004453861584754263,
      "loss": 3.1769,
      "theoretical_loss": 4.023484555972794,
      "tokens_seen": 389889024
    },
    {
      "epoch": 1.03,
      "objective/train/docs_used": 492346,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.5270936489105225,
      "objective/train/theoretical_loss": 4.0234275678438545,
      "objective/train/tokens_used": 410398176,
      "theoretical_loss": 4.0234275678438545,
      "tokens_seen": 389938176
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044537612838515547,
      "loss": 3.1988,
      "theoretical_loss": 4.023408573844117,
      "tokens_seen": 389954560
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044536609829488465,
      "loss": 2.8549,
      "theoretical_loss": 4.023332608058775,
      "tokens_seen": 390020096
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044535606820461383,
      "loss": 3.1868,
      "theoretical_loss": 4.023256658610505,
      "tokens_seen": 390085632
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044534603811434306,
      "loss": 3.2635,
      "theoretical_loss": 4.02318072549305,
      "tokens_seen": 390151168
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004453360080240722,
      "loss": 3.1127,
      "theoretical_loss": 4.0231048087001575,
      "tokens_seen": 390216704
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004453259779338014,
      "loss": 3.2178,
      "theoretical_loss": 4.023028908225577,
      "tokens_seen": 390282240
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044531594784353055,
      "loss": 3.1681,
      "theoretical_loss": 4.02295302406306,
      "tokens_seen": 390347776
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004453059177532598,
      "loss": 3.1644,
      "theoretical_loss": 4.0228771562063645,
      "tokens_seen": 390413312
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044529588766298897,
      "loss": 3.3908,
      "theoretical_loss": 4.022801304649251,
      "tokens_seen": 390478848
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044528585757271815,
      "loss": 3.2459,
      "theoretical_loss": 4.022725469385481,
      "tokens_seen": 390544384
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044527582748244733,
      "loss": 2.9103,
      "theoretical_loss": 4.022649650408823,
      "tokens_seen": 390609920
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044526579739217657,
      "loss": 3.2225,
      "theoretical_loss": 4.022573847713048,
      "tokens_seen": 390675456
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004452557673019057,
      "loss": 2.9859,
      "theoretical_loss": 4.022498061291927,
      "tokens_seen": 390740992
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044524573721163493,
      "loss": 3.1762,
      "theoretical_loss": 4.022422291139238,
      "tokens_seen": 390806528
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044523570712136406,
      "loss": 3.3121,
      "theoretical_loss": 4.022346537248763,
      "tokens_seen": 390872064
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004452256770310933,
      "loss": 3.2159,
      "theoretical_loss": 4.022270799614285,
      "tokens_seen": 390937600
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044521564694082247,
      "loss": 3.1719,
      "theoretical_loss": 4.02219507822959,
      "tokens_seen": 391003136
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044520561685055165,
      "loss": 3.325,
      "theoretical_loss": 4.022119373088472,
      "tokens_seen": 391068672
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044519558676028083,
      "loss": 3.2354,
      "theoretical_loss": 4.022043684184723,
      "tokens_seen": 391134208
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044518555667001,
      "loss": 3.0924,
      "theoretical_loss": 4.021968011512141,
      "tokens_seen": 391199744
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044517552657973925,
      "loss": 3.1403,
      "theoretical_loss": 4.0218923550645265,
      "tokens_seen": 391265280
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044516549648946843,
      "loss": 3.2618,
      "theoretical_loss": 4.021816714835685,
      "tokens_seen": 391330816
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004451554663991976,
      "loss": 3.1097,
      "theoretical_loss": 4.0217410908194235,
      "tokens_seen": 391396352
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004451454363089268,
      "loss": 3.0826,
      "theoretical_loss": 4.021665483009555,
      "tokens_seen": 391461888
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044513540621865603,
      "loss": 3.325,
      "theoretical_loss": 4.021589891399891,
      "tokens_seen": 391527424
    },
    {
      "epoch": 1.03,
      "objective/train/docs_used": 493102,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.2480216026306152,
      "objective/train/theoretical_loss": 4.02153320832031,
      "objective/train/tokens_used": 412036576,
      "theoretical_loss": 4.02153320832031,
      "tokens_seen": 391576576
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044512537612838516,
      "loss": 3.1819,
      "theoretical_loss": 4.021514315984252,
      "tokens_seen": 391592960
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004451153460381144,
      "loss": 3.0459,
      "theoretical_loss": 4.021438756756459,
      "tokens_seen": 391658496
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004451053159478435,
      "loss": 3.2661,
      "theoretical_loss": 4.021363213710337,
      "tokens_seen": 391724032
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044509528585757275,
      "loss": 3.1196,
      "theoretical_loss": 4.021287686839712,
      "tokens_seen": 391789568
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044508525576730194,
      "loss": 3.0289,
      "theoretical_loss": 4.021212176138419,
      "tokens_seen": 391855104
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004450752256770311,
      "loss": 3.4092,
      "theoretical_loss": 4.02113668160029,
      "tokens_seen": 391920640
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004450651955867603,
      "loss": 3.0095,
      "theoretical_loss": 4.0210612032191655,
      "tokens_seen": 391986176
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004450551654964895,
      "loss": 3.417,
      "theoretical_loss": 4.020985740988887,
      "tokens_seen": 392051712
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044504513540621866,
      "loss": 3.3539,
      "theoretical_loss": 4.020910294903298,
      "tokens_seen": 392117248
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004450351053159479,
      "loss": 3.3295,
      "theoretical_loss": 4.020834864956248,
      "tokens_seen": 392182784
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.000445025075225677,
      "loss": 3.4475,
      "theoretical_loss": 4.020759451141589,
      "tokens_seen": 392248320
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044501504513540626,
      "loss": 3.405,
      "theoretical_loss": 4.020684053453176,
      "tokens_seen": 392313856
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004450050150451354,
      "loss": 3.3016,
      "theoretical_loss": 4.020608671884868,
      "tokens_seen": 392379392
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004449949849548646,
      "loss": 3.223,
      "theoretical_loss": 4.020533306430527,
      "tokens_seen": 392444928
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004449849548645938,
      "loss": 3.0813,
      "theoretical_loss": 4.020457957084017,
      "tokens_seen": 392510464
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.000444974924774323,
      "loss": 3.2733,
      "theoretical_loss": 4.0203826238392075,
      "tokens_seen": 392576000
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044496489468405216,
      "loss": 3.4012,
      "theoretical_loss": 4.020307306689972,
      "tokens_seen": 392641536
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004449548645937814,
      "loss": 2.9446,
      "theoretical_loss": 4.020232005630183,
      "tokens_seen": 392707072
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004449448345035105,
      "loss": 2.8994,
      "theoretical_loss": 4.020156720653722,
      "tokens_seen": 392772608
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044493480441323976,
      "loss": 3.1714,
      "theoretical_loss": 4.020081451754469,
      "tokens_seen": 392838144
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004449247743229689,
      "loss": 3.33,
      "theoretical_loss": 4.0200061989263105,
      "tokens_seen": 392903680
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004449147442326981,
      "loss": 3.2939,
      "theoretical_loss": 4.019930962163135,
      "tokens_seen": 392969216
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004449047141424273,
      "loss": 3.3044,
      "theoretical_loss": 4.019855741458835,
      "tokens_seen": 393034752
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004448946840521565,
      "loss": 3.1703,
      "theoretical_loss": 4.019780536807305,
      "tokens_seen": 393100288
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044488465396188567,
      "loss": 3.1619,
      "theoretical_loss": 4.019705348202446,
      "tokens_seen": 393165824
    },
    {
      "epoch": 1.03,
      "objective/train/docs_used": 494542,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9207966327667236,
      "objective/train/theoretical_loss": 4.019648967275759,
      "objective/train/tokens_used": 413674976,
      "theoretical_loss": 4.019648967275759,
      "tokens_seen": 393214976
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044487462387161485,
      "loss": 3.1161,
      "theoretical_loss": 4.019630175638158,
      "tokens_seen": 393231360
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044486459378134403,
      "loss": 3.1258,
      "theoretical_loss": 4.0195550191083464,
      "tokens_seen": 393296896
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044485456369107326,
      "loss": 3.12,
      "theoretical_loss": 4.019479878606921,
      "tokens_seen": 393362432
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004448445336008024,
      "loss": 3.2797,
      "theoretical_loss": 4.019404754127793,
      "tokens_seen": 393427968
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004448345035105316,
      "loss": 3.1239,
      "theoretical_loss": 4.01932964566488,
      "tokens_seen": 393493504
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044482447342026075,
      "loss": 3.118,
      "theoretical_loss": 4.0192545532120985,
      "tokens_seen": 393559040
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044481444332999,
      "loss": 3.1753,
      "theoretical_loss": 4.01917947676337,
      "tokens_seen": 393624576
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044480441323971917,
      "loss": 3.1646,
      "theoretical_loss": 4.019104416312623,
      "tokens_seen": 393690112
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044479438314944835,
      "loss": 3.0934,
      "theoretical_loss": 4.019029371853784,
      "tokens_seen": 393755648
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044478435305917753,
      "loss": 3.1208,
      "theoretical_loss": 4.018954343380786,
      "tokens_seen": 393821184
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044477432296890677,
      "loss": 3.0607,
      "theoretical_loss": 4.018879330887563,
      "tokens_seen": 393886720
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004447642928786359,
      "loss": 3.2782,
      "theoretical_loss": 4.018804334368055,
      "tokens_seen": 393952256
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044475426278836513,
      "loss": 3.16,
      "theoretical_loss": 4.018729353816203,
      "tokens_seen": 394017792
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044474423269809426,
      "loss": 3.2467,
      "theoretical_loss": 4.018654389225954,
      "tokens_seen": 394083328
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004447342026078235,
      "loss": 3.3955,
      "theoretical_loss": 4.018579440591255,
      "tokens_seen": 394148864
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044472417251755267,
      "loss": 3.2859,
      "theoretical_loss": 4.018504507906059,
      "tokens_seen": 394214400
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044471414242728185,
      "loss": 3.3852,
      "theoretical_loss": 4.0184295911643195,
      "tokens_seen": 394279936
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044470411233701103,
      "loss": 3.437,
      "theoretical_loss": 4.018354690359996,
      "tokens_seen": 394345472
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004446940822467402,
      "loss": 3.277,
      "theoretical_loss": 4.01827980548705,
      "tokens_seen": 394411008
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004446840521564694,
      "loss": 3.3371,
      "theoretical_loss": 4.018204936539448,
      "tokens_seen": 394476544
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044467402206619863,
      "loss": 3.2478,
      "theoretical_loss": 4.018130083511156,
      "tokens_seen": 394542080
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044466399197592776,
      "loss": 3.2044,
      "theoretical_loss": 4.018055246396147,
      "tokens_seen": 394607616
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.000444653961885657,
      "loss": 3.2246,
      "theoretical_loss": 4.017980425188396,
      "tokens_seen": 394673152
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004446439317953861,
      "loss": 3.0988,
      "theoretical_loss": 4.017905619881881,
      "tokens_seen": 394738688
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044463390170511536,
      "loss": 3.0712,
      "theoretical_loss": 4.017830830470583,
      "tokens_seen": 394804224
    },
    {
      "epoch": 1.03,
      "objective/train/docs_used": 495175,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.2577452659606934,
      "objective/train/theoretical_loss": 4.017774748839727,
      "objective/train/tokens_used": 415313376,
      "theoretical_loss": 4.017774748839727,
      "tokens_seen": 394853376
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044462387161484454,
      "loss": 3.1142,
      "theoretical_loss": 4.017756056948487,
      "tokens_seen": 394869760
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004446138415245737,
      "loss": 3.3469,
      "theoretical_loss": 4.017681299309582,
      "tokens_seen": 394935296
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004446038114343029,
      "loss": 3.1336,
      "theoretical_loss": 4.017606557547858,
      "tokens_seen": 395000832
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044459378134403214,
      "loss": 2.8472,
      "theoretical_loss": 4.01753183165731,
      "tokens_seen": 395066368
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044458375125376126,
      "loss": 3.1497,
      "theoretical_loss": 4.017457121631936,
      "tokens_seen": 395131904
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004445737211634905,
      "loss": 3.2117,
      "theoretical_loss": 4.017382427465737,
      "tokens_seen": 395197440
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004445636910732196,
      "loss": 2.946,
      "theoretical_loss": 4.017307749152717,
      "tokens_seen": 395262976
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044455366098294886,
      "loss": 3.2121,
      "theoretical_loss": 4.017233086686884,
      "tokens_seen": 395328512
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044454363089267804,
      "loss": 2.9614,
      "theoretical_loss": 4.017158440062249,
      "tokens_seen": 395394048
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004445336008024072,
      "loss": 3.1452,
      "theoretical_loss": 4.017083809272826,
      "tokens_seen": 395459584
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004445235707121364,
      "loss": 2.9447,
      "theoretical_loss": 4.017009194312633,
      "tokens_seen": 395525120
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004445135406218656,
      "loss": 2.99,
      "theoretical_loss": 4.016934595175689,
      "tokens_seen": 395590656
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044450351053159477,
      "loss": 3.0288,
      "theoretical_loss": 4.01686001185602,
      "tokens_seen": 395656192
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.000444493480441324,
      "loss": 3.3945,
      "theoretical_loss": 4.016785444347652,
      "tokens_seen": 395721728
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044448345035105313,
      "loss": 3.2868,
      "theoretical_loss": 4.016710892644614,
      "tokens_seen": 395787264
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044447342026078236,
      "loss": 3.281,
      "theoretical_loss": 4.016636356740943,
      "tokens_seen": 395852800
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004444633901705115,
      "loss": 3.1532,
      "theoretical_loss": 4.016561836630672,
      "tokens_seen": 395918336
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004444533600802407,
      "loss": 3.2487,
      "theoretical_loss": 4.016487332307844,
      "tokens_seen": 395983872
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044444332998996996,
      "loss": 3.1453,
      "theoretical_loss": 4.016412843766502,
      "tokens_seen": 396049408
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004444332998996991,
      "loss": 3.1682,
      "theoretical_loss": 4.016338371000691,
      "tokens_seen": 396114944
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004444232698094283,
      "loss": 3.1292,
      "theoretical_loss": 4.016263914004462,
      "tokens_seen": 396180480
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004444132397191575,
      "loss": 3.4037,
      "theoretical_loss": 4.016189472771868,
      "tokens_seen": 396246016
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004444032096288867,
      "loss": 3.0772,
      "theoretical_loss": 4.016115047296965,
      "tokens_seen": 396311552
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044439317953861587,
      "loss": 3.2491,
      "theoretical_loss": 4.016040637573812,
      "tokens_seen": 396377088
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044438314944834505,
      "loss": 3.2143,
      "theoretical_loss": 4.0159662435964725,
      "tokens_seen": 396442624
    },
    {
      "epoch": 1.03,
      "objective/train/docs_used": 496403,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9757843017578125,
      "objective/train/theoretical_loss": 4.015910458443088,
      "objective/train/tokens_used": 416951776,
      "theoretical_loss": 4.015910458443088,
      "tokens_seen": 396491776
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044437311935807423,
      "loss": 2.9239,
      "theoretical_loss": 4.015891865359012,
      "tokens_seen": 396508160
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044436308926780346,
      "loss": 2.9201,
      "theoretical_loss": 4.0158175028555,
      "tokens_seen": 396573696
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004443530591775326,
      "loss": 3.2929,
      "theoretical_loss": 4.015743156080008,
      "tokens_seen": 396639232
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004443430290872618,
      "loss": 2.9639,
      "theoretical_loss": 4.015668825026613,
      "tokens_seen": 396704768
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044433299899699095,
      "loss": 3.1691,
      "theoretical_loss": 4.015594509689393,
      "tokens_seen": 396770304
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004443229689067202,
      "loss": 3.0509,
      "theoretical_loss": 4.015520210062429,
      "tokens_seen": 396835840
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044431293881644937,
      "loss": 3.3591,
      "theoretical_loss": 4.015445926139808,
      "tokens_seen": 396901376
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044430290872617855,
      "loss": 3.1861,
      "theoretical_loss": 4.015371657915617,
      "tokens_seen": 396966912
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044429287863590773,
      "loss": 3.2149,
      "theoretical_loss": 4.015297405383948,
      "tokens_seen": 397032448
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044428284854563697,
      "loss": 3.2117,
      "theoretical_loss": 4.015223168538896,
      "tokens_seen": 397097984
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004442728184553661,
      "loss": 3.1278,
      "theoretical_loss": 4.015148947374559,
      "tokens_seen": 397163520
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044426278836509533,
      "loss": 3.1822,
      "theoretical_loss": 4.015074741885038,
      "tokens_seen": 397229056
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044425275827482446,
      "loss": 3.0979,
      "theoretical_loss": 4.015000552064437,
      "tokens_seen": 397294592
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004442427281845537,
      "loss": 3.0718,
      "theoretical_loss": 4.014926377906864,
      "tokens_seen": 397360128
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044423269809428287,
      "loss": 3.0584,
      "theoretical_loss": 4.01485221940643,
      "tokens_seen": 397425664
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044422266800401205,
      "loss": 3.2676,
      "theoretical_loss": 4.014778076557249,
      "tokens_seen": 397491200
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044421263791374123,
      "loss": 3.1287,
      "theoretical_loss": 4.0147039493534376,
      "tokens_seen": 397556736
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004442026078234704,
      "loss": 3.184,
      "theoretical_loss": 4.014629837789116,
      "tokens_seen": 397622272
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004441925777331996,
      "loss": 3.3226,
      "theoretical_loss": 4.014555741858407,
      "tokens_seen": 397687808
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044418254764292883,
      "loss": 3.1903,
      "theoretical_loss": 4.014481661555439,
      "tokens_seen": 397753344
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044417251755265796,
      "loss": 3.1214,
      "theoretical_loss": 4.0144075968743405,
      "tokens_seen": 397818880
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004441624874623872,
      "loss": 3.2588,
      "theoretical_loss": 4.014333547809246,
      "tokens_seen": 397884416
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004441524573721163,
      "loss": 3.0024,
      "theoretical_loss": 4.01425951435429,
      "tokens_seen": 397949952
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044414242728184556,
      "loss": 3.0515,
      "theoretical_loss": 4.014185496503612,
      "tokens_seen": 398015488
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044413239719157474,
      "loss": 3.0877,
      "theoretical_loss": 4.014111494251355,
      "tokens_seen": 398081024
    },
    {
      "epoch": 1.03,
      "objective/train/docs_used": 497345,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9470760822296143,
      "objective/train/theoretical_loss": 4.014056002795105,
      "objective/train/tokens_used": 418590176,
      "theoretical_loss": 4.014056002795105,
      "tokens_seen": 398130176
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004441223671013039,
      "loss": 3.09,
      "theoretical_loss": 4.014037507591665,
      "tokens_seen": 398146560
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004441123370110331,
      "loss": 3.3322,
      "theoretical_loss": 4.01396353651869,
      "tokens_seen": 398212096
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044410230692076234,
      "loss": 3.3269,
      "theoretical_loss": 4.013889581026583,
      "tokens_seen": 398277632
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044409227683049146,
      "loss": 3.1256,
      "theoretical_loss": 4.013815641109498,
      "tokens_seen": 398343168
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004440822467402207,
      "loss": 2.9061,
      "theoretical_loss": 4.0137417167615945,
      "tokens_seen": 398408704
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004440722166499498,
      "loss": 2.7541,
      "theoretical_loss": 4.013667807977033,
      "tokens_seen": 398474240
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044406218655967906,
      "loss": 2.8525,
      "theoretical_loss": 4.013593914749977,
      "tokens_seen": 398539776
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044405215646940824,
      "loss": 3.2677,
      "theoretical_loss": 4.013520037074597,
      "tokens_seen": 398605312
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004440421263791374,
      "loss": 3.069,
      "theoretical_loss": 4.013446174945062,
      "tokens_seen": 398670848
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004440320962888666,
      "loss": 3.0376,
      "theoretical_loss": 4.013372328355547,
      "tokens_seen": 398736384
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004440220661985958,
      "loss": 3.0581,
      "theoretical_loss": 4.0132984973002275,
      "tokens_seen": 398801920
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044401203610832497,
      "loss": 3.1568,
      "theoretical_loss": 4.013224681773286,
      "tokens_seen": 398867456
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004440020060180542,
      "loss": 3.2621,
      "theoretical_loss": 4.0131508817689046,
      "tokens_seen": 398932992
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044399197592778333,
      "loss": 3.0157,
      "theoretical_loss": 4.013077097281269,
      "tokens_seen": 398998528
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044398194583751256,
      "loss": 3.3744,
      "theoretical_loss": 4.013003328304571,
      "tokens_seen": 399064064
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004439719157472417,
      "loss": 3.2167,
      "theoretical_loss": 4.012929574833002,
      "tokens_seen": 399129600
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004439618856569709,
      "loss": 3.2892,
      "theoretical_loss": 4.012855836860759,
      "tokens_seen": 399195136
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004439518555667001,
      "loss": 3.3375,
      "theoretical_loss": 4.01278211438204,
      "tokens_seen": 399260672
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004439418254764293,
      "loss": 3.1236,
      "theoretical_loss": 4.012708407391049,
      "tokens_seen": 399326208
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044393179538615847,
      "loss": 2.9568,
      "theoretical_loss": 4.012634715881989,
      "tokens_seen": 399391744
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004439217652958877,
      "loss": 3.2874,
      "theoretical_loss": 4.012561039849069,
      "tokens_seen": 399457280
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044391173520561683,
      "loss": 3.1477,
      "theoretical_loss": 4.012487379286502,
      "tokens_seen": 399522816
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044390170511534607,
      "loss": 2.9509,
      "theoretical_loss": 4.012413734188501,
      "tokens_seen": 399588352
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004438916750250752,
      "loss": 3.1483,
      "theoretical_loss": 4.012340104549285,
      "tokens_seen": 399653888
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044388164493480443,
      "loss": 3.11,
      "theoretical_loss": 4.012266490363073,
      "tokens_seen": 399719424
    },
    {
      "epoch": 1.03,
      "objective/train/docs_used": 502454,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.3102781772613525,
      "objective/train/theoretical_loss": 4.012211289860975,
      "objective/train/tokens_used": 420228576,
      "theoretical_loss": 4.012211289860975,
      "tokens_seen": 399768576
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004438716148445336,
      "loss": 3.2869,
      "theoretical_loss": 4.012192891624092,
      "tokens_seen": 399784960
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004438615847542628,
      "loss": 2.9235,
      "theoretical_loss": 4.012119308326567,
      "tokens_seen": 399850496
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044385155466399197,
      "loss": 3.0863,
      "theoretical_loss": 4.012045740464728,
      "tokens_seen": 399916032
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044384152457372115,
      "loss": 3.0504,
      "theoretical_loss": 4.01197218803281,
      "tokens_seen": 399981568
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044383149448345033,
      "loss": 3.1842,
      "theoretical_loss": 4.011898651025048,
      "tokens_seen": 400047104
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044382146439317957,
      "loss": 3.1272,
      "theoretical_loss": 4.011825129435683,
      "tokens_seen": 400112640
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004438114343029087,
      "loss": 3.1282,
      "theoretical_loss": 4.011751623258955,
      "tokens_seen": 400178176
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044380140421263793,
      "loss": 3.0844,
      "theoretical_loss": 4.011678132489112,
      "tokens_seen": 400243712
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004437913741223671,
      "loss": 2.8656,
      "theoretical_loss": 4.011604657120401,
      "tokens_seen": 400309248
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004437813440320963,
      "loss": 3.2295,
      "theoretical_loss": 4.011531197147076,
      "tokens_seen": 400374784
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004437713139418255,
      "loss": 3.0765,
      "theoretical_loss": 4.011457752563391,
      "tokens_seen": 400440320
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044376128385155466,
      "loss": 3.2233,
      "theoretical_loss": 4.011384323363603,
      "tokens_seen": 400505856
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044375125376128384,
      "loss": 3.3812,
      "theoretical_loss": 4.011310909541975,
      "tokens_seen": 400571392
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044374122367101307,
      "loss": 3.0616,
      "theoretical_loss": 4.011237511092769,
      "tokens_seen": 400636928
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004437311935807422,
      "loss": 3.1718,
      "theoretical_loss": 4.011164128010255,
      "tokens_seen": 400702464
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044372116349047143,
      "loss": 3.0474,
      "theoretical_loss": 4.011090760288701,
      "tokens_seen": 400768000
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044371113340020056,
      "loss": 3.0441,
      "theoretical_loss": 4.011017407922381,
      "tokens_seen": 400833536
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004437011033099298,
      "loss": 3.1871,
      "theoretical_loss": 4.010944070905572,
      "tokens_seen": 400899072
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044369107321965903,
      "loss": 3.3112,
      "theoretical_loss": 4.010870749232554,
      "tokens_seen": 400964608
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044368104312938816,
      "loss": 2.9837,
      "theoretical_loss": 4.0107974428976085,
      "tokens_seen": 401030144
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004436710130391174,
      "loss": 3.4095,
      "theoretical_loss": 4.010724151895022,
      "tokens_seen": 401095680
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004436609829488465,
      "loss": 3.2873,
      "theoretical_loss": 4.010650876219083,
      "tokens_seen": 401161216
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044365095285857576,
      "loss": 3.1268,
      "theoretical_loss": 4.010577615864083,
      "tokens_seen": 401226752
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044364092276830494,
      "loss": 3.0925,
      "theoretical_loss": 4.010504370824316,
      "tokens_seen": 401292288
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004436308926780341,
      "loss": 3.0146,
      "theoretical_loss": 4.010431141094083,
      "tokens_seen": 401357824
    },
    {
      "epoch": 1.03,
      "objective/train/docs_used": 507639,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.0078208446502686,
      "objective/train/theoretical_loss": 4.01037622883986,
      "objective/train/tokens_used": 421866976,
      "theoretical_loss": 4.01037622883986,
      "tokens_seen": 401406976
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004436208625877633,
      "loss": 2.9291,
      "theoretical_loss": 4.010357926667682,
      "tokens_seen": 401423360
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044361083249749254,
      "loss": 3.3317,
      "theoretical_loss": 4.0102847275394184,
      "tokens_seen": 401488896
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044360080240722166,
      "loss": 2.7376,
      "theoretical_loss": 4.0102115437035994,
      "tokens_seen": 401554432
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004435907723169509,
      "loss": 3.2031,
      "theoretical_loss": 4.0101383751545345,
      "tokens_seen": 401619968
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044358074222668,
      "loss": 3.1482,
      "theoretical_loss": 4.010065221886537,
      "tokens_seen": 401685504
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044357071213640926,
      "loss": 3.0891,
      "theoretical_loss": 4.0099920838939225,
      "tokens_seen": 401751040
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044356068204613844,
      "loss": 3.0714,
      "theoretical_loss": 4.009918961171012,
      "tokens_seen": 401816576
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004435506519558676,
      "loss": 3.1246,
      "theoretical_loss": 4.009845853712126,
      "tokens_seen": 401882112
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004435406218655968,
      "loss": 3.2701,
      "theoretical_loss": 4.0097727615115915,
      "tokens_seen": 401947648
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.000443530591775326,
      "loss": 3.082,
      "theoretical_loss": 4.009699684563736,
      "tokens_seen": 402013184
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044352056168505517,
      "loss": 3.2898,
      "theoretical_loss": 4.00962662286289,
      "tokens_seen": 402078720
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004435105315947844,
      "loss": 3.0541,
      "theoretical_loss": 4.00955357640339,
      "tokens_seen": 402144256
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044350050150451353,
      "loss": 2.9924,
      "theoretical_loss": 4.009480545179572,
      "tokens_seen": 402209792
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044349047141424276,
      "loss": 2.8212,
      "theoretical_loss": 4.0094075291857765,
      "tokens_seen": 402275328
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004434804413239719,
      "loss": 2.8057,
      "theoretical_loss": 4.009334528416347,
      "tokens_seen": 402340864
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004434704112337011,
      "loss": 3.1605,
      "theoretical_loss": 4.009261542865631,
      "tokens_seen": 402406400
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004434603811434303,
      "loss": 3.3208,
      "theoretical_loss": 4.009188572527977,
      "tokens_seen": 402471936
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004434503510531595,
      "loss": 3.2758,
      "theoretical_loss": 4.009115617397739,
      "tokens_seen": 402537472
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044344032096288867,
      "loss": 3.1012,
      "theoretical_loss": 4.009042677469271,
      "tokens_seen": 402603008
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004434302908726179,
      "loss": 3.0349,
      "theoretical_loss": 4.008969752736932,
      "tokens_seen": 402668544
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044342026078234703,
      "loss": 3.3604,
      "theoretical_loss": 4.0088968431950835,
      "tokens_seen": 402734080
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044341023069207627,
      "loss": 3.0932,
      "theoretical_loss": 4.008823948838091,
      "tokens_seen": 402799616
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004434002006018054,
      "loss": 3.0842,
      "theoretical_loss": 4.008751069660322,
      "tokens_seen": 402865152
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044339017051153463,
      "loss": 3.1979,
      "theoretical_loss": 4.008678205656146,
      "tokens_seen": 402930688
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004433801404212638,
      "loss": 3.2418,
      "theoretical_loss": 4.008605356819937,
      "tokens_seen": 402996224
    },
    {
      "epoch": 1.03,
      "objective/train/docs_used": 512662,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.06799578666687,
      "objective/train/theoretical_loss": 4.0085507301433765,
      "objective/train/tokens_used": 423505376,
      "theoretical_loss": 4.0085507301433765,
      "tokens_seen": 403045376
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.000443370110330993,
      "loss": 3.2713,
      "theoretical_loss": 4.008532523146073,
      "tokens_seen": 403061760
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044336008024072217,
      "loss": 3.0465,
      "theoretical_loss": 4.008459704628931,
      "tokens_seen": 403127296
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044335005015045135,
      "loss": 3.1517,
      "theoretical_loss": 4.008386901262897,
      "tokens_seen": 403192832
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044334002006018053,
      "loss": 3.1538,
      "theoretical_loss": 4.0083141130423545,
      "tokens_seen": 403258368
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044332998996990977,
      "loss": 2.9468,
      "theoretical_loss": 4.008241339961692,
      "tokens_seen": 403323904
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004433199598796389,
      "loss": 3.0336,
      "theoretical_loss": 4.008168582015301,
      "tokens_seen": 403389440
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044330992978936813,
      "loss": 3.1535,
      "theoretical_loss": 4.008095839197578,
      "tokens_seen": 403454976
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004432998996990973,
      "loss": 3.3492,
      "theoretical_loss": 4.008023111502919,
      "tokens_seen": 403520512
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004432898696088265,
      "loss": 3.0574,
      "theoretical_loss": 4.007950398925725,
      "tokens_seen": 403586048
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004432798395185557,
      "loss": 3.1388,
      "theoretical_loss": 4.007877701460398,
      "tokens_seen": 403651584
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044326980942828486,
      "loss": 3.1729,
      "theoretical_loss": 4.0078050191013475,
      "tokens_seen": 403717120
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044325977933801404,
      "loss": 2.9521,
      "theoretical_loss": 4.007732351842981,
      "tokens_seen": 403782656
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044324974924774327,
      "loss": 3.2514,
      "theoretical_loss": 4.007659699679711,
      "tokens_seen": 403848192
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004432397191574724,
      "loss": 2.8846,
      "theoretical_loss": 4.007587062605954,
      "tokens_seen": 403913728
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044322968906720163,
      "loss": 3.2135,
      "theoretical_loss": 4.007514440616128,
      "tokens_seen": 403979264
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044321965897693076,
      "loss": 2.8579,
      "theoretical_loss": 4.007441833704654,
      "tokens_seen": 404044800
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044320962888666,
      "loss": 2.7827,
      "theoretical_loss": 4.0073692418659554,
      "tokens_seen": 404110336
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004431995987963892,
      "loss": 3.0201,
      "theoretical_loss": 4.007296665094462,
      "tokens_seen": 404175872
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044318956870611836,
      "loss": 3.2041,
      "theoretical_loss": 4.007224103384603,
      "tokens_seen": 404241408
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044317953861584754,
      "loss": 2.8284,
      "theoretical_loss": 4.007151556730811,
      "tokens_seen": 404306944
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004431695085255767,
      "loss": 3.0129,
      "theoretical_loss": 4.007079025127523,
      "tokens_seen": 404372480
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004431594784353059,
      "loss": 3.0203,
      "theoretical_loss": 4.007006508569178,
      "tokens_seen": 404438016
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044314944834503514,
      "loss": 3.1886,
      "theoretical_loss": 4.0069340070502175,
      "tokens_seen": 404503552
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044313941825476427,
      "loss": 3.0306,
      "theoretical_loss": 4.006861520565088,
      "tokens_seen": 404569088
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004431293881644935,
      "loss": 3.1072,
      "theoretical_loss": 4.006789049108235,
      "tokens_seen": 404634624
    },
    {
      "epoch": 1.03,
      "objective/train/docs_used": 517837,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.787900924682617,
      "objective/train/theoretical_loss": 4.006734705374566,
      "objective/train/tokens_used": 425143776,
      "theoretical_loss": 4.006734705374566,
      "tokens_seen": 404683776
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004431193580742227,
      "loss": 2.9616,
      "theoretical_loss": 4.006716592674112,
      "tokens_seen": 404700160
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044310932798395186,
      "loss": 3.1183,
      "theoretical_loss": 4.006644151257173,
      "tokens_seen": 404765696
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044309929789368104,
      "loss": 2.8431,
      "theoretical_loss": 4.006571724851873,
      "tokens_seen": 404831232
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004430892678034102,
      "loss": 2.9166,
      "theoretical_loss": 4.0064993134526725,
      "tokens_seen": 404896768
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004430792377131394,
      "loss": 3.0856,
      "theoretical_loss": 4.006426917054036,
      "tokens_seen": 404962304
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044306920762286864,
      "loss": 3.0765,
      "theoretical_loss": 4.006354535650425,
      "tokens_seen": 405027840
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044305917753259777,
      "loss": 3.1055,
      "theoretical_loss": 4.0062821692363135,
      "tokens_seen": 405093376
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.000443049147442327,
      "loss": 3.0614,
      "theoretical_loss": 4.00620981780617,
      "tokens_seen": 405158912
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044303911735205613,
      "loss": 3.0428,
      "theoretical_loss": 4.006137481354468,
      "tokens_seen": 405224448
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044302908726178537,
      "loss": 3.0897,
      "theoretical_loss": 4.006065159875687,
      "tokens_seen": 405289984
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044301905717151455,
      "loss": 2.9146,
      "theoretical_loss": 4.005992853364307,
      "tokens_seen": 405355520
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044300902708124373,
      "loss": 3.0938,
      "theoretical_loss": 4.005920561814811,
      "tokens_seen": 405421056
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004429989969909729,
      "loss": 3.0032,
      "theoretical_loss": 4.005848285221685,
      "tokens_seen": 405486592
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004429889669007021,
      "loss": 2.8133,
      "theoretical_loss": 4.005776023579418,
      "tokens_seen": 405552128
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044297893681043127,
      "loss": 2.9431,
      "theoretical_loss": 4.005703776882503,
      "tokens_seen": 405617664
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004429689067201605,
      "loss": 3.0047,
      "theoretical_loss": 4.005631545125434,
      "tokens_seen": 405683200
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044295887662988963,
      "loss": 2.989,
      "theoretical_loss": 4.00555932830271,
      "tokens_seen": 405748736
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044294884653961887,
      "loss": 3.2523,
      "theoretical_loss": 4.0054871264088305,
      "tokens_seen": 405814272
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004429388164493481,
      "loss": 3.0664,
      "theoretical_loss": 4.0054149394383,
      "tokens_seen": 405879808
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044292878635907723,
      "loss": 3.2111,
      "theoretical_loss": 4.005342767385625,
      "tokens_seen": 405945344
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044291875626880647,
      "loss": 2.8902,
      "theoretical_loss": 4.005270610245315,
      "tokens_seen": 406010880
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004429087261785356,
      "loss": 3.4227,
      "theoretical_loss": 4.005198468011882,
      "tokens_seen": 406076416
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044289869608826483,
      "loss": 3.1243,
      "theoretical_loss": 4.005126340679842,
      "tokens_seen": 406141952
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.000442888665997994,
      "loss": 3.1989,
      "theoretical_loss": 4.005054228243713,
      "tokens_seen": 406207488
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004428786359077232,
      "loss": 2.9508,
      "theoretical_loss": 4.004982130698017,
      "tokens_seen": 406273024
    },
    {
      "epoch": 1.03,
      "objective/train/docs_used": 518449,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.878129482269287,
      "objective/train/theoretical_loss": 4.004928067307296,
      "objective/train/tokens_used": 426782176,
      "theoretical_loss": 4.004928067307296,
      "tokens_seen": 406322176
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044286860581745237,
      "loss": 3.0946,
      "theoretical_loss": 4.004910048037276,
      "tokens_seen": 406338560
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044285857572718155,
      "loss": 2.8832,
      "theoretical_loss": 4.00483798025602,
      "tokens_seen": 406404096
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044284854563691073,
      "loss": 3.1023,
      "theoretical_loss": 4.0047659273487755,
      "tokens_seen": 406469632
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044283851554663997,
      "loss": 3.0547,
      "theoretical_loss": 4.004693889310077,
      "tokens_seen": 406535168
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004428284854563691,
      "loss": 3.1845,
      "theoretical_loss": 4.00462186613446,
      "tokens_seen": 406600704
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044281845536609833,
      "loss": 3.0654,
      "theoretical_loss": 4.004549857816463,
      "tokens_seen": 406666240
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004428084252758275,
      "loss": 3.0538,
      "theoretical_loss": 4.004477864350627,
      "tokens_seen": 406731776
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004427983951855567,
      "loss": 3.0571,
      "theoretical_loss": 4.004405885731497,
      "tokens_seen": 406797312
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004427883650952859,
      "loss": 3.0495,
      "theoretical_loss": 4.00433392195362,
      "tokens_seen": 406862848
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044277833500501506,
      "loss": 3.1538,
      "theoretical_loss": 4.004261973011546,
      "tokens_seen": 406928384
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044276830491474424,
      "loss": 2.9353,
      "theoretical_loss": 4.004190038899827,
      "tokens_seen": 406993920
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044275827482447347,
      "loss": 2.8766,
      "theoretical_loss": 4.004118119613019,
      "tokens_seen": 407059456
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004427482447342026,
      "loss": 3.0604,
      "theoretical_loss": 4.004046215145682,
      "tokens_seen": 407124992
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044273821464393184,
      "loss": 3.2647,
      "theoretical_loss": 4.003974325492377,
      "tokens_seen": 407190528
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044272818455366096,
      "loss": 3.1432,
      "theoretical_loss": 4.003902450647669,
      "tokens_seen": 407256064
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004427181544633902,
      "loss": 3.0507,
      "theoretical_loss": 4.003830590606123,
      "tokens_seen": 407321600
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004427081243731194,
      "loss": 3.0707,
      "theoretical_loss": 4.003758745362311,
      "tokens_seen": 407387136
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044269809428284856,
      "loss": 3.1622,
      "theoretical_loss": 4.003686914910807,
      "tokens_seen": 407452672
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044268806419257774,
      "loss": 3.146,
      "theoretical_loss": 4.003615099246184,
      "tokens_seen": 407518208
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004426780341023069,
      "loss": 3.2283,
      "theoretical_loss": 4.003543298363022,
      "tokens_seen": 407583744
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004426680040120361,
      "loss": 2.921,
      "theoretical_loss": 4.003471512255905,
      "tokens_seen": 407649280
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044265797392176534,
      "loss": 3.1143,
      "theoretical_loss": 4.003399740919413,
      "tokens_seen": 407714816
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044264794383149447,
      "loss": 3.0341,
      "theoretical_loss": 4.003327984348138,
      "tokens_seen": 407780352
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004426379137412237,
      "loss": 3.1449,
      "theoretical_loss": 4.003256242536667,
      "tokens_seen": 407845888
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004426278836509529,
      "loss": 2.9538,
      "theoretical_loss": 4.003184515479592,
      "tokens_seen": 407911424
    },
    {
      "epoch": 1.03,
      "objective/train/docs_used": 519624,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.0527055263519287,
      "objective/train/theoretical_loss": 4.00313072986611,
      "objective/train/tokens_used": 428420576,
      "theoretical_loss": 4.00313072986611,
      "tokens_seen": 407960576
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044261785356068206,
      "loss": 3.0662,
      "theoretical_loss": 4.003112803171513,
      "tokens_seen": 407976960
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044260782347041124,
      "loss": 2.9138,
      "theoretical_loss": 4.003041105607026,
      "tokens_seen": 408042496
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004425977933801404,
      "loss": 3.1145,
      "theoretical_loss": 4.002969422780733,
      "tokens_seen": 408108032
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004425877632898696,
      "loss": 3.0576,
      "theoretical_loss": 4.002897754687238,
      "tokens_seen": 408173568
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044257773319959884,
      "loss": 3.2281,
      "theoretical_loss": 4.0028261013211495,
      "tokens_seen": 408239104
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044256770310932797,
      "loss": 3.1774,
      "theoretical_loss": 4.002754462677077,
      "tokens_seen": 408304640
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004425576730190572,
      "loss": 3.0006,
      "theoretical_loss": 4.002682838749632,
      "tokens_seen": 408370176
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044254764292878633,
      "loss": 3.1701,
      "theoretical_loss": 4.002611229533432,
      "tokens_seen": 408435712
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044253761283851557,
      "loss": 3.0723,
      "theoretical_loss": 4.002539635023095,
      "tokens_seen": 408501248
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044252758274824475,
      "loss": 3.0007,
      "theoretical_loss": 4.002468055213243,
      "tokens_seen": 408566784
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044251755265797393,
      "loss": 3.1507,
      "theoretical_loss": 4.0023964900985,
      "tokens_seen": 408632320
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004425075225677031,
      "loss": 3.3785,
      "theoretical_loss": 4.002324939673492,
      "tokens_seen": 408697856
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004424974924774323,
      "loss": 3.0969,
      "theoretical_loss": 4.0022534039328495,
      "tokens_seen": 408763392
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044248746238716147,
      "loss": 3.1206,
      "theoretical_loss": 4.002181882871206,
      "tokens_seen": 408828928
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004424774322968907,
      "loss": 2.8046,
      "theoretical_loss": 4.002110376483197,
      "tokens_seen": 408894464
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044246740220661983,
      "loss": 2.9415,
      "theoretical_loss": 4.002038884763459,
      "tokens_seen": 408960000
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044245737211634907,
      "loss": 2.9706,
      "theoretical_loss": 4.001967407706635,
      "tokens_seen": 409025536
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044244734202607825,
      "loss": 2.9634,
      "theoretical_loss": 4.0018959453073695,
      "tokens_seen": 409091072
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044243731193580743,
      "loss": 3.2558,
      "theoretical_loss": 4.001824497560308,
      "tokens_seen": 409156608
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004424272818455366,
      "loss": 3.1996,
      "theoretical_loss": 4.0017530644601,
      "tokens_seen": 409222144
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004424172517552658,
      "loss": 2.9816,
      "theoretical_loss": 4.001681646001399,
      "tokens_seen": 409287680
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.000442407221664995,
      "loss": 3.027,
      "theoretical_loss": 4.0016102421788595,
      "tokens_seen": 409353216
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004423971915747242,
      "loss": 2.993,
      "theoretical_loss": 4.00153885298714,
      "tokens_seen": 409418752
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044238716148445334,
      "loss": 2.9679,
      "theoretical_loss": 4.001467478420901,
      "tokens_seen": 409484288
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044237713139418257,
      "loss": 3.3059,
      "theoretical_loss": 4.001396118474806,
      "tokens_seen": 409549824
    },
    {
      "epoch": 1.03,
      "objective/train/docs_used": 520212,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.317723512649536,
      "objective/train/theoretical_loss": 4.001342608106496,
      "objective/train/tokens_used": 430058976,
      "theoretical_loss": 4.001342608106496,
      "tokens_seen": 409598976
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004423671013039117,
      "loss": 2.9615,
      "theoretical_loss": 4.001324773143522,
      "tokens_seen": 409615360
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044235707121364093,
      "loss": 3.0724,
      "theoretical_loss": 4.001253442421718,
      "tokens_seen": 409680896
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004423470411233701,
      "loss": 2.965,
      "theoretical_loss": 4.001182126304067,
      "tokens_seen": 409746432
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004423370110330993,
      "loss": 3.0159,
      "theoretical_loss": 4.001110824785242,
      "tokens_seen": 409811968
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004423269809428285,
      "loss": 2.9357,
      "theoretical_loss": 4.001039537859921,
      "tokens_seen": 409877504
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004423169508525577,
      "loss": 3.4018,
      "theoretical_loss": 4.000968265522786,
      "tokens_seen": 409943040
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044230692076228684,
      "loss": 3.1943,
      "theoretical_loss": 4.000897007768518,
      "tokens_seen": 410008576
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004422968906720161,
      "loss": 2.9675,
      "theoretical_loss": 4.000825764591807,
      "tokens_seen": 410074112
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004422868605817452,
      "loss": 3.1719,
      "theoretical_loss": 4.000754535987337,
      "tokens_seen": 410139648
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044227683049147444,
      "loss": 3.1564,
      "theoretical_loss": 4.0006833219498015,
      "tokens_seen": 410205184
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004422668004012036,
      "loss": 3.1727,
      "theoretical_loss": 4.000612122473896,
      "tokens_seen": 410270720
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004422567703109328,
      "loss": 3.0706,
      "theoretical_loss": 4.000540937554316,
      "tokens_seen": 410336256
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.000442246740220662,
      "loss": 3.193,
      "theoretical_loss": 4.000469767185763,
      "tokens_seen": 410401792
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044223671013039116,
      "loss": 2.8826,
      "theoretical_loss": 4.000398611362938,
      "tokens_seen": 410467328
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044222668004012034,
      "loss": 3.0742,
      "theoretical_loss": 4.0003274700805465,
      "tokens_seen": 410532864
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004422166499498496,
      "loss": 3.1663,
      "theoretical_loss": 4.000256343333298,
      "tokens_seen": 410598400
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004422066198595787,
      "loss": 3.1183,
      "theoretical_loss": 4.000185231115903,
      "tokens_seen": 410663936
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044219658976930794,
      "loss": 3.0318,
      "theoretical_loss": 4.000114133423077,
      "tokens_seen": 410729472
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004421865596790371,
      "loss": 2.8639,
      "theoretical_loss": 4.000043050249533,
      "tokens_seen": 410795008
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004421765295887663,
      "loss": 3.0937,
      "theoretical_loss": 3.999971981589993,
      "tokens_seen": 410860544
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044216649949849554,
      "loss": 2.9955,
      "theoretical_loss": 3.9999009274391772,
      "tokens_seen": 410926080
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044215646940822467,
      "loss": 3.2867,
      "theoretical_loss": 3.9998298877918126,
      "tokens_seen": 410991616
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004421464393179539,
      "loss": 3.0194,
      "theoretical_loss": 3.999758862642625,
      "tokens_seen": 411057152
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004421364092276831,
      "loss": 3.0622,
      "theoretical_loss": 3.9996878519863452,
      "tokens_seen": 411122688
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044212637913741226,
      "loss": 3.0314,
      "theoretical_loss": 3.999616855817707,
      "tokens_seen": 411188224
    },
    {
      "epoch": 1.03,
      "objective/train/docs_used": 521565,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.3985636234283447,
      "objective/train/theoretical_loss": 3.9995636181955763,
      "objective/train/tokens_used": 431697376,
      "theoretical_loss": 3.9995636181955763,
      "tokens_seen": 411237376
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044211634904714144,
      "loss": 3.0621,
      "theoretical_loss": 3.999545874131446,
      "tokens_seen": 411253760
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004421063189568706,
      "loss": 3.194,
      "theoretical_loss": 3.9994749069223,
      "tokens_seen": 411319296
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004420962888665998,
      "loss": 2.8937,
      "theoretical_loss": 3.9994039541850115,
      "tokens_seen": 411384832
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044208625877632904,
      "loss": 2.9052,
      "theoretical_loss": 3.9993330159143237,
      "tokens_seen": 411450368
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044207622868605817,
      "loss": 3.1371,
      "theoretical_loss": 3.999262092104984,
      "tokens_seen": 411515904
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004420661985957874,
      "loss": 3.1621,
      "theoretical_loss": 3.9991911827517415,
      "tokens_seen": 411581440
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044205616850551653,
      "loss": 2.9365,
      "theoretical_loss": 3.9991202878493493,
      "tokens_seen": 411646976
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044204613841524577,
      "loss": 3.1542,
      "theoretical_loss": 3.9990494073925618,
      "tokens_seen": 411712512
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044203610832497495,
      "loss": 3.4033,
      "theoretical_loss": 3.9989785413761365,
      "tokens_seen": 411778048
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044202607823470413,
      "loss": 3.1086,
      "theoretical_loss": 3.9989076897948355,
      "tokens_seen": 411843584
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004420160481444333,
      "loss": 2.9042,
      "theoretical_loss": 3.99883685264342,
      "tokens_seen": 411909120
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004420060180541625,
      "loss": 3.1093,
      "theoretical_loss": 3.9987660299166574,
      "tokens_seen": 411974656
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044199598796389167,
      "loss": 3.225,
      "theoretical_loss": 3.998695221609317,
      "tokens_seen": 412040192
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0004419859578736209,
      "loss": 3.0794,
      "theoretical_loss": 3.9986244277161687,
      "tokens_seen": 412105728
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044197592778335003,
      "loss": 2.9221,
      "theoretical_loss": 3.9985536482319874,
      "tokens_seen": 412171264
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.00044196589769307927,
      "loss": 3.2153,
      "theoretical_loss": 3.99848288315155,
      "tokens_seen": 412236800
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00044195586760280845,
      "loss": 3.1729,
      "theoretical_loss": 3.998412132469637,
      "tokens_seen": 412302336
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00044194583751253763,
      "loss": 3.0724,
      "theoretical_loss": 3.9983413961810292,
      "tokens_seen": 412367872
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004419358074222668,
      "loss": 2.8081,
      "theoretical_loss": 3.9982706742805125,
      "tokens_seen": 412433408
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.000441925777331996,
      "loss": 2.9037,
      "theoretical_loss": 3.998199966762875,
      "tokens_seen": 412498944
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004419157472417252,
      "loss": 3.138,
      "theoretical_loss": 3.998129273622907,
      "tokens_seen": 412564480
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004419057171514544,
      "loss": 2.9746,
      "theoretical_loss": 3.9980585948554013,
      "tokens_seen": 412630016
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00044189568706118354,
      "loss": 2.8928,
      "theoretical_loss": 3.9979879304551544,
      "tokens_seen": 412695552
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00044188565697091277,
      "loss": 3.0935,
      "theoretical_loss": 3.9979172804169654,
      "tokens_seen": 412761088
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004418756268806419,
      "loss": 3.2098,
      "theoretical_loss": 3.997846644735634,
      "tokens_seen": 412826624
    },
    {
      "epoch": 1.04,
      "objective/train/docs_used": 522883,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.032809019088745,
      "objective/train/theoretical_loss": 3.997793677393199,
      "objective/train/tokens_used": 433335776,
      "theoretical_loss": 3.997793677393199,
      "tokens_seen": 412875776
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00044186559679037113,
      "loss": 2.9367,
      "theoretical_loss": 3.997776023405966,
      "tokens_seen": 412892160
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004418555667001003,
      "loss": 3.0454,
      "theoretical_loss": 3.997705416422767,
      "tokens_seen": 412957696
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004418455366098295,
      "loss": 2.9081,
      "theoretical_loss": 3.9976348237808477,
      "tokens_seen": 413023232
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004418355065195587,
      "loss": 3.3134,
      "theoretical_loss": 3.9975642454750195,
      "tokens_seen": 413088768
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004418254764292879,
      "loss": 3.0579,
      "theoretical_loss": 3.9974936815000977,
      "tokens_seen": 413154304
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00044181544633901704,
      "loss": 3.1252,
      "theoretical_loss": 3.997423131850899,
      "tokens_seen": 413219840
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004418054162487463,
      "loss": 3.1003,
      "theoretical_loss": 3.9973525965222443,
      "tokens_seen": 413285376
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004417953861584754,
      "loss": 3.0599,
      "theoretical_loss": 3.9972820755089566,
      "tokens_seen": 413350912
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00044178535606820464,
      "loss": 2.923,
      "theoretical_loss": 3.9972115688058616,
      "tokens_seen": 413416448
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004417753259779338,
      "loss": 3.0775,
      "theoretical_loss": 3.9971410764077877,
      "tokens_seen": 413481984
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.000441765295887663,
      "loss": 2.9165,
      "theoretical_loss": 3.997070598309566,
      "tokens_seen": 413547520
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004417552657973922,
      "loss": 2.9974,
      "theoretical_loss": 3.9970001345060293,
      "tokens_seen": 413613056
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00044174523570712136,
      "loss": 3.001,
      "theoretical_loss": 3.9969296849920153,
      "tokens_seen": 413678592
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00044173520561685054,
      "loss": 3.0008,
      "theoretical_loss": 3.9968592497623625,
      "tokens_seen": 413744128
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004417251755265798,
      "loss": 3.11,
      "theoretical_loss": 3.996788828811913,
      "tokens_seen": 413809664
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004417151454363089,
      "loss": 3.2571,
      "theoretical_loss": 3.996718422135511,
      "tokens_seen": 413875200
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00044170511534603814,
      "loss": 2.9652,
      "theoretical_loss": 3.996648029728003,
      "tokens_seen": 413940736
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00044169508525576727,
      "loss": 3.0139,
      "theoretical_loss": 3.9965776515842397,
      "tokens_seen": 414006272
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004416850551654965,
      "loss": 2.9294,
      "theoretical_loss": 3.9965072876990737,
      "tokens_seen": 414071808
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004416750250752257,
      "loss": 3.0109,
      "theoretical_loss": 3.99643693806736,
      "tokens_seen": 414137344
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00044166499498495487,
      "loss": 3.1063,
      "theoretical_loss": 3.9963666026839557,
      "tokens_seen": 414202880
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00044165496489468405,
      "loss": 3.2242,
      "theoretical_loss": 3.996296281543722,
      "tokens_seen": 414268416
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004416449348044133,
      "loss": 2.8582,
      "theoretical_loss": 3.996225974641522,
      "tokens_seen": 414333952
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004416349047141424,
      "loss": 2.8527,
      "theoretical_loss": 3.9961556819722217,
      "tokens_seen": 414399488
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00044162487462387164,
      "loss": 3.037,
      "theoretical_loss": 3.9960854035306888,
      "tokens_seen": 414465024
    },
    {
      "epoch": 1.04,
      "objective/train/docs_used": 523694,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.1503584384918213,
      "objective/train/theoretical_loss": 3.9960327040334263,
      "objective/train/tokens_used": 434974176,
      "theoretical_loss": 3.9960327040334263,
      "tokens_seen": 414514176
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00044161484453360077,
      "loss": 3.2747,
      "theoretical_loss": 3.996015139311795,
      "tokens_seen": 414530560
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00044160481444333,
      "loss": 3.2075,
      "theoretical_loss": 3.9959448893104144,
      "tokens_seen": 414596096
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004415947843530592,
      "loss": 3.0925,
      "theoretical_loss": 3.9958746535214233,
      "tokens_seen": 414661632
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00044158475426278837,
      "loss": 3.201,
      "theoretical_loss": 3.9958044319397006,
      "tokens_seen": 414727168
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00044157472417251755,
      "loss": 2.9766,
      "theoretical_loss": 3.9957342245601275,
      "tokens_seen": 414792704
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00044156469408224673,
      "loss": 3.1882,
      "theoretical_loss": 3.99566403137759,
      "tokens_seen": 414858240
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004415546639919759,
      "loss": 3.0593,
      "theoretical_loss": 3.995593852386974,
      "tokens_seen": 414923776
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00044154463390170515,
      "loss": 3.1353,
      "theoretical_loss": 3.995523687583169,
      "tokens_seen": 414989312
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004415346038114343,
      "loss": 3.1151,
      "theoretical_loss": 3.9954535369610684,
      "tokens_seen": 415054848
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004415245737211635,
      "loss": 3.1027,
      "theoretical_loss": 3.995383400515567,
      "tokens_seen": 415120384
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00044151454363089264,
      "loss": 3.0921,
      "theoretical_loss": 3.9953132782415617,
      "tokens_seen": 415185920
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00044150451354062187,
      "loss": 3.1593,
      "theoretical_loss": 3.9952431701339535,
      "tokens_seen": 415251456
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00044149448345035105,
      "loss": 3.2781,
      "theoretical_loss": 3.995173076187645,
      "tokens_seen": 415316992
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00044148445336008023,
      "loss": 2.9717,
      "theoretical_loss": 3.995102996397542,
      "tokens_seen": 415382528
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004414744232698094,
      "loss": 3.0149,
      "theoretical_loss": 3.995032930758552,
      "tokens_seen": 415448064
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00044146439317953865,
      "loss": 3.0351,
      "theoretical_loss": 3.9949628792655876,
      "tokens_seen": 415513600
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004414543630892678,
      "loss": 3.1212,
      "theoretical_loss": 3.9948928419135603,
      "tokens_seen": 415579136
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.000441444332998997,
      "loss": 2.9529,
      "theoretical_loss": 3.9948228186973873,
      "tokens_seen": 415644672
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004414343029087262,
      "loss": 3.1814,
      "theoretical_loss": 3.9947528096119873,
      "tokens_seen": 415710208
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004414242728184554,
      "loss": 3.1078,
      "theoretical_loss": 3.9946828146522817,
      "tokens_seen": 415775744
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004414142427281846,
      "loss": 2.9443,
      "theoretical_loss": 3.9946128338131945,
      "tokens_seen": 415841280
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00044140421263791374,
      "loss": 2.9845,
      "theoretical_loss": 3.9945428670896517,
      "tokens_seen": 415906816
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00044139418254764297,
      "loss": 3.0643,
      "theoretical_loss": 3.9944729144765834,
      "tokens_seen": 415972352
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004413841524573721,
      "loss": 2.9436,
      "theoretical_loss": 3.9944029759689204,
      "tokens_seen": 416037888
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00044137412236710133,
      "loss": 2.895,
      "theoretical_loss": 3.9943330515615987,
      "tokens_seen": 416103424
    },
    {
      "epoch": 1.04,
      "objective/train/docs_used": 524726,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.1443862915039062,
      "objective/train/theoretical_loss": 3.9942806175064094,
      "objective/train/tokens_used": 436612576,
      "theoretical_loss": 3.9942806175064094,
      "tokens_seen": 416152576
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004413640922768305,
      "loss": 2.8181,
      "theoretical_loss": 3.994263141249554,
      "tokens_seen": 416168960
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004413540621865597,
      "loss": 3.0632,
      "theoretical_loss": 3.994193245027726,
      "tokens_seen": 416234496
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004413440320962889,
      "loss": 3.2087,
      "theoretical_loss": 3.9941233628910586,
      "tokens_seen": 416300032
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004413340020060181,
      "loss": 3.0843,
      "theoretical_loss": 3.9940534948344952,
      "tokens_seen": 416365568
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00044132397191574724,
      "loss": 3.115,
      "theoretical_loss": 3.993983640852984,
      "tokens_seen": 416431104
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004413139418254765,
      "loss": 3.1099,
      "theoretical_loss": 3.993913800941475,
      "tokens_seen": 416496640
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004413039117352056,
      "loss": 3.0741,
      "theoretical_loss": 3.9938439750949213,
      "tokens_seen": 416562176
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00044129388164493484,
      "loss": 3.0151,
      "theoretical_loss": 3.9937741633082777,
      "tokens_seen": 416627712
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.000441283851554664,
      "loss": 3.0126,
      "theoretical_loss": 3.9937043655765025,
      "tokens_seen": 416693248
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004412738214643932,
      "loss": 3.0725,
      "theoretical_loss": 3.993634581894556,
      "tokens_seen": 416758784
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004412637913741224,
      "loss": 3.2187,
      "theoretical_loss": 3.9935648122574015,
      "tokens_seen": 416824320
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00044125376128385156,
      "loss": 3.131,
      "theoretical_loss": 3.993495056660005,
      "tokens_seen": 416889856
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00044124373119358074,
      "loss": 3.0721,
      "theoretical_loss": 3.993425315097335,
      "tokens_seen": 416955392
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00044123370110331,
      "loss": 2.9879,
      "theoretical_loss": 3.993355587564362,
      "tokens_seen": 417020928
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004412236710130391,
      "loss": 3.079,
      "theoretical_loss": 3.9932858740560597,
      "tokens_seen": 417086464
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00044121364092276834,
      "loss": 3.1582,
      "theoretical_loss": 3.9932161745674035,
      "tokens_seen": 417152000
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00044120361083249747,
      "loss": 3.2366,
      "theoretical_loss": 3.9931464890933737,
      "tokens_seen": 417217536
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004411935807422267,
      "loss": 3.0629,
      "theoretical_loss": 3.9930768176289506,
      "tokens_seen": 417283072
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004411835506519559,
      "loss": 3.1617,
      "theoretical_loss": 3.9930071601691184,
      "tokens_seen": 417348608
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00044117352056168507,
      "loss": 3.0412,
      "theoretical_loss": 3.9929375167088637,
      "tokens_seen": 417414144
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00044116349047141425,
      "loss": 2.9656,
      "theoretical_loss": 3.9928678872431744,
      "tokens_seen": 417479680
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004411534603811435,
      "loss": 2.9363,
      "theoretical_loss": 3.992798271767044,
      "tokens_seen": 417545216
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004411434302908726,
      "loss": 2.9418,
      "theoretical_loss": 3.992728670275465,
      "tokens_seen": 417610752
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00044113340020060184,
      "loss": 3.0558,
      "theoretical_loss": 3.992659082763436,
      "tokens_seen": 417676288
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00044112337011033097,
      "loss": 3.0002,
      "theoretical_loss": 3.9925895092259545,
      "tokens_seen": 417741824
    },
    {
      "epoch": 1.04,
      "objective/train/docs_used": 525508,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.890538454055786,
      "objective/train/theoretical_loss": 3.992537338240634,
      "objective/train/tokens_used": 438250976,
      "theoretical_loss": 3.992537338240634,
      "tokens_seen": 417790976
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004411133400200602,
      "loss": 2.9902,
      "theoretical_loss": 3.9925199496580235,
      "tokens_seen": 417807360
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004411033099297894,
      "loss": 3.0079,
      "theoretical_loss": 3.9924504040546474,
      "tokens_seen": 417872896
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00044109327983951857,
      "loss": 3.1488,
      "theoretical_loss": 3.992380872410833,
      "tokens_seen": 417938432
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00044108324974924775,
      "loss": 3.2423,
      "theoretical_loss": 3.9923113547215907,
      "tokens_seen": 418003968
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00044107321965897693,
      "loss": 3.121,
      "theoretical_loss": 3.992241850981932,
      "tokens_seen": 418069504
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004410631895687061,
      "loss": 2.8882,
      "theoretical_loss": 3.9921723611868716,
      "tokens_seen": 418135040
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00044105315947843535,
      "loss": 3.145,
      "theoretical_loss": 3.992102885331427,
      "tokens_seen": 418200576
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004410431293881645,
      "loss": 2.9919,
      "theoretical_loss": 3.9920334234106187,
      "tokens_seen": 418266112
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004410330992978937,
      "loss": 2.9014,
      "theoretical_loss": 3.991963975419468,
      "tokens_seen": 418331648
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00044102306920762284,
      "loss": 3.1452,
      "theoretical_loss": 3.991894541353001,
      "tokens_seen": 418397184
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00044101303911735207,
      "loss": 3.098,
      "theoretical_loss": 3.9918251212062454,
      "tokens_seen": 418462720
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00044100300902708125,
      "loss": 2.9439,
      "theoretical_loss": 3.99175571497423,
      "tokens_seen": 418528256
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00044099297893681043,
      "loss": 2.8596,
      "theoretical_loss": 3.991686322651989,
      "tokens_seen": 418593792
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004409829488465396,
      "loss": 2.8567,
      "theoretical_loss": 3.9916169442345564,
      "tokens_seen": 418659328
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00044097291875626885,
      "loss": 3.0336,
      "theoretical_loss": 3.991547579716971,
      "tokens_seen": 418724864
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.000440962888665998,
      "loss": 3.01,
      "theoretical_loss": 3.9914782290942723,
      "tokens_seen": 418790400
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004409528585757272,
      "loss": 2.718,
      "theoretical_loss": 3.9914088923615036,
      "tokens_seen": 418855936
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00044094282848545634,
      "loss": 2.9002,
      "theoretical_loss": 3.9913395695137104,
      "tokens_seen": 418921472
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004409327983951856,
      "loss": 3.1815,
      "theoretical_loss": 3.99127026054594,
      "tokens_seen": 418987008
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00044092276830491476,
      "loss": 3.1064,
      "theoretical_loss": 3.991200965453244,
      "tokens_seen": 419052544
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00044091273821464394,
      "loss": 2.9756,
      "theoretical_loss": 3.991131684230675,
      "tokens_seen": 419118080
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004409027081243731,
      "loss": 3.1044,
      "theoretical_loss": 3.991062416873288,
      "tokens_seen": 419183616
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004408926780341023,
      "loss": 2.9907,
      "theoretical_loss": 3.990993163376142,
      "tokens_seen": 419249152
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004408826479438315,
      "loss": 2.5876,
      "theoretical_loss": 3.990923923734297,
      "tokens_seen": 419314688
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004408726178535607,
      "loss": 2.9581,
      "theoretical_loss": 3.990854697942816,
      "tokens_seen": 419380224
    },
    {
      "epoch": 1.04,
      "objective/train/docs_used": 526274,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7906696796417236,
      "objective/train/theoretical_loss": 3.9908027876855394,
      "objective/train/tokens_used": 439889376,
      "theoretical_loss": 3.9908027876855394,
      "tokens_seen": 419429376
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00044086258776328984,
      "loss": 2.9342,
      "theoretical_loss": 3.990785485996766,
      "tokens_seen": 419445760
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004408525576730191,
      "loss": 2.8886,
      "theoretical_loss": 3.990716287891214,
      "tokens_seen": 419511296
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004408425275827482,
      "loss": 2.9543,
      "theoretical_loss": 3.9906471036212308,
      "tokens_seen": 419576832
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00044083249749247744,
      "loss": 3.0434,
      "theoretical_loss": 3.9905779331818905,
      "tokens_seen": 419642368
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004408224674022066,
      "loss": 2.7933,
      "theoretical_loss": 3.9905087765682676,
      "tokens_seen": 419707904
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004408124373119358,
      "loss": 2.9713,
      "theoretical_loss": 3.990439633775442,
      "tokens_seen": 419773440
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.000440802407221665,
      "loss": 3.2049,
      "theoretical_loss": 3.9903705047984936,
      "tokens_seen": 419838976
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004407923771313942,
      "loss": 3.0042,
      "theoretical_loss": 3.9903013896325064,
      "tokens_seen": 419904512
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00044078234704112335,
      "loss": 3.002,
      "theoretical_loss": 3.9902322882725656,
      "tokens_seen": 419970048
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004407723169508526,
      "loss": 3.2834,
      "theoretical_loss": 3.9901632007137593,
      "tokens_seen": 420035584
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004407622868605817,
      "loss": 3.0347,
      "theoretical_loss": 3.9900941269511794,
      "tokens_seen": 420101120
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00044075225677031094,
      "loss": 3.0477,
      "theoretical_loss": 3.990025066979919,
      "tokens_seen": 420166656
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004407422266800401,
      "loss": 2.8861,
      "theoretical_loss": 3.9899560207950744,
      "tokens_seen": 420232192
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004407321965897693,
      "loss": 2.878,
      "theoretical_loss": 3.989886988391743,
      "tokens_seen": 420297728
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004407221664994985,
      "loss": 3.1336,
      "theoretical_loss": 3.989817969765027,
      "tokens_seen": 420363264
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00044071213640922767,
      "loss": 3.0215,
      "theoretical_loss": 3.9897489649100284,
      "tokens_seen": 420428800
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00044070210631895685,
      "loss": 2.9364,
      "theoretical_loss": 3.9896799738218545,
      "tokens_seen": 420494336
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004406920762286861,
      "loss": 3.0076,
      "theoretical_loss": 3.9896109964956135,
      "tokens_seen": 420559872
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00044068204613841527,
      "loss": 2.8764,
      "theoretical_loss": 3.9895420329264155,
      "tokens_seen": 420625408
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00044067201604814445,
      "loss": 3.2049,
      "theoretical_loss": 3.989473083109375,
      "tokens_seen": 420690944
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004406619859578737,
      "loss": 3.239,
      "theoretical_loss": 3.9894041470396076,
      "tokens_seen": 420756480
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004406519558676028,
      "loss": 3.1395,
      "theoretical_loss": 3.989335224712232,
      "tokens_seen": 420822016
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00044064192577733204,
      "loss": 3.0907,
      "theoretical_loss": 3.989266316122369,
      "tokens_seen": 420887552
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00044063189568706117,
      "loss": 3.1008,
      "theoretical_loss": 3.989197421265141,
      "tokens_seen": 420953088
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004406218655967904,
      "loss": 3.0425,
      "theoretical_loss": 3.9891285401356757,
      "tokens_seen": 421018624
    },
    {
      "epoch": 1.04,
      "objective/train/docs_used": 527775,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.216331958770752,
      "objective/train/theoretical_loss": 3.9890768882944894,
      "objective/train/tokens_used": 441527776,
      "theoretical_loss": 3.9890768882944894,
      "tokens_seen": 421067776
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004406118355065196,
      "loss": 3.0228,
      "theoretical_loss": 3.9890596727291,
      "tokens_seen": 421084160
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00044060180541624877,
      "loss": 2.7891,
      "theoretical_loss": 3.988990819040546,
      "tokens_seen": 421149696
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00044059177532597795,
      "loss": 3.0307,
      "theoretical_loss": 3.9889219790651462,
      "tokens_seen": 421215232
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00044058174523570713,
      "loss": 3.0988,
      "theoretical_loss": 3.988853152798037,
      "tokens_seen": 421280768
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004405717151454363,
      "loss": 2.8487,
      "theoretical_loss": 3.9887843402343566,
      "tokens_seen": 421346304
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00044056168505516555,
      "loss": 3.0812,
      "theoretical_loss": 3.9887155413692463,
      "tokens_seen": 421411840
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004405516549648947,
      "loss": 3.0347,
      "theoretical_loss": 3.9886467561978476,
      "tokens_seen": 421477376
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004405416248746239,
      "loss": 2.7433,
      "theoretical_loss": 3.9885779847153087,
      "tokens_seen": 421542912
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00044053159478435304,
      "loss": 3.1027,
      "theoretical_loss": 3.9885092269167766,
      "tokens_seen": 421608448
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00044052156469408227,
      "loss": 3.1315,
      "theoretical_loss": 3.9884404827974023,
      "tokens_seen": 421673984
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00044051153460381145,
      "loss": 2.9011,
      "theoretical_loss": 3.9883717523523385,
      "tokens_seen": 421739520
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00044050150451354063,
      "loss": 2.9213,
      "theoretical_loss": 3.9883030355767417,
      "tokens_seen": 421805056
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004404914744232698,
      "loss": 3.0712,
      "theoretical_loss": 3.98823433246577,
      "tokens_seen": 421870592
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00044048144433299905,
      "loss": 3.0033,
      "theoretical_loss": 3.9881656430145833,
      "tokens_seen": 421936128
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004404714142427282,
      "loss": 3.0171,
      "theoretical_loss": 3.988096967218345,
      "tokens_seen": 422001664
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004404613841524574,
      "loss": 2.9677,
      "theoretical_loss": 3.9880283050722207,
      "tokens_seen": 422067200
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00044045135406218654,
      "loss": 2.9678,
      "theoretical_loss": 3.9879596565713786,
      "tokens_seen": 422132736
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004404413239719158,
      "loss": 2.8646,
      "theoretical_loss": 3.987891021710989,
      "tokens_seen": 422198272
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00044043129388164496,
      "loss": 3.17,
      "theoretical_loss": 3.987822400486225,
      "tokens_seen": 422263808
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00044042126379137414,
      "loss": 3.1096,
      "theoretical_loss": 3.9877537928922613,
      "tokens_seen": 422329344
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004404112337011033,
      "loss": 3.0704,
      "theoretical_loss": 3.9876851989242765,
      "tokens_seen": 422394880
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004404012036108325,
      "loss": 3.0149,
      "theoretical_loss": 3.9876166185774506,
      "tokens_seen": 422460416
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004403911735205617,
      "loss": 3.1474,
      "theoretical_loss": 3.987548051846966,
      "tokens_seen": 422525952
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004403811434302909,
      "loss": 2.9211,
      "theoretical_loss": 3.987479498728009,
      "tokens_seen": 422591488
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00044037111334002004,
      "loss": 3.144,
      "theoretical_loss": 3.987410959215766,
      "tokens_seen": 422657024
    },
    {
      "epoch": 1.04,
      "objective/train/docs_used": 528406,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4659383296966553,
      "objective/train/theoretical_loss": 3.9873595635080967,
      "objective/train/tokens_used": 443166176,
      "theoretical_loss": 3.9873595635080967,
      "tokens_seen": 422706176
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004403610832497493,
      "loss": 3.073,
      "theoretical_loss": 3.987342433305428,
      "tokens_seen": 422722560
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004403510531594784,
      "loss": 3.0885,
      "theoretical_loss": 3.9872739209921866,
      "tokens_seen": 422788096
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00044034102306920764,
      "loss": 3.0353,
      "theoretical_loss": 3.987205422271238,
      "tokens_seen": 422853632
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004403309929789368,
      "loss": 3.1106,
      "theoretical_loss": 3.9871369371377785,
      "tokens_seen": 422919168
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.000440320962888666,
      "loss": 2.91,
      "theoretical_loss": 3.987068465587009,
      "tokens_seen": 422984704
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004403109327983952,
      "loss": 2.991,
      "theoretical_loss": 3.9870000076141303,
      "tokens_seen": 423050240
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004403009027081244,
      "loss": 3.0972,
      "theoretical_loss": 3.9869315632143483,
      "tokens_seen": 423115776
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00044029087261785355,
      "loss": 3.0585,
      "theoretical_loss": 3.9868631323828705,
      "tokens_seen": 423181312
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004402808425275828,
      "loss": 2.8196,
      "theoretical_loss": 3.9867947151149052,
      "tokens_seen": 423246848
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004402708124373119,
      "loss": 3.1558,
      "theoretical_loss": 3.9867263114056657,
      "tokens_seen": 423312384
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00044026078234704114,
      "loss": 3.2312,
      "theoretical_loss": 3.9866579212503654,
      "tokens_seen": 423377920
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004402507522567703,
      "loss": 3.2391,
      "theoretical_loss": 3.986589544644222,
      "tokens_seen": 423443456
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004402407221664995,
      "loss": 3.1316,
      "theoretical_loss": 3.986521181582455,
      "tokens_seen": 423508992
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004402306920762287,
      "loss": 3.2101,
      "theoretical_loss": 3.9864528320602854,
      "tokens_seen": 423574528
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00044022066198595787,
      "loss": 2.9653,
      "theoretical_loss": 3.9863844960729375,
      "tokens_seen": 423640064
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00044021063189568705,
      "loss": 3.0989,
      "theoretical_loss": 3.9863161736156387,
      "tokens_seen": 423705600
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004402006018054163,
      "loss": 2.7849,
      "theoretical_loss": 3.9862478646836172,
      "tokens_seen": 423771136
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004401905717151454,
      "loss": 2.9124,
      "theoretical_loss": 3.986179569272105,
      "tokens_seen": 423836672
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00044018054162487465,
      "loss": 3.0429,
      "theoretical_loss": 3.9861112873763354,
      "tokens_seen": 423902208
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00044017051153460383,
      "loss": 2.9653,
      "theoretical_loss": 3.986043018991545,
      "tokens_seen": 423967744
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.000440160481444333,
      "loss": 3.0573,
      "theoretical_loss": 3.9859747641129726,
      "tokens_seen": 424033280
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004401504513540622,
      "loss": 3.0092,
      "theoretical_loss": 3.9859065227358585,
      "tokens_seen": 424098816
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00044014042126379137,
      "loss": 2.9568,
      "theoretical_loss": 3.985838294855448,
      "tokens_seen": 424164352
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00044013039117352055,
      "loss": 2.8587,
      "theoretical_loss": 3.9857700804669847,
      "tokens_seen": 424229888
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004401203610832498,
      "loss": 2.9795,
      "theoretical_loss": 3.9857018795657186,
      "tokens_seen": 424295424
    },
    {
      "epoch": 1.04,
      "objective/train/docs_used": 530052,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.4916634559631348,
      "objective/train/theoretical_loss": 3.985650737737885,
      "objective/train/tokens_used": 444804576,
      "theoretical_loss": 3.985650737737885,
      "tokens_seen": 424344576
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004401103309929789,
      "loss": 3.1533,
      "theoretical_loss": 3.9856336921469007,
      "tokens_seen": 424360960
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00044010030090270815,
      "loss": 2.9523,
      "theoretical_loss": 3.9855655182057825,
      "tokens_seen": 424426496
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004400902708124373,
      "loss": 2.8567,
      "theoretical_loss": 3.9854973577376214,
      "tokens_seen": 424492032
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004400802407221665,
      "loss": 2.8743,
      "theoretical_loss": 3.9854292107376734,
      "tokens_seen": 424557568
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004400702106318957,
      "loss": 2.8946,
      "theoretical_loss": 3.9853610772012007,
      "tokens_seen": 424623104
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004400601805416249,
      "loss": 2.9646,
      "theoretical_loss": 3.985292957123465,
      "tokens_seen": 424688640
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00044005015045135406,
      "loss": 2.9647,
      "theoretical_loss": 3.9852248504997325,
      "tokens_seen": 424754176
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00044004012036108324,
      "loss": 2.936,
      "theoretical_loss": 3.985156757325269,
      "tokens_seen": 424819712
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004400300902708124,
      "loss": 2.8415,
      "theoretical_loss": 3.9850886775953462,
      "tokens_seen": 424885248
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00044002006018054165,
      "loss": 3.0282,
      "theoretical_loss": 3.9850206113052353,
      "tokens_seen": 424950784
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004400100300902708,
      "loss": 3.0696,
      "theoretical_loss": 3.9849525584502112,
      "tokens_seen": 425016320
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00044,
      "loss": 3.1165,
      "theoretical_loss": 3.984884519025552,
      "tokens_seen": 425081856
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004399899699097292,
      "loss": 3.2719,
      "theoretical_loss": 3.9848164930265364,
      "tokens_seen": 425147392
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004399799398194584,
      "loss": 2.9398,
      "theoretical_loss": 3.9847484804484457,
      "tokens_seen": 425212928
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00043996990972918756,
      "loss": 2.7438,
      "theoretical_loss": 3.9846804812865653,
      "tokens_seen": 425278464
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00043995987963891674,
      "loss": 3.0575,
      "theoretical_loss": 3.9846124955361812,
      "tokens_seen": 425344000
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004399498495486459,
      "loss": 2.8866,
      "theoretical_loss": 3.9845445231925827,
      "tokens_seen": 425409536
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00043993981945837516,
      "loss": 3.0092,
      "theoretical_loss": 3.984476564251061,
      "tokens_seen": 425475072
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00043992978936810434,
      "loss": 3.1991,
      "theoretical_loss": 3.9844086187069108,
      "tokens_seen": 425540608
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004399197592778335,
      "loss": 3.1595,
      "theoretical_loss": 3.9843406865554267,
      "tokens_seen": 425606144
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004399097291875627,
      "loss": 3.1975,
      "theoretical_loss": 3.984272767791908,
      "tokens_seen": 425671680
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004398996990972919,
      "loss": 2.9193,
      "theoretical_loss": 3.984204862411657,
      "tokens_seen": 425737216
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004398896690070211,
      "loss": 2.92,
      "theoretical_loss": 3.9841369704099745,
      "tokens_seen": 425802752
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00043987963891675024,
      "loss": 2.9934,
      "theoretical_loss": 3.9840690917821675,
      "tokens_seen": 425868288
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004398696088264795,
      "loss": 3.1759,
      "theoretical_loss": 3.984001226523545,
      "tokens_seen": 425933824
    },
    {
      "epoch": 1.04,
      "objective/train/docs_used": 530759,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.535398006439209,
      "objective/train/theoretical_loss": 3.983950336350283,
      "objective/train/tokens_used": 446442976,
      "theoretical_loss": 3.983950336350283,
      "tokens_seen": 425982976
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004398595787362086,
      "loss": 2.9939,
      "theoretical_loss": 3.983933374629416,
      "tokens_seen": 425999360
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00043984954864593784,
      "loss": 3.0605,
      "theoretical_loss": 3.9838655360950925,
      "tokens_seen": 426064896
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.000439839518555667,
      "loss": 3.1724,
      "theoretical_loss": 3.983797710915892,
      "tokens_seen": 426130432
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004398294884653962,
      "loss": 2.8873,
      "theoretical_loss": 3.9837298990871304,
      "tokens_seen": 426195968
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004398194583751254,
      "loss": 2.9109,
      "theoretical_loss": 3.9836621006041284,
      "tokens_seen": 426261504
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004398094282848546,
      "loss": 3.051,
      "theoretical_loss": 3.9835943154622075,
      "tokens_seen": 426327040
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00043979939819458375,
      "loss": 2.8254,
      "theoretical_loss": 3.9835265436566925,
      "tokens_seen": 426392576
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.000439789368104313,
      "loss": 3.2306,
      "theoretical_loss": 3.9834587851829104,
      "tokens_seen": 426458112
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004397793380140421,
      "loss": 3.1756,
      "theoretical_loss": 3.9833910400361905,
      "tokens_seen": 426523648
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00043976930792377134,
      "loss": 2.9631,
      "theoretical_loss": 3.9833233082118644,
      "tokens_seen": 426589184
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004397592778335005,
      "loss": 3.1489,
      "theoretical_loss": 3.9832555897052666,
      "tokens_seen": 426654720
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004397492477432297,
      "loss": 3.0471,
      "theoretical_loss": 3.9831878845117323,
      "tokens_seen": 426720256
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004397392176529589,
      "loss": 3.0495,
      "theoretical_loss": 3.9831201926266018,
      "tokens_seen": 426785792
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00043972918756268807,
      "loss": 3.1702,
      "theoretical_loss": 3.983052514045215,
      "tokens_seen": 426851328
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00043971915747241725,
      "loss": 2.9261,
      "theoretical_loss": 3.982984848762915,
      "tokens_seen": 426916864
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004397091273821465,
      "loss": 3.0344,
      "theoretical_loss": 3.982917196775049,
      "tokens_seen": 426982400
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004396990972918756,
      "loss": 3.1101,
      "theoretical_loss": 3.9828495580769636,
      "tokens_seen": 427047936
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00043968906720160485,
      "loss": 2.8813,
      "theoretical_loss": 3.98278193266401,
      "tokens_seen": 427113472
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00043967903711133403,
      "loss": 2.9827,
      "theoretical_loss": 3.9827143205315405,
      "tokens_seen": 427179008
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004396690070210632,
      "loss": 2.8908,
      "theoretical_loss": 3.982646721674911,
      "tokens_seen": 427244544
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004396589769307924,
      "loss": 3.0251,
      "theoretical_loss": 3.9825791360894787,
      "tokens_seen": 427310080
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00043964894684052157,
      "loss": 2.885,
      "theoretical_loss": 3.9825115637706023,
      "tokens_seen": 427375616
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00043963891675025075,
      "loss": 2.9362,
      "theoretical_loss": 3.982444004713645,
      "tokens_seen": 427441152
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00043962888665998,
      "loss": 2.9634,
      "theoretical_loss": 3.982376458913971,
      "tokens_seen": 427506688
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004396188565697091,
      "loss": 3.0015,
      "theoretical_loss": 3.982308926366947,
      "tokens_seen": 427572224
    },
    {
      "epoch": 1.04,
      "objective/train/docs_used": 532253,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.377031087875366,
      "objective/train/theoretical_loss": 3.9822582856509445,
      "objective/train/tokens_used": 448081376,
      "theoretical_loss": 3.9822582856509445,
      "tokens_seen": 427621376
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00043960882647943835,
      "loss": 2.8872,
      "theoretical_loss": 3.982241407067942,
      "tokens_seen": 427637760
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004395987963891675,
      "loss": 3.1085,
      "theoretical_loss": 3.9821739010123274,
      "tokens_seen": 427703296
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004395887662988967,
      "loss": 3.1265,
      "theoretical_loss": 3.9821064081954773,
      "tokens_seen": 427768832
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004395787362086259,
      "loss": 2.9826,
      "theoretical_loss": 3.9820389286127678,
      "tokens_seen": 427834368
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004395687061183551,
      "loss": 3.0233,
      "theoretical_loss": 3.981971462259577,
      "tokens_seen": 427899904
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00043955867602808426,
      "loss": 2.8907,
      "theoretical_loss": 3.981904009131285,
      "tokens_seen": 427965440
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00043954864593781344,
      "loss": 2.9427,
      "theoretical_loss": 3.9818365692232764,
      "tokens_seen": 428030976
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004395386158475426,
      "loss": 2.6778,
      "theoretical_loss": 3.981769142530936,
      "tokens_seen": 428096512
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00043952858575727185,
      "loss": 3.1932,
      "theoretical_loss": 3.98170172904965,
      "tokens_seen": 428162048
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.000439518555667001,
      "loss": 2.9245,
      "theoretical_loss": 3.9816343287748106,
      "tokens_seen": 428227584
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004395085255767302,
      "loss": 2.8477,
      "theoretical_loss": 3.9815669417018085,
      "tokens_seen": 428293120
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004394984954864594,
      "loss": 3.1402,
      "theoretical_loss": 3.981499567826039,
      "tokens_seen": 428358656
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004394884653961886,
      "loss": 2.8769,
      "theoretical_loss": 3.981432207142899,
      "tokens_seen": 428424192
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00043947843530591776,
      "loss": 2.8358,
      "theoretical_loss": 3.9813648596477877,
      "tokens_seen": 428489728
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00043946840521564694,
      "loss": 3.0459,
      "theoretical_loss": 3.9812975253361067,
      "tokens_seen": 428555264
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004394583751253761,
      "loss": 3.0235,
      "theoretical_loss": 3.9812302042032597,
      "tokens_seen": 428620800
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00043944834503510536,
      "loss": 3.1087,
      "theoretical_loss": 3.981162896244653,
      "tokens_seen": 428686336
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004394383149448345,
      "loss": 3.0048,
      "theoretical_loss": 3.9810956014556944,
      "tokens_seen": 428751872
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004394282848545637,
      "loss": 3.1698,
      "theoretical_loss": 3.981028319831796,
      "tokens_seen": 428817408
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00043941825476429285,
      "loss": 3.0103,
      "theoretical_loss": 3.9809610513683698,
      "tokens_seen": 428882944
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004394082246740221,
      "loss": 3.2602,
      "theoretical_loss": 3.9808937960608315,
      "tokens_seen": 428948480
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00043939819458375126,
      "loss": 2.9979,
      "theoretical_loss": 3.9808265539045986,
      "tokens_seen": 429014016
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00043938816449348044,
      "loss": 2.995,
      "theoretical_loss": 3.9807593248950917,
      "tokens_seen": 429079552
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004393781344032096,
      "loss": 3.0497,
      "theoretical_loss": 3.980692109027732,
      "tokens_seen": 429145088
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004393681043129388,
      "loss": 3.0166,
      "theoretical_loss": 3.980624906297945,
      "tokens_seen": 429210624
    },
    {
      "epoch": 1.04,
      "objective/train/docs_used": 533007,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.870386838912964,
      "objective/train/theoretical_loss": 3.980574512869385,
      "objective/train/tokens_used": 449719776,
      "theoretical_loss": 3.980574512869385,
      "tokens_seen": 429259776
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.000439358074222668,
      "loss": 2.9974,
      "theoretical_loss": 3.9805577167011563,
      "tokens_seen": 429276160
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004393480441323972,
      "loss": 2.8626,
      "theoretical_loss": 3.980490540232797,
      "tokens_seen": 429341696
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00043933801404212635,
      "loss": 2.9723,
      "theoretical_loss": 3.9804233768882966,
      "tokens_seen": 429407232
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004393279839518556,
      "loss": 2.8119,
      "theoretical_loss": 3.9803562266630896,
      "tokens_seen": 429472768
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00043931795386158477,
      "loss": 2.9868,
      "theoretical_loss": 3.9802890895526124,
      "tokens_seen": 429538304
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00043930792377131395,
      "loss": 2.909,
      "theoretical_loss": 3.9802219655523023,
      "tokens_seen": 429603840
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00043929789368104313,
      "loss": 3.1078,
      "theoretical_loss": 3.9801548546576004,
      "tokens_seen": 429669376
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004392878635907723,
      "loss": 2.9401,
      "theoretical_loss": 3.9800877568639494,
      "tokens_seen": 429734912
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004392778335005015,
      "loss": 3.1711,
      "theoretical_loss": 3.9800206721667952,
      "tokens_seen": 429800448
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004392678034102307,
      "loss": 3.2001,
      "theoretical_loss": 3.979953600561584,
      "tokens_seen": 429865984
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00043925777331995985,
      "loss": 3.0095,
      "theoretical_loss": 3.979886542043766,
      "tokens_seen": 429931520
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004392477432296891,
      "loss": 3.0156,
      "theoretical_loss": 3.979819496608793,
      "tokens_seen": 429997056
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004392377131394182,
      "loss": 2.8932,
      "theoretical_loss": 3.9797524642521194,
      "tokens_seen": 430062592
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00043922768304914745,
      "loss": 3.1111,
      "theoretical_loss": 3.979685444969202,
      "tokens_seen": 430128128
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00043921765295887663,
      "loss": 3.0072,
      "theoretical_loss": 3.979618438755498,
      "tokens_seen": 430193664
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004392076228686058,
      "loss": 2.9538,
      "theoretical_loss": 3.9795514456064702,
      "tokens_seen": 430259200
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.000439197592778335,
      "loss": 3.0761,
      "theoretical_loss": 3.979484465517581,
      "tokens_seen": 430324736
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00043918756268806423,
      "loss": 2.9912,
      "theoretical_loss": 3.979417498484297,
      "tokens_seen": 430390272
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004391775325977934,
      "loss": 2.8828,
      "theoretical_loss": 3.9793505445020845,
      "tokens_seen": 430455808
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004391675025075226,
      "loss": 2.9818,
      "theoretical_loss": 3.9792836035664148,
      "tokens_seen": 430521344
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00043915747241725177,
      "loss": 3.1237,
      "theoretical_loss": 3.9792166756727596,
      "tokens_seen": 430586880
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00043914744232698095,
      "loss": 3.0967,
      "theoretical_loss": 3.9791497608165933,
      "tokens_seen": 430652416
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004391374122367102,
      "loss": 3.1895,
      "theoretical_loss": 3.9790828589933933,
      "tokens_seen": 430717952
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004391273821464393,
      "loss": 2.8996,
      "theoretical_loss": 3.9790159701986387,
      "tokens_seen": 430783488
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00043911735205616855,
      "loss": 3.0491,
      "theoretical_loss": 3.978949094427811,
      "tokens_seen": 430849024
    },
    {
      "epoch": 1.04,
      "objective/train/docs_used": 534288,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8886990547180176,
      "objective/train/theoretical_loss": 3.9788989461439246,
      "objective/train/tokens_used": 451358176,
      "theoretical_loss": 3.9788989461439246,
      "tokens_seen": 430898176
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004391073219658977,
      "loss": 2.7626,
      "theoretical_loss": 3.978882231676393,
      "tokens_seen": 430914560
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004390972918756269,
      "loss": 2.8555,
      "theoretical_loss": 3.9788153819398717,
      "tokens_seen": 430980096
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004390872617853561,
      "loss": 3.1994,
      "theoretical_loss": 3.978748545213734,
      "tokens_seen": 431045632
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004390772316950853,
      "loss": 3.0032,
      "theoretical_loss": 3.9786817214934715,
      "tokens_seen": 431111168
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00043906720160481446,
      "loss": 2.9658,
      "theoretical_loss": 3.9786149107745765,
      "tokens_seen": 431176704
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00043905717151454364,
      "loss": 3.1315,
      "theoretical_loss": 3.9785481130525433,
      "tokens_seen": 431242240
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004390471414242728,
      "loss": 3.1294,
      "theoretical_loss": 3.9784813283228693,
      "tokens_seen": 431307776
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00043903711133400205,
      "loss": 3.1198,
      "theoretical_loss": 3.978414556581054,
      "tokens_seen": 431373312
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004390270812437312,
      "loss": 3.0879,
      "theoretical_loss": 3.9783477978225994,
      "tokens_seen": 431438848
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004390170511534604,
      "loss": 3.0557,
      "theoretical_loss": 3.9782810520430085,
      "tokens_seen": 431504384
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004390070210631896,
      "loss": 2.8589,
      "theoretical_loss": 3.9782143192377886,
      "tokens_seen": 431569920
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004389969909729188,
      "loss": 2.7922,
      "theoretical_loss": 3.9781475994024467,
      "tokens_seen": 431635456
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00043898696088264796,
      "loss": 2.9462,
      "theoretical_loss": 3.9780808925324935,
      "tokens_seen": 431700992
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00043897693079237714,
      "loss": 3.2187,
      "theoretical_loss": 3.978014198623443,
      "tokens_seen": 431766528
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004389669007021063,
      "loss": 2.6678,
      "theoretical_loss": 3.977947517670809,
      "tokens_seen": 431832064
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00043895687061183556,
      "loss": 3.0396,
      "theoretical_loss": 3.9778808496701092,
      "tokens_seen": 431897600
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004389468405215647,
      "loss": 3.0952,
      "theoretical_loss": 3.9778141946168635,
      "tokens_seen": 431963136
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004389368104312939,
      "loss": 2.9717,
      "theoretical_loss": 3.9777475525065933,
      "tokens_seen": 432028672
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00043892678034102305,
      "loss": 3.0584,
      "theoretical_loss": 3.9776809233348223,
      "tokens_seen": 432094208
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004389167502507523,
      "loss": 2.9005,
      "theoretical_loss": 3.977614307097077,
      "tokens_seen": 432159744
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00043890672016048146,
      "loss": 2.6508,
      "theoretical_loss": 3.977547703788886,
      "tokens_seen": 432225280
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00043889669007021064,
      "loss": 3.1334,
      "theoretical_loss": 3.9774811134057795,
      "tokens_seen": 432290816
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004388866599799398,
      "loss": 3.1029,
      "theoretical_loss": 3.97741453594329,
      "tokens_seen": 432356352
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.000438876629889669,
      "loss": 2.8582,
      "theoretical_loss": 3.977347971396954,
      "tokens_seen": 432421888
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004388665997993982,
      "loss": 3.0586,
      "theoretical_loss": 3.977281419762307,
      "tokens_seen": 432487424
    },
    {
      "epoch": 1.04,
      "objective/train/docs_used": 535120,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.702747344970703,
      "objective/train/theoretical_loss": 3.9772315145069355,
      "objective/train/tokens_used": 452996576,
      "theoretical_loss": 3.9772315145069355,
      "tokens_seen": 432536576
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004388565697091274,
      "loss": 2.9321,
      "theoretical_loss": 3.9772148810348904,
      "tokens_seen": 432552960
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00043884653961885655,
      "loss": 2.9524,
      "theoretical_loss": 3.9771483552102445,
      "tokens_seen": 432618496
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004388365095285858,
      "loss": 2.8339,
      "theoretical_loss": 3.977081842283913,
      "tokens_seen": 432684032
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00043882647943831497,
      "loss": 2.6309,
      "theoretical_loss": 3.977015342251444,
      "tokens_seen": 432749568
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00043881644934804415,
      "loss": 2.8433,
      "theoretical_loss": 3.976948855108384,
      "tokens_seen": 432815104
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00043880641925777333,
      "loss": 3.1073,
      "theoretical_loss": 3.9768823808502836,
      "tokens_seen": 432880640
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004387963891675025,
      "loss": 2.8288,
      "theoretical_loss": 3.976815919472697,
      "tokens_seen": 432946176
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004387863590772317,
      "loss": 2.9447,
      "theoretical_loss": 3.976749470971178,
      "tokens_seen": 433011712
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004387763289869609,
      "loss": 2.929,
      "theoretical_loss": 3.976683035341284,
      "tokens_seen": 433077248
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00043876629889669005,
      "loss": 3.1009,
      "theoretical_loss": 3.9766166125785753,
      "tokens_seen": 433142784
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004387562688064193,
      "loss": 2.8277,
      "theoretical_loss": 3.976550202678612,
      "tokens_seen": 433208320
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004387462387161484,
      "loss": 2.8715,
      "theoretical_loss": 3.9764838056369594,
      "tokens_seen": 433273856
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00043873620862587765,
      "loss": 2.9434,
      "theoretical_loss": 3.9764174214491828,
      "tokens_seen": 433339392
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00043872617853560683,
      "loss": 3.0301,
      "theoretical_loss": 3.97635105011085,
      "tokens_seen": 433404928
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.000438716148445336,
      "loss": 2.9667,
      "theoretical_loss": 3.976284691617532,
      "tokens_seen": 433470464
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004387061183550652,
      "loss": 2.8421,
      "theoretical_loss": 3.976218345964801,
      "tokens_seen": 433536000
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00043869608826479443,
      "loss": 3.0121,
      "theoretical_loss": 3.9761520131482326,
      "tokens_seen": 433601536
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00043868605817452356,
      "loss": 3.2531,
      "theoretical_loss": 3.976085693163403,
      "tokens_seen": 433667072
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004386760280842528,
      "loss": 3.0608,
      "theoretical_loss": 3.9760193860058917,
      "tokens_seen": 433732608
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004386659979939819,
      "loss": 3.1406,
      "theoretical_loss": 3.97595309167128,
      "tokens_seen": 433798144
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00043865596790371115,
      "loss": 2.7294,
      "theoretical_loss": 3.975886810155152,
      "tokens_seen": 433863680
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00043864593781344033,
      "loss": 2.9871,
      "theoretical_loss": 3.9758205414530923,
      "tokens_seen": 433929216
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004386359077231695,
      "loss": 2.9434,
      "theoretical_loss": 3.9757542855606895,
      "tokens_seen": 433994752
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004386258776328987,
      "loss": 2.856,
      "theoretical_loss": 3.975688042473534,
      "tokens_seen": 434060288
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004386158475426279,
      "loss": 3.0601,
      "theoretical_loss": 3.975621812187218,
      "tokens_seen": 434125824
    },
    {
      "epoch": 1.04,
      "objective/train/docs_used": 536351,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.160634756088257,
      "objective/train/theoretical_loss": 3.9755721478703814,
      "objective/train/tokens_used": 454634976,
      "theoretical_loss": 3.9755721478703814,
      "tokens_seen": 434174976
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00043860581745235706,
      "loss": 3.0205,
      "theoretical_loss": 3.9755555946973358,
      "tokens_seen": 434191360
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004385957873620863,
      "loss": 2.9008,
      "theoretical_loss": 3.975489389999484,
      "tokens_seen": 434256896
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004385857572718154,
      "loss": 2.9809,
      "theoretical_loss": 3.9754231980892616,
      "tokens_seen": 434322432
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00043857572718154466,
      "loss": 3.0585,
      "theoretical_loss": 3.97535701896227,
      "tokens_seen": 434387968
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004385656970912738,
      "loss": 3.105,
      "theoretical_loss": 3.9752908526141115,
      "tokens_seen": 434453504
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.000438555667001003,
      "loss": 3.0696,
      "theoretical_loss": 3.9752246990403926,
      "tokens_seen": 434519040
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004385456369107322,
      "loss": 3.1542,
      "theoretical_loss": 3.9751585582367195,
      "tokens_seen": 434584576
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004385356068204614,
      "loss": 2.8699,
      "theoretical_loss": 3.9750924301987034,
      "tokens_seen": 434650112
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00043852557673019056,
      "loss": 3.0068,
      "theoretical_loss": 3.975026314921955,
      "tokens_seen": 434715648
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004385155466399198,
      "loss": 2.9832,
      "theoretical_loss": 3.974960212402089,
      "tokens_seen": 434781184
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004385055165496489,
      "loss": 2.9908,
      "theoretical_loss": 3.974894122634722,
      "tokens_seen": 434846720
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00043849548645937816,
      "loss": 2.838,
      "theoretical_loss": 3.9748280456154714,
      "tokens_seen": 434912256
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004384854563691073,
      "loss": 3.1155,
      "theoretical_loss": 3.9747619813399586,
      "tokens_seen": 434977792
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004384754262788365,
      "loss": 3.0884,
      "theoretical_loss": 3.974695929803806,
      "tokens_seen": 435043328
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004384653961885657,
      "loss": 3.1434,
      "theoretical_loss": 3.9746298910026385,
      "tokens_seen": 435108864
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004384553660982949,
      "loss": 3.0162,
      "theoretical_loss": 3.974563864932083,
      "tokens_seen": 435174400
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00043844533600802407,
      "loss": 2.9452,
      "theoretical_loss": 3.9744978515877696,
      "tokens_seen": 435239936
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00043843530591775325,
      "loss": 3.0573,
      "theoretical_loss": 3.9744318509653285,
      "tokens_seen": 435305472
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004384252758274825,
      "loss": 3.0258,
      "theoretical_loss": 3.9743658630603935,
      "tokens_seen": 435371008
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00043841524573721166,
      "loss": 2.9923,
      "theoretical_loss": 3.9742998878686007,
      "tokens_seen": 435436544
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00043840521564694084,
      "loss": 3.0141,
      "theoretical_loss": 3.974233925385588,
      "tokens_seen": 435502080
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00043839518555667,
      "loss": 2.7964,
      "theoretical_loss": 3.974167975606995,
      "tokens_seen": 435567616
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004383851554663992,
      "loss": 2.9255,
      "theoretical_loss": 3.974102038528464,
      "tokens_seen": 435633152
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004383751253761284,
      "loss": 2.8229,
      "theoretical_loss": 3.9740361141456395,
      "tokens_seen": 435698688
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004383650952858576,
      "loss": 3.1127,
      "theoretical_loss": 3.9739702024541677,
      "tokens_seen": 435764224
    },
    {
      "epoch": 1.04,
      "objective/train/docs_used": 537149,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.811138868331909,
      "objective/train/theoretical_loss": 3.9739207770116467,
      "objective/train/tokens_used": 456273376,
      "theoretical_loss": 3.9739207770116467,
      "tokens_seen": 435813376
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00043835506519558675,
      "loss": 2.8946,
      "theoretical_loss": 3.9739043034496975,
      "tokens_seen": 435829760
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.000438345035105316,
      "loss": 2.9828,
      "theoretical_loss": 3.9738384171278787,
      "tokens_seen": 435895296
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00043833500501504517,
      "loss": 3.0393,
      "theoretical_loss": 3.9737725434843654,
      "tokens_seen": 435960832
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00043832497492477435,
      "loss": 2.9571,
      "theoretical_loss": 3.9737066825148126,
      "tokens_seen": 436026368
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00043831494483450353,
      "loss": 3.052,
      "theoretical_loss": 3.9736408342148763,
      "tokens_seen": 436091904
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004383049147442327,
      "loss": 2.8974,
      "theoretical_loss": 3.973574998580217,
      "tokens_seen": 436157440
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004382948846539619,
      "loss": 3.1101,
      "theoretical_loss": 3.973509175606495,
      "tokens_seen": 436222976
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004382848545636911,
      "loss": 2.8409,
      "theoretical_loss": 3.9734433652893753,
      "tokens_seen": 436288512
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00043827482447342025,
      "loss": 3.0663,
      "theoretical_loss": 3.9733775676245227,
      "tokens_seen": 436354048
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004382647943831495,
      "loss": 3.0505,
      "theoretical_loss": 3.973311782607605,
      "tokens_seen": 436419584
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004382547642928786,
      "loss": 3.0826,
      "theoretical_loss": 3.973246010234292,
      "tokens_seen": 436485120
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00043824473420260785,
      "loss": 3.0302,
      "theoretical_loss": 3.973180250500257,
      "tokens_seen": 436550656
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00043823470411233703,
      "loss": 2.9014,
      "theoretical_loss": 3.973114503401174,
      "tokens_seen": 436616192
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004382246740220662,
      "loss": 3.0467,
      "theoretical_loss": 3.973048768932718,
      "tokens_seen": 436681728
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004382146439317954,
      "loss": 3.1691,
      "theoretical_loss": 3.972983047090569,
      "tokens_seen": 436747264
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00043820461384152463,
      "loss": 2.8118,
      "theoretical_loss": 3.972917337870407,
      "tokens_seen": 436812800
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00043819458375125376,
      "loss": 2.9634,
      "theoretical_loss": 3.9728516412679147,
      "tokens_seen": 436878336
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.000438184553660983,
      "loss": 3.2248,
      "theoretical_loss": 3.9727859572787763,
      "tokens_seen": 436943872
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004381745235707121,
      "loss": 2.7478,
      "theoretical_loss": 3.9727202858986805,
      "tokens_seen": 437009408
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00043816449348044135,
      "loss": 2.9089,
      "theoretical_loss": 3.972654627123316,
      "tokens_seen": 437074944
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00043815446339017053,
      "loss": 3.0526,
      "theoretical_loss": 3.972588980948373,
      "tokens_seen": 437140480
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004381444332998997,
      "loss": 3.0026,
      "theoretical_loss": 3.9725233473695454,
      "tokens_seen": 437206016
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004381344032096289,
      "loss": 2.8952,
      "theoretical_loss": 3.9724577263825296,
      "tokens_seen": 437271552
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004381243731193581,
      "loss": 2.9139,
      "theoretical_loss": 3.9723921179830217,
      "tokens_seen": 437337088
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00043811434302908726,
      "loss": 2.9151,
      "theoretical_loss": 3.972326522166722,
      "tokens_seen": 437402624
    },
    {
      "epoch": 1.04,
      "objective/train/docs_used": 538542,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.2445921897888184,
      "objective/train/theoretical_loss": 3.972277333559642,
      "objective/train/tokens_used": 457911776,
      "theoretical_loss": 3.972277333559642,
      "tokens_seen": 437451776
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004381043129388165,
      "loss": 3.0233,
      "theoretical_loss": 3.972260938929333,
      "tokens_seen": 437468160
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004380942828485456,
      "loss": 2.7278,
      "theoretical_loss": 3.972195368266558,
      "tokens_seen": 437533696
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00043808425275827486,
      "loss": 3.0123,
      "theoretical_loss": 3.9721298101741027,
      "tokens_seen": 437599232
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.000438074222668004,
      "loss": 2.9663,
      "theoretical_loss": 3.972064264647676,
      "tokens_seen": 437664768
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004380641925777332,
      "loss": 3.1004,
      "theoretical_loss": 3.971998731682988,
      "tokens_seen": 437730304
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004380541624874624,
      "loss": 2.9891,
      "theoretical_loss": 3.9719332112757515,
      "tokens_seen": 437795840
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004380441323971916,
      "loss": 3.0848,
      "theoretical_loss": 3.9718677034216796,
      "tokens_seen": 437861376
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00043803410230692076,
      "loss": 2.9775,
      "theoretical_loss": 3.97180220811649,
      "tokens_seen": 437926912
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00043802407221665,
      "loss": 2.7554,
      "theoretical_loss": 3.9717367253559015,
      "tokens_seen": 437992448
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004380140421263791,
      "loss": 3.0661,
      "theoretical_loss": 3.971671255135634,
      "tokens_seen": 438057984
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00043800401203610836,
      "loss": 3.0481,
      "theoretical_loss": 3.971605797451411,
      "tokens_seen": 438123520
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004379939819458375,
      "loss": 2.7857,
      "theoretical_loss": 3.971540352298958,
      "tokens_seen": 438189056
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004379839518555667,
      "loss": 3.0102,
      "theoretical_loss": 3.971474919674001,
      "tokens_seen": 438254592
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004379739217652959,
      "loss": 2.9069,
      "theoretical_loss": 3.97140949957227,
      "tokens_seen": 438320128
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004379638916750251,
      "loss": 3.0839,
      "theoretical_loss": 3.9713440919894962,
      "tokens_seen": 438385664
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00043795386158475427,
      "loss": 2.8571,
      "theoretical_loss": 3.971278696921412,
      "tokens_seen": 438451200
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00043794383149448345,
      "loss": 3.0024,
      "theoretical_loss": 3.971213314363754,
      "tokens_seen": 438516736
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00043793380140421263,
      "loss": 2.8872,
      "theoretical_loss": 3.9711479443122593,
      "tokens_seen": 438582272
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00043792377131394186,
      "loss": 3.0489,
      "theoretical_loss": 3.9710825867626673,
      "tokens_seen": 438647808
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.000437913741223671,
      "loss": 2.8011,
      "theoretical_loss": 3.9710172417107206,
      "tokens_seen": 438713344
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004379037111334002,
      "loss": 2.8608,
      "theoretical_loss": 3.970951909152162,
      "tokens_seen": 438778880
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00043789368104312935,
      "loss": 3.0229,
      "theoretical_loss": 3.9708865890827383,
      "tokens_seen": 438844416
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004378836509528586,
      "loss": 3.0305,
      "theoretical_loss": 3.9708212814981962,
      "tokens_seen": 438909952
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00043787362086258777,
      "loss": 2.5924,
      "theoretical_loss": 3.970755986394287,
      "tokens_seen": 438975488
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00043786359077231695,
      "loss": 2.9254,
      "theoretical_loss": 3.970690703766763,
      "tokens_seen": 439041024
    },
    {
      "epoch": 1.04,
      "objective/train/docs_used": 539178,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.423835277557373,
      "objective/train/theoretical_loss": 3.9706417499811923,
      "objective/train/tokens_used": 459550176,
      "theoretical_loss": 3.9706417499811923,
      "tokens_seen": 439090176
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00043785356068204613,
      "loss": 2.8201,
      "theoretical_loss": 3.970625433611377,
      "tokens_seen": 439106560
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00043784353059177537,
      "loss": 3.0262,
      "theoretical_loss": 3.9705601759238864,
      "tokens_seen": 439172096
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004378335005015045,
      "loss": 3.1224,
      "theoretical_loss": 3.97049493070005,
      "tokens_seen": 439237632
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00043782347041123373,
      "loss": 3.0464,
      "theoretical_loss": 3.970429697935627,
      "tokens_seen": 439303168
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00043781344032096286,
      "loss": 2.9987,
      "theoretical_loss": 3.9703644776263802,
      "tokens_seen": 439368704
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004378034102306921,
      "loss": 3.0031,
      "theoretical_loss": 3.9702992697680752,
      "tokens_seen": 439434240
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00043779338014042127,
      "loss": 2.8794,
      "theoretical_loss": 3.970234074356477,
      "tokens_seen": 439499776
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00043778335005015045,
      "loss": 3.1284,
      "theoretical_loss": 3.9701688913873565,
      "tokens_seen": 439565312
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00043777331995987963,
      "loss": 2.988,
      "theoretical_loss": 3.970103720856483,
      "tokens_seen": 439630848
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004377632898696088,
      "loss": 2.8583,
      "theoretical_loss": 3.9700385627596297,
      "tokens_seen": 439696384
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.000437753259779338,
      "loss": 3.0441,
      "theoretical_loss": 3.969973417092571,
      "tokens_seen": 439761920
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00043774322968906723,
      "loss": 2.8557,
      "theoretical_loss": 3.969908283851085,
      "tokens_seen": 439827456
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00043773319959879636,
      "loss": 2.9863,
      "theoretical_loss": 3.9698431630309505,
      "tokens_seen": 439892992
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004377231695085256,
      "loss": 3.1955,
      "theoretical_loss": 3.969778054627948,
      "tokens_seen": 439958528
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004377131394182547,
      "loss": 2.8304,
      "theoretical_loss": 3.969712958637862,
      "tokens_seen": 440024064
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00043770310932798396,
      "loss": 2.9701,
      "theoretical_loss": 3.969647875056476,
      "tokens_seen": 440089600
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00043769307923771314,
      "loss": 2.954,
      "theoretical_loss": 3.9695828038795784,
      "tokens_seen": 440155136
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004376830491474423,
      "loss": 2.7975,
      "theoretical_loss": 3.969517745102958,
      "tokens_seen": 440220672
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00043767301905717155,
      "loss": 2.7506,
      "theoretical_loss": 3.9694526987224075,
      "tokens_seen": 440286208
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00043766298896690073,
      "loss": 2.9735,
      "theoretical_loss": 3.9693876647337185,
      "tokens_seen": 440351744
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004376529588766299,
      "loss": 2.7676,
      "theoretical_loss": 3.969322643132688,
      "tokens_seen": 440417280
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004376429287863591,
      "loss": 2.9377,
      "theoretical_loss": 3.969257633915113,
      "tokens_seen": 440482816
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004376328986960883,
      "loss": 2.7283,
      "theoretical_loss": 3.9691926370767936,
      "tokens_seen": 440548352
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00043762286860581746,
      "loss": 2.9292,
      "theoretical_loss": 3.969127652613531,
      "tokens_seen": 440613888
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004376128385155467,
      "loss": 3.1896,
      "theoretical_loss": 3.9690626805211293,
      "tokens_seen": 440679424
    },
    {
      "epoch": 1.04,
      "objective/train/docs_used": 540524,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.3463034629821777,
      "objective/train/theoretical_loss": 3.9690139595676817,
      "objective/train/tokens_used": 461188576,
      "theoretical_loss": 3.9690139595676817,
      "tokens_seen": 440728576
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004376028084252758,
      "loss": 3.082,
      "theoretical_loss": 3.9689977207953935,
      "tokens_seen": 440744960
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00043759277833500506,
      "loss": 2.7782,
      "theoretical_loss": 3.9689327734321327,
      "tokens_seen": 440810496
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004375827482447342,
      "loss": 3.0849,
      "theoretical_loss": 3.9688678384271556,
      "tokens_seen": 440876032
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004375727181544634,
      "loss": 3.0616,
      "theoretical_loss": 3.9688029157762745,
      "tokens_seen": 440941568
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004375626880641926,
      "loss": 3.1015,
      "theoretical_loss": 3.968738005475304,
      "tokens_seen": 441007104
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004375526579739218,
      "loss": 2.913,
      "theoretical_loss": 3.9686731075200594,
      "tokens_seen": 441072640
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00043754262788365096,
      "loss": 2.9542,
      "theoretical_loss": 3.9686082219063588,
      "tokens_seen": 441138176
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004375325977933802,
      "loss": 2.8682,
      "theoretical_loss": 3.968543348630022,
      "tokens_seen": 441203712
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004375225677031093,
      "loss": 2.7976,
      "theoretical_loss": 3.968478487686872,
      "tokens_seen": 441269248
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00043751253761283856,
      "loss": 2.8125,
      "theoretical_loss": 3.9684136390727325,
      "tokens_seen": 441334784
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004375025075225677,
      "loss": 2.6787,
      "theoretical_loss": 3.9683488027834293,
      "tokens_seen": 441400320
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004374924774322969,
      "loss": 2.9182,
      "theoretical_loss": 3.9682839788147906,
      "tokens_seen": 441465856
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004374824473420261,
      "loss": 2.6642,
      "theoretical_loss": 3.9682191671626477,
      "tokens_seen": 441531392
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004374724172517553,
      "loss": 2.9396,
      "theoretical_loss": 3.968154367822832,
      "tokens_seen": 441596928
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00043746238716148447,
      "loss": 2.8495,
      "theoretical_loss": 3.968089580791178,
      "tokens_seen": 441662464
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00043745235707121365,
      "loss": 3.2537,
      "theoretical_loss": 3.9680248060635215,
      "tokens_seen": 441728000
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00043744232698094283,
      "loss": 3.1128,
      "theoretical_loss": 3.967960043635702,
      "tokens_seen": 441793536
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00043743229689067206,
      "loss": 2.9745,
      "theoretical_loss": 3.9678952935035587,
      "tokens_seen": 441859072
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004374222668004012,
      "loss": 2.9838,
      "theoretical_loss": 3.967830555662935,
      "tokens_seen": 441924608
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004374122367101304,
      "loss": 3.0596,
      "theoretical_loss": 3.9677658301096743,
      "tokens_seen": 441990144
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00043740220661985955,
      "loss": 3.0076,
      "theoretical_loss": 3.967701116839624,
      "tokens_seen": 442055680
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004373921765295888,
      "loss": 3.0611,
      "theoretical_loss": 3.9676364158486326,
      "tokens_seen": 442121216
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00043738214643931797,
      "loss": 2.8529,
      "theoretical_loss": 3.9675717271325492,
      "tokens_seen": 442186752
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00043737211634904715,
      "loss": 2.7567,
      "theoretical_loss": 3.967507050687228,
      "tokens_seen": 442252288
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00043736208625877633,
      "loss": 3.0194,
      "theoretical_loss": 3.967442386508522,
      "tokens_seen": 442317824
    },
    {
      "epoch": 1.04,
      "objective/train/docs_used": 541315,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.549229145050049,
      "objective/train/theoretical_loss": 3.9673938964219673,
      "objective/train/tokens_used": 462826976,
      "theoretical_loss": 3.9673938964219673,
      "tokens_seen": 442366976
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00043735205616850557,
      "loss": 2.6802,
      "theoretical_loss": 3.967377734592289,
      "tokens_seen": 442383360
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004373420260782347,
      "loss": 2.9142,
      "theoretical_loss": 3.9673130949343873,
      "tokens_seen": 442448896
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00043733199598796393,
      "loss": 3.0676,
      "theoretical_loss": 3.967248467530677,
      "tokens_seen": 442514432
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00043732196589769306,
      "loss": 3.0618,
      "theoretical_loss": 3.967183852377021,
      "tokens_seen": 442579968
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004373119358074223,
      "loss": 3.0316,
      "theoretical_loss": 3.9671192494692837,
      "tokens_seen": 442645504
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00043730190571715147,
      "loss": 3.0938,
      "theoretical_loss": 3.9670546588033315,
      "tokens_seen": 442711040
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00043729187562688065,
      "loss": 2.8994,
      "theoretical_loss": 3.9669900803750338,
      "tokens_seen": 442776576
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00043728184553660983,
      "loss": 3.0618,
      "theoretical_loss": 3.96692551418026,
      "tokens_seen": 442842112
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.000437271815446339,
      "loss": 3.1426,
      "theoretical_loss": 3.966860960214883,
      "tokens_seen": 442907648
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004372617853560682,
      "loss": 2.9978,
      "theoretical_loss": 3.966796418474779,
      "tokens_seen": 442973184
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00043725175526579743,
      "loss": 2.6247,
      "theoretical_loss": 3.966731888955823,
      "tokens_seen": 443038720
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00043724172517552656,
      "loss": 2.8093,
      "theoretical_loss": 3.966667371653893,
      "tokens_seen": 443104256
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004372316950852558,
      "loss": 2.9878,
      "theoretical_loss": 3.9666028665648714,
      "tokens_seen": 443169792
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004372216649949849,
      "loss": 2.8664,
      "theoretical_loss": 3.96653837368464,
      "tokens_seen": 443235328
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00043721163490471416,
      "loss": 2.9273,
      "theoretical_loss": 3.9664738930090833,
      "tokens_seen": 443300864
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00043720160481444334,
      "loss": 2.9889,
      "theoretical_loss": 3.9664094245340875,
      "tokens_seen": 443366400
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004371915747241725,
      "loss": 2.9439,
      "theoretical_loss": 3.9663449682555423,
      "tokens_seen": 443431936
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004371815446339017,
      "loss": 3.0401,
      "theoretical_loss": 3.966280524169337,
      "tokens_seen": 443497472
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00043717151454363093,
      "loss": 3.0757,
      "theoretical_loss": 3.966216092271365,
      "tokens_seen": 443563008
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00043716148445336006,
      "loss": 2.9564,
      "theoretical_loss": 3.966151672557521,
      "tokens_seen": 443628544
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004371514543630893,
      "loss": 2.8941,
      "theoretical_loss": 3.9660872650237016,
      "tokens_seen": 443694080
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004371414242728184,
      "loss": 3.0021,
      "theoretical_loss": 3.966022869665804,
      "tokens_seen": 443759616
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00043713139418254766,
      "loss": 3.0318,
      "theoretical_loss": 3.96595848647973,
      "tokens_seen": 443825152
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00043712136409227684,
      "loss": 3.0716,
      "theoretical_loss": 3.9658941154613823,
      "tokens_seen": 443890688
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.000437111334002006,
      "loss": 2.9925,
      "theoretical_loss": 3.9658297566066647,
      "tokens_seen": 443956224
    },
    {
      "epoch": 1.04,
      "objective/train/docs_used": 542939,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.203843355178833,
      "objective/train/theoretical_loss": 3.965781495445546,
      "objective/train/tokens_used": 464465376,
      "theoretical_loss": 3.965781495445546,
      "tokens_seen": 444005376
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004371013039117352,
      "loss": 3.0367,
      "theoretical_loss": 3.9657654099114836,
      "tokens_seen": 444021760
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004370912738214644,
      "loss": 3.0506,
      "theoretical_loss": 3.965701075371748,
      "tokens_seen": 444087296
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00043708124373119356,
      "loss": 2.9351,
      "theoretical_loss": 3.965636752983368,
      "tokens_seen": 444152832
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004370712136409228,
      "loss": 2.8604,
      "theoretical_loss": 3.965572442742256,
      "tokens_seen": 444218368
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004370611835506519,
      "loss": 3.1753,
      "theoretical_loss": 3.9655081446443265,
      "tokens_seen": 444283904
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00043705115346038116,
      "loss": 2.8183,
      "theoretical_loss": 3.9654438586854965,
      "tokens_seen": 444349440
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004370411233701103,
      "loss": 3.1016,
      "theoretical_loss": 3.965379584861684,
      "tokens_seen": 444414976
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004370310932798395,
      "loss": 2.6801,
      "theoretical_loss": 3.965315323168808,
      "tokens_seen": 444480512
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004370210631895687,
      "loss": 2.9384,
      "theoretical_loss": 3.9652510736027926,
      "tokens_seen": 444546048
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004370110330992979,
      "loss": 2.8253,
      "theoretical_loss": 3.9651868361595612,
      "tokens_seen": 444611584
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00043700100300902707,
      "loss": 2.925,
      "theoretical_loss": 3.9651226108350395,
      "tokens_seen": 444677120
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004369909729187563,
      "loss": 2.9335,
      "theoretical_loss": 3.965058397625157,
      "tokens_seen": 444742656
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00043698094282848543,
      "loss": 2.9976,
      "theoretical_loss": 3.9649941965258435,
      "tokens_seen": 444808192
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00043697091273821467,
      "loss": 2.7717,
      "theoretical_loss": 3.9649300075330305,
      "tokens_seen": 444873728
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004369608826479438,
      "loss": 2.8541,
      "theoretical_loss": 3.964865830642653,
      "tokens_seen": 444939264
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.00043695085255767303,
      "loss": 3.0042,
      "theoretical_loss": 3.964801665850646,
      "tokens_seen": 445004800
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004369408224674022,
      "loss": 3.0289,
      "theoretical_loss": 3.964737513152949,
      "tokens_seen": 445070336
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004369307923771314,
      "loss": 3.0322,
      "theoretical_loss": 3.9646733725455006,
      "tokens_seen": 445135872
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0004369207622868606,
      "loss": 3.0371,
      "theoretical_loss": 3.964609244024243,
      "tokens_seen": 445201408
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043691073219658975,
      "loss": 2.8602,
      "theoretical_loss": 3.9645451275851205,
      "tokens_seen": 445266944
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.000436900702106319,
      "loss": 2.9065,
      "theoretical_loss": 3.9644810232240792,
      "tokens_seen": 445332480
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043689067201604817,
      "loss": 2.9665,
      "theoretical_loss": 3.9644169309370665,
      "tokens_seen": 445398016
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043688064192577735,
      "loss": 2.9056,
      "theoretical_loss": 3.964352850720032,
      "tokens_seen": 445463552
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043687061183550653,
      "loss": 2.8584,
      "theoretical_loss": 3.9642887825689286,
      "tokens_seen": 445529088
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043686058174523577,
      "loss": 2.9096,
      "theoretical_loss": 3.964224726479708,
      "tokens_seen": 445594624
    },
    {
      "epoch": 1.05,
      "objective/train/docs_used": 543579,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.705620765686035,
      "objective/train/theoretical_loss": 3.964176692325972,
      "objective/train/tokens_used": 466103776,
      "theoretical_loss": 3.964176692325972,
      "tokens_seen": 445643776
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004368505516549649,
      "loss": 3.0128,
      "theoretical_loss": 3.9641606824483278,
      "tokens_seen": 445660160
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043684052156469413,
      "loss": 3.0941,
      "theoretical_loss": 3.9640966504707444,
      "tokens_seen": 445725696
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043683049147442326,
      "loss": 2.9789,
      "theoretical_loss": 3.964032630542918,
      "tokens_seen": 445791232
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004368204613841525,
      "loss": 2.81,
      "theoretical_loss": 3.96396862266081,
      "tokens_seen": 445856768
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043681043129388167,
      "loss": 2.9512,
      "theoretical_loss": 3.9639046268203835,
      "tokens_seen": 445922304
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043680040120361085,
      "loss": 3.0781,
      "theoretical_loss": 3.963840643017604,
      "tokens_seen": 445987840
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043679037111334003,
      "loss": 2.7061,
      "theoretical_loss": 3.9637766712484384,
      "tokens_seen": 446053376
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004367803410230692,
      "loss": 3.125,
      "theoretical_loss": 3.9637127115088573,
      "tokens_seen": 446118912
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004367703109327984,
      "loss": 2.981,
      "theoretical_loss": 3.963648763794831,
      "tokens_seen": 446184448
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043676028084252763,
      "loss": 2.7527,
      "theoretical_loss": 3.9635848281023325,
      "tokens_seen": 446249984
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043675025075225676,
      "loss": 2.8724,
      "theoretical_loss": 3.9635209044273365,
      "tokens_seen": 446315520
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.000436740220661986,
      "loss": 2.9677,
      "theoretical_loss": 3.9634569927658214,
      "tokens_seen": 446381056
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004367301905717151,
      "loss": 2.9814,
      "theoretical_loss": 3.9633930931137655,
      "tokens_seen": 446446592
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043672016048144436,
      "loss": 2.9977,
      "theoretical_loss": 3.963329205467149,
      "tokens_seen": 446512128
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043671013039117354,
      "loss": 2.8533,
      "theoretical_loss": 3.9632653298219562,
      "tokens_seen": 446577664
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004367001003009027,
      "loss": 2.7676,
      "theoretical_loss": 3.96320146617417,
      "tokens_seen": 446643200
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004366900702106319,
      "loss": 3.0384,
      "theoretical_loss": 3.9631376145197796,
      "tokens_seen": 446708736
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043668004012036113,
      "loss": 3.0348,
      "theoretical_loss": 3.963073774854771,
      "tokens_seen": 446774272
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043667001003009026,
      "loss": 3.0659,
      "theoretical_loss": 3.9630099471751365,
      "tokens_seen": 446839808
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004366599799398195,
      "loss": 3.0177,
      "theoretical_loss": 3.962946131476868,
      "tokens_seen": 446905344
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004366499498495486,
      "loss": 2.9298,
      "theoretical_loss": 3.9628823277559597,
      "tokens_seen": 446970880
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043663991975927786,
      "loss": 2.9698,
      "theoretical_loss": 3.9628185360084087,
      "tokens_seen": 447036416
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043662988966900704,
      "loss": 3.0568,
      "theoretical_loss": 3.962754756230213,
      "tokens_seen": 447101952
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004366198595787362,
      "loss": 2.9263,
      "theoretical_loss": 3.9626909884173727,
      "tokens_seen": 447167488
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004366098294884654,
      "loss": 3.1906,
      "theoretical_loss": 3.9626272325658896,
      "tokens_seen": 447233024
    },
    {
      "epoch": 1.05,
      "objective/train/docs_used": 544310,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.0421924591064453,
      "objective/train/theoretical_loss": 3.962579423524514,
      "objective/train/tokens_used": 467742176,
      "theoretical_loss": 3.962579423524514,
      "tokens_seen": 447282176
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004365997993981946,
      "loss": 3.0931,
      "theoretical_loss": 3.9625634886717678,
      "tokens_seen": 447298560
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043658976930792376,
      "loss": 2.9716,
      "theoretical_loss": 3.962499756731014,
      "tokens_seen": 447364096
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.000436579739217653,
      "loss": 3.0392,
      "theoretical_loss": 3.9624360367396347,
      "tokens_seen": 447429632
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043656970912738213,
      "loss": 2.9594,
      "theoretical_loss": 3.9623723286936414,
      "tokens_seen": 447495168
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043655967903711136,
      "loss": 3.1369,
      "theoretical_loss": 3.9623086325890444,
      "tokens_seen": 447560704
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004365496489468405,
      "loss": 2.9239,
      "theoretical_loss": 3.9622449484218585,
      "tokens_seen": 447626240
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004365396188565697,
      "loss": 3.2243,
      "theoretical_loss": 3.9621812761880975,
      "tokens_seen": 447691776
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004365295887662989,
      "loss": 2.8872,
      "theoretical_loss": 3.9621176158837814,
      "tokens_seen": 447757312
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004365195586760281,
      "loss": 3.1265,
      "theoretical_loss": 3.9620539675049278,
      "tokens_seen": 447822848
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043650952858575727,
      "loss": 2.9101,
      "theoretical_loss": 3.9619903310475575,
      "tokens_seen": 447888384
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004364994984954865,
      "loss": 2.9973,
      "theoretical_loss": 3.9619267065076955,
      "tokens_seen": 447953920
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043648946840521563,
      "loss": 2.9611,
      "theoretical_loss": 3.961863093881366,
      "tokens_seen": 448019456
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043647943831494487,
      "loss": 3.2598,
      "theoretical_loss": 3.9617994931645955,
      "tokens_seen": 448084992
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.000436469408224674,
      "loss": 2.9044,
      "theoretical_loss": 3.9617359043534135,
      "tokens_seen": 448150528
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043645937813440323,
      "loss": 2.9648,
      "theoretical_loss": 3.961672327443851,
      "tokens_seen": 448216064
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004364493480441324,
      "loss": 2.9154,
      "theoretical_loss": 3.96160876243194,
      "tokens_seen": 448281600
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004364393179538616,
      "loss": 2.9511,
      "theoretical_loss": 3.961545209313715,
      "tokens_seen": 448347136
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043642928786359077,
      "loss": 2.8572,
      "theoretical_loss": 3.961481668085214,
      "tokens_seen": 448412672
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043641925777331995,
      "loss": 3.2968,
      "theoretical_loss": 3.9614181387424745,
      "tokens_seen": 448478208
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043640922768304913,
      "loss": 2.8314,
      "theoretical_loss": 3.9613546212815365,
      "tokens_seen": 448543744
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043639919759277837,
      "loss": 2.9011,
      "theoretical_loss": 3.961291115698442,
      "tokens_seen": 448609280
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004363891675025075,
      "loss": 2.757,
      "theoretical_loss": 3.9612276219892366,
      "tokens_seen": 448674816
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043637913741223673,
      "loss": 2.8699,
      "theoretical_loss": 3.9611641401499647,
      "tokens_seen": 448740352
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004363691073219659,
      "loss": 3.0423,
      "theoretical_loss": 3.961100670176675,
      "tokens_seen": 448805888
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004363590772316951,
      "loss": 2.9675,
      "theoretical_loss": 3.961037212065418,
      "tokens_seen": 448871424
    },
    {
      "epoch": 1.05,
      "objective/train/docs_used": 544864,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.588275909423828,
      "objective/train/theoretical_loss": 3.9609896262640576,
      "objective/train/tokens_used": 469380576,
      "theoretical_loss": 3.9609896262640576,
      "tokens_seen": 448920576
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004363490471414243,
      "loss": 2.9927,
      "theoretical_loss": 3.960973765812244,
      "tokens_seen": 448936960
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043633901705115346,
      "loss": 3.0827,
      "theoretical_loss": 3.9609103314132064,
      "tokens_seen": 449002496
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043632898696088264,
      "loss": 3.0377,
      "theoretical_loss": 3.960846908864362,
      "tokens_seen": 449068032
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043631895687061187,
      "loss": 2.8994,
      "theoretical_loss": 3.9607834981617676,
      "tokens_seen": 449133568
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.000436308926780341,
      "loss": 2.9719,
      "theoretical_loss": 3.9607200993014824,
      "tokens_seen": 449199104
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043629889669007023,
      "loss": 2.9095,
      "theoretical_loss": 3.960656712279567,
      "tokens_seen": 449264640
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043628886659979936,
      "loss": 2.9354,
      "theoretical_loss": 3.960593337092086,
      "tokens_seen": 449330176
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004362788365095286,
      "loss": 3.012,
      "theoretical_loss": 3.9605299737351025,
      "tokens_seen": 449395712
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004362688064192578,
      "loss": 3.1365,
      "theoretical_loss": 3.9604666222046845,
      "tokens_seen": 449461248
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043625877632898696,
      "loss": 2.8716,
      "theoretical_loss": 3.9604032824968995,
      "tokens_seen": 449526784
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043624874623871614,
      "loss": 2.8419,
      "theoretical_loss": 3.960339954607819,
      "tokens_seen": 449592320
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004362387161484453,
      "loss": 3.0657,
      "theoretical_loss": 3.9602766385335153,
      "tokens_seen": 449657856
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004362286860581745,
      "loss": 3.0267,
      "theoretical_loss": 3.9602133342700623,
      "tokens_seen": 449723392
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043621865596790374,
      "loss": 3.0393,
      "theoretical_loss": 3.9601500418135367,
      "tokens_seen": 449788928
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043620862587763286,
      "loss": 3.1022,
      "theoretical_loss": 3.960086761160016,
      "tokens_seen": 449854464
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004361985957873621,
      "loss": 2.9125,
      "theoretical_loss": 3.960023492305581,
      "tokens_seen": 449920000
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004361885656970913,
      "loss": 2.9061,
      "theoretical_loss": 3.959960235246312,
      "tokens_seen": 449985536
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043617853560682046,
      "loss": 3.0702,
      "theoretical_loss": 3.9598969899782936,
      "tokens_seen": 450051072
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004361685055165497,
      "loss": 2.9372,
      "theoretical_loss": 3.9598337564976114,
      "tokens_seen": 450116608
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004361584754262788,
      "loss": 3.0303,
      "theoretical_loss": 3.9597705348003527,
      "tokens_seen": 450182144
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043614844533600806,
      "loss": 2.8191,
      "theoretical_loss": 3.959707324882607,
      "tokens_seen": 450247680
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043613841524573724,
      "loss": 3.0286,
      "theoretical_loss": 3.9596441267404647,
      "tokens_seen": 450313216
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004361283851554664,
      "loss": 2.9917,
      "theoretical_loss": 3.95958094037002,
      "tokens_seen": 450378752
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004361183550651956,
      "loss": 3.238,
      "theoretical_loss": 3.959517765767366,
      "tokens_seen": 450444288
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004361083249749248,
      "loss": 2.9412,
      "theoretical_loss": 3.9594546029286013,
      "tokens_seen": 450509824
    },
    {
      "epoch": 1.05,
      "objective/train/docs_used": 544864,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.3638360500335693,
      "objective/train/theoretical_loss": 3.9594072385172328,
      "objective/train/tokens_used": 471018976,
      "theoretical_loss": 3.9594072385172328,
      "tokens_seen": 450558976
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043609829488465397,
      "loss": 2.9778,
      "theoretical_loss": 3.9593914518498234,
      "tokens_seen": 450575360
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004360882647943832,
      "loss": 3.1076,
      "theoretical_loss": 3.9593283125271332,
      "tokens_seen": 450640896
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043607823470411233,
      "loss": 3.0533,
      "theoretical_loss": 3.959265184956633,
      "tokens_seen": 450706432
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043606820461384156,
      "loss": 3.0076,
      "theoretical_loss": 3.9592020691344265,
      "tokens_seen": 450771968
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004360581745235707,
      "loss": 3.2244,
      "theoretical_loss": 3.9591389650566207,
      "tokens_seen": 450837504
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004360481444332999,
      "loss": 3.0192,
      "theoretical_loss": 3.959075872719322,
      "tokens_seen": 450903040
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004360381143430291,
      "loss": 3.082,
      "theoretical_loss": 3.9590127921186413,
      "tokens_seen": 450968576
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004360280842527583,
      "loss": 3.1096,
      "theoretical_loss": 3.95894972325069,
      "tokens_seen": 451034112
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043601805416248747,
      "loss": 2.8023,
      "theoretical_loss": 3.9588866661115816,
      "tokens_seen": 451099648
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004360080240722167,
      "loss": 3.124,
      "theoretical_loss": 3.958823620697431,
      "tokens_seen": 451165184
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043599799398194583,
      "loss": 2.9392,
      "theoretical_loss": 3.958760587004355,
      "tokens_seen": 451230720
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043598796389167507,
      "loss": 3.0512,
      "theoretical_loss": 3.958697565028474,
      "tokens_seen": 451296256
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004359779338014042,
      "loss": 2.9383,
      "theoretical_loss": 3.958634554765908,
      "tokens_seen": 451361792
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043596790371113343,
      "loss": 3.0894,
      "theoretical_loss": 3.95857155621278,
      "tokens_seen": 451427328
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004359578736208626,
      "loss": 3.1975,
      "theoretical_loss": 3.958508569365214,
      "tokens_seen": 451492864
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004359478435305918,
      "loss": 3.0222,
      "theoretical_loss": 3.9584455942193366,
      "tokens_seen": 451558400
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043593781344032097,
      "loss": 2.9316,
      "theoretical_loss": 3.9583826307712764,
      "tokens_seen": 451623936
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043592778335005015,
      "loss": 3.0144,
      "theoretical_loss": 3.958319679017163,
      "tokens_seen": 451689472
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043591775325977933,
      "loss": 2.9815,
      "theoretical_loss": 3.9582567389531285,
      "tokens_seen": 451755008
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043590772316950857,
      "loss": 2.9248,
      "theoretical_loss": 3.9581938105753065,
      "tokens_seen": 451820544
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004358976930792377,
      "loss": 2.924,
      "theoretical_loss": 3.958130893879833,
      "tokens_seen": 451886080
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043588766298896693,
      "loss": 2.9269,
      "theoretical_loss": 3.9580679888628456,
      "tokens_seen": 451951616
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004358776328986961,
      "loss": 3.0172,
      "theoretical_loss": 3.958005095520483,
      "tokens_seen": 452017152
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004358676028084253,
      "loss": 3.0345,
      "theoretical_loss": 3.9579422138488862,
      "tokens_seen": 452082688
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004358575727181545,
      "loss": 3.2033,
      "theoretical_loss": 3.9578793438441986,
      "tokens_seen": 452148224
    },
    {
      "epoch": 1.05,
      "objective/train/docs_used": 545628,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9914121627807617,
      "objective/train/theoretical_loss": 3.9578321989947725,
      "objective/train/tokens_used": 472657376,
      "theoretical_loss": 3.9578321989947725,
      "tokens_seen": 452197376
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043584754262788366,
      "loss": 2.9125,
      "theoretical_loss": 3.9578164855025646,
      "tokens_seen": 452213760
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043583751253761284,
      "loss": 2.7379,
      "theoretical_loss": 3.9577536388201313,
      "tokens_seen": 452279296
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043582748244734207,
      "loss": 3.1926,
      "theoretical_loss": 3.957690803793047,
      "tokens_seen": 452344832
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004358174523570712,
      "loss": 3.1275,
      "theoretical_loss": 3.9576279804174614,
      "tokens_seen": 452410368
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043580742226680043,
      "loss": 3.0179,
      "theoretical_loss": 3.957565168689528,
      "tokens_seen": 452475904
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043579739217652956,
      "loss": 3.0056,
      "theoretical_loss": 3.9575023686053985,
      "tokens_seen": 452541440
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004357873620862588,
      "loss": 3.1728,
      "theoretical_loss": 3.957439580161231,
      "tokens_seen": 452606976
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.000435777331995988,
      "loss": 3.027,
      "theoretical_loss": 3.9573768033531813,
      "tokens_seen": 452672512
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043576730190571716,
      "loss": 2.8647,
      "theoretical_loss": 3.95731403817741,
      "tokens_seen": 452738048
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043575727181544634,
      "loss": 2.8473,
      "theoretical_loss": 3.9572512846300776,
      "tokens_seen": 452803584
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004357472417251755,
      "loss": 2.6547,
      "theoretical_loss": 3.9571885427073474,
      "tokens_seen": 452869120
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004357372116349047,
      "loss": 2.7046,
      "theoretical_loss": 3.9571258124053843,
      "tokens_seen": 452934656
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043572718154463394,
      "loss": 2.9029,
      "theoretical_loss": 3.957063093720355,
      "tokens_seen": 453000192
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043571715145436306,
      "loss": 3.0426,
      "theoretical_loss": 3.957000386648428,
      "tokens_seen": 453065728
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004357071213640923,
      "loss": 2.8844,
      "theoretical_loss": 3.9569376911857734,
      "tokens_seen": 453131264
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004356970912738215,
      "loss": 3.0411,
      "theoretical_loss": 3.9568750073285637,
      "tokens_seen": 453196800
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043568706118355066,
      "loss": 2.8895,
      "theoretical_loss": 3.9568123350729727,
      "tokens_seen": 453262336
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043567703109327984,
      "loss": 3.0225,
      "theoretical_loss": 3.956749674415176,
      "tokens_seen": 453327872
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.000435667001003009,
      "loss": 3.0718,
      "theoretical_loss": 3.9566870253513513,
      "tokens_seen": 453393408
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004356569709127382,
      "loss": 3.0213,
      "theoretical_loss": 3.956624387877678,
      "tokens_seen": 453458944
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043564694082246744,
      "loss": 3.0282,
      "theoretical_loss": 3.956561761990338,
      "tokens_seen": 453524480
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043563691073219657,
      "loss": 2.9475,
      "theoretical_loss": 3.9564991476855136,
      "tokens_seen": 453590016
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004356268806419258,
      "loss": 2.8315,
      "theoretical_loss": 3.956436544959389,
      "tokens_seen": 453655552
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043561685055165493,
      "loss": 3.0489,
      "theoretical_loss": 3.9563739538081517,
      "tokens_seen": 453721088
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043560682046138417,
      "loss": 2.9775,
      "theoretical_loss": 3.9563113742279907,
      "tokens_seen": 453786624
    },
    {
      "epoch": 1.05,
      "objective/train/docs_used": 546923,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7837202548980713,
      "objective/train/theoretical_loss": 3.956264447134096,
      "objective/train/tokens_used": 474295776,
      "theoretical_loss": 3.956264447134096,
      "tokens_seen": 453835776
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043559679037111335,
      "loss": 2.8202,
      "theoretical_loss": 3.956248806215095,
      "tokens_seen": 453852160
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043558676028084253,
      "loss": 2.867,
      "theoretical_loss": 3.9561862497656572,
      "tokens_seen": 453917696
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004355767301905717,
      "loss": 2.8822,
      "theoretical_loss": 3.9561237048758713,
      "tokens_seen": 453983232
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004355667001003009,
      "loss": 2.8726,
      "theoretical_loss": 3.9560611715419327,
      "tokens_seen": 454048768
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043555667001003007,
      "loss": 3.0744,
      "theoretical_loss": 3.955998649760039,
      "tokens_seen": 454114304
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004355466399197593,
      "loss": 2.7415,
      "theoretical_loss": 3.9559361395263895,
      "tokens_seen": 454179840
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043553660982948843,
      "loss": 2.862,
      "theoretical_loss": 3.9558736408371855,
      "tokens_seen": 454245376
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043552657973921767,
      "loss": 2.7807,
      "theoretical_loss": 3.9558111536886287,
      "tokens_seen": 454310912
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043551654964894685,
      "loss": 2.9495,
      "theoretical_loss": 3.955748678076925,
      "tokens_seen": 454376448
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043550651955867603,
      "loss": 3.0138,
      "theoretical_loss": 3.9556862139982805,
      "tokens_seen": 454441984
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004354964894684052,
      "loss": 2.9964,
      "theoretical_loss": 3.955623761448903,
      "tokens_seen": 454507520
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004354864593781344,
      "loss": 2.9618,
      "theoretical_loss": 3.955561320425004,
      "tokens_seen": 454573056
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004354764292878636,
      "loss": 2.9674,
      "theoretical_loss": 3.9554988909227933,
      "tokens_seen": 454638592
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004354663991975928,
      "loss": 2.9339,
      "theoretical_loss": 3.955436472938486,
      "tokens_seen": 454704128
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043545636910732194,
      "loss": 2.8827,
      "theoretical_loss": 3.9553740664682966,
      "tokens_seen": 454769664
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043544633901705117,
      "loss": 2.8341,
      "theoretical_loss": 3.9553116715084427,
      "tokens_seen": 454835200
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004354363089267803,
      "loss": 3.045,
      "theoretical_loss": 3.955249288055143,
      "tokens_seen": 454900736
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043542627883650953,
      "loss": 3.152,
      "theoretical_loss": 3.9551869161046187,
      "tokens_seen": 454966272
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043541624874623877,
      "loss": 2.8714,
      "theoretical_loss": 3.955124555653092,
      "tokens_seen": 455031808
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004354062186559679,
      "loss": 2.9861,
      "theoretical_loss": 3.9550622066967875,
      "tokens_seen": 455097344
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043539618856569713,
      "loss": 3.0678,
      "theoretical_loss": 3.9549998692319317,
      "tokens_seen": 455162880
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004353861584754263,
      "loss": 3.1052,
      "theoretical_loss": 3.9549375432547516,
      "tokens_seen": 455228416
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004353761283851555,
      "loss": 3.0939,
      "theoretical_loss": 3.9548752287614772,
      "tokens_seen": 455293952
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004353660982948847,
      "loss": 3.2468,
      "theoretical_loss": 3.9548129257483406,
      "tokens_seen": 455359488
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043535606820461386,
      "loss": 2.7976,
      "theoretical_loss": 3.9547506342115737,
      "tokens_seen": 455425024
    },
    {
      "epoch": 1.05,
      "objective/train/docs_used": 547608,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.1719532012939453,
      "objective/train/theoretical_loss": 3.9547039230881023,
      "objective/train/tokens_used": 475934176,
      "theoretical_loss": 3.9547039230881023,
      "tokens_seen": 455474176
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043534603811434304,
      "loss": 3.0834,
      "theoretical_loss": 3.954688354147413,
      "tokens_seen": 455490560
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043533600802407227,
      "loss": 2.8907,
      "theoretical_loss": 3.9546260855520945,
      "tokens_seen": 455556096
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004353259779338014,
      "loss": 3.0901,
      "theoretical_loss": 3.9545638284218567,
      "tokens_seen": 455621632
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043531594784353063,
      "loss": 2.9834,
      "theoretical_loss": 3.9545015827529406,
      "tokens_seen": 455687168
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043530591775325976,
      "loss": 2.9154,
      "theoretical_loss": 3.9544393485415874,
      "tokens_seen": 455752704
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.000435295887662989,
      "loss": 3.0887,
      "theoretical_loss": 3.954377125784042,
      "tokens_seen": 455818240
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004352858575727182,
      "loss": 2.8017,
      "theoretical_loss": 3.954314914476549,
      "tokens_seen": 455883776
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043527582748244736,
      "loss": 2.9674,
      "theoretical_loss": 3.9542527146153565,
      "tokens_seen": 455949312
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043526579739217654,
      "loss": 2.8448,
      "theoretical_loss": 3.954190526196714,
      "tokens_seen": 456014848
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004352557673019057,
      "loss": 2.9596,
      "theoretical_loss": 3.9541283492168717,
      "tokens_seen": 456080384
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004352457372116349,
      "loss": 3.2475,
      "theoretical_loss": 3.954066183672083,
      "tokens_seen": 456145920
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043523570712136414,
      "loss": 2.9562,
      "theoretical_loss": 3.9540040295586016,
      "tokens_seen": 456211456
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043522567703109326,
      "loss": 3.0325,
      "theoretical_loss": 3.9539418868726846,
      "tokens_seen": 456276992
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004352156469408225,
      "loss": 2.9111,
      "theoretical_loss": 3.95387975561059,
      "tokens_seen": 456342528
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004352056168505517,
      "loss": 3.1916,
      "theoretical_loss": 3.9538176357685764,
      "tokens_seen": 456408064
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043519558676028086,
      "loss": 2.9947,
      "theoretical_loss": 3.953755527342907,
      "tokens_seen": 456473600
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043518555667001004,
      "loss": 2.9115,
      "theoretical_loss": 3.953693430329844,
      "tokens_seen": 456539136
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004351755265797392,
      "loss": 3.0481,
      "theoretical_loss": 3.953631344725653,
      "tokens_seen": 456604672
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004351654964894684,
      "loss": 2.8957,
      "theoretical_loss": 3.953569270526601,
      "tokens_seen": 456670208
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043515546639919764,
      "loss": 2.9168,
      "theoretical_loss": 3.953507207728956,
      "tokens_seen": 456735744
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043514543630892677,
      "loss": 2.9412,
      "theoretical_loss": 3.9534451563289883,
      "tokens_seen": 456801280
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.000435135406218656,
      "loss": 2.9884,
      "theoretical_loss": 3.9533831163229705,
      "tokens_seen": 456866816
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043512537612838513,
      "loss": 3.1086,
      "theoretical_loss": 3.953321087707177,
      "tokens_seen": 456932352
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043511534603811437,
      "loss": 3.0378,
      "theoretical_loss": 3.9532590704778823,
      "tokens_seen": 456997888
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043510531594784355,
      "loss": 3.1419,
      "theoretical_loss": 3.9531970646313637,
      "tokens_seen": 457063424
    },
    {
      "epoch": 1.05,
      "objective/train/docs_used": 548950,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.220716714859009,
      "objective/train/theoretical_loss": 3.9531505677141836,
      "objective/train/tokens_used": 477572576,
      "theoretical_loss": 3.9531505677141836,
      "tokens_seen": 457112576
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043509528585757273,
      "loss": 3.0836,
      "theoretical_loss": 3.953135070163901,
      "tokens_seen": 457128960
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004350852557673019,
      "loss": 2.9593,
      "theoretical_loss": 3.9530730870717745,
      "tokens_seen": 457194496
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004350752256770311,
      "loss": 3.2264,
      "theoretical_loss": 3.9530111153512677,
      "tokens_seen": 457260032
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043506519558676027,
      "loss": 3.0847,
      "theoretical_loss": 3.9529491549986644,
      "tokens_seen": 457325568
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004350551654964895,
      "loss": 2.9372,
      "theoretical_loss": 3.9528872060102502,
      "tokens_seen": 457391104
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043504513540621863,
      "loss": 3.154,
      "theoretical_loss": 3.952825268382314,
      "tokens_seen": 457456640
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043503510531594787,
      "loss": 3.0226,
      "theoretical_loss": 3.9527633421111448,
      "tokens_seen": 457522176
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043502507522567705,
      "loss": 3.0597,
      "theoretical_loss": 3.952701427193033,
      "tokens_seen": 457587712
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043501504513540623,
      "loss": 3.0086,
      "theoretical_loss": 3.9526395236242737,
      "tokens_seen": 457653248
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004350050150451354,
      "loss": 3.2069,
      "theoretical_loss": 3.9525776314011605,
      "tokens_seen": 457718784
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004349949849548646,
      "loss": 3.0413,
      "theoretical_loss": 3.95251575051999,
      "tokens_seen": 457784320
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004349849548645938,
      "loss": 2.9921,
      "theoretical_loss": 3.9524538809770604,
      "tokens_seen": 457849856
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.000434974924774323,
      "loss": 3.0149,
      "theoretical_loss": 3.952392022768672,
      "tokens_seen": 457915392
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043496489468405214,
      "loss": 2.9887,
      "theoretical_loss": 3.952330175891127,
      "tokens_seen": 457980928
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043495486459378137,
      "loss": 2.9086,
      "theoretical_loss": 3.952268340340728,
      "tokens_seen": 458046464
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004349448345035105,
      "loss": 3.1181,
      "theoretical_loss": 3.952206516113781,
      "tokens_seen": 458112000
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043493480441323973,
      "loss": 3.0618,
      "theoretical_loss": 3.952144703206592,
      "tokens_seen": 458177536
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004349247743229689,
      "loss": 2.952,
      "theoretical_loss": 3.952082901615471,
      "tokens_seen": 458243072
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004349147442326981,
      "loss": 3.0284,
      "theoretical_loss": 3.9520211113367276,
      "tokens_seen": 458308608
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004349047141424273,
      "loss": 3.1181,
      "theoretical_loss": 3.9519593323666746,
      "tokens_seen": 458374144
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004348946840521565,
      "loss": 2.913,
      "theoretical_loss": 3.951897564701625,
      "tokens_seen": 458439680
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043488465396188564,
      "loss": 2.6729,
      "theoretical_loss": 3.951835808337895,
      "tokens_seen": 458505216
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004348746238716149,
      "loss": 3.3398,
      "theoretical_loss": 3.9517740632718015,
      "tokens_seen": 458570752
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.000434864593781344,
      "loss": 2.8925,
      "theoretical_loss": 3.9517123294996646,
      "tokens_seen": 458636288
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043485456369107324,
      "loss": 2.9762,
      "theoretical_loss": 3.951650607017804,
      "tokens_seen": 458701824
    },
    {
      "epoch": 1.05,
      "objective/train/docs_used": 550265,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.620121955871582,
      "objective/train/theoretical_loss": 3.9516043225634405,
      "objective/train/tokens_used": 479210976,
      "theoretical_loss": 3.9516043225634405,
      "tokens_seen": 458750976
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004348445336008024,
      "loss": 3.1754,
      "theoretical_loss": 3.9515888958225425,
      "tokens_seen": 458767360
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004348345035105316,
      "loss": 2.9674,
      "theoretical_loss": 3.951527195910205,
      "tokens_seen": 458832896
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004348244734202608,
      "loss": 3.0026,
      "theoretical_loss": 3.9514655072771165,
      "tokens_seen": 458898432
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043481444332998996,
      "loss": 2.9806,
      "theoretical_loss": 3.951403829919606,
      "tokens_seen": 458963968
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043480441323971914,
      "loss": 3.0238,
      "theoretical_loss": 3.951342163834001,
      "tokens_seen": 459029504
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004347943831494484,
      "loss": 2.9712,
      "theoretical_loss": 3.9512805090166347,
      "tokens_seen": 459095040
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004347843530591775,
      "loss": 2.9981,
      "theoretical_loss": 3.9512188654638387,
      "tokens_seen": 459160576
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043477432296890674,
      "loss": 2.9099,
      "theoretical_loss": 3.951157233171948,
      "tokens_seen": 459226112
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043476429287863587,
      "loss": 3.0506,
      "theoretical_loss": 3.9510956121372978,
      "tokens_seen": 459291648
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004347542627883651,
      "loss": 2.9484,
      "theoretical_loss": 3.951034002356228,
      "tokens_seen": 459357184
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004347442326980943,
      "loss": 2.9222,
      "theoretical_loss": 3.9509724038250775,
      "tokens_seen": 459422720
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043473420260782346,
      "loss": 3.1718,
      "theoretical_loss": 3.950910816540187,
      "tokens_seen": 459488256
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043472417251755265,
      "loss": 3.1192,
      "theoretical_loss": 3.9508492404979005,
      "tokens_seen": 459553792
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004347141424272819,
      "loss": 3.248,
      "theoretical_loss": 3.950787675694562,
      "tokens_seen": 459619328
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.000434704112337011,
      "loss": 2.9839,
      "theoretical_loss": 3.950726122126519,
      "tokens_seen": 459684864
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043469408224674024,
      "loss": 3.1868,
      "theoretical_loss": 3.9506645797901196,
      "tokens_seen": 459750400
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004346840521564694,
      "loss": 2.8167,
      "theoretical_loss": 3.950603048681714,
      "tokens_seen": 459815936
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004346740220661986,
      "loss": 3.07,
      "theoretical_loss": 3.950541528797652,
      "tokens_seen": 459881472
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043466399197592784,
      "loss": 2.9742,
      "theoretical_loss": 3.9504800201342896,
      "tokens_seen": 459947008
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043465396188565697,
      "loss": 3.0656,
      "theoretical_loss": 3.9504185226879804,
      "tokens_seen": 460012544
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004346439317953862,
      "loss": 3.0949,
      "theoretical_loss": 3.950357036455081,
      "tokens_seen": 460078080
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043463390170511533,
      "loss": 2.9103,
      "theoretical_loss": 3.950295561431951,
      "tokens_seen": 460143616
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043462387161484457,
      "loss": 3.0709,
      "theoretical_loss": 3.95023409761495,
      "tokens_seen": 460209152
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043461384152457375,
      "loss": 2.99,
      "theoretical_loss": 3.950172645000439,
      "tokens_seen": 460274688
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043460381143430293,
      "loss": 3.1795,
      "theoretical_loss": 3.9501112035847834,
      "tokens_seen": 460340224
    },
    {
      "epoch": 1.05,
      "objective/train/docs_used": 550910,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.578948736190796,
      "objective/train/theoretical_loss": 3.950065129870103,
      "objective/train/tokens_used": 480849376,
      "theoretical_loss": 3.950065129870103,
      "tokens_seen": 460389376
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004345937813440321,
      "loss": 2.9974,
      "theoretical_loss": 3.950049773364347,
      "tokens_seen": 460405760
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004345837512537613,
      "loss": 3.1578,
      "theoretical_loss": 3.9499883543354977,
      "tokens_seen": 460471296
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043457372116349047,
      "loss": 3.0666,
      "theoretical_loss": 3.9499269464946036,
      "tokens_seen": 460536832
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004345636910732197,
      "loss": 3.0071,
      "theoretical_loss": 3.949865549838035,
      "tokens_seen": 460602368
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043455366098294883,
      "loss": 3.25,
      "theoretical_loss": 3.9498041643621646,
      "tokens_seen": 460667904
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043454363089267807,
      "loss": 3.1709,
      "theoretical_loss": 3.949742790063366,
      "tokens_seen": 460733440
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043453360080240725,
      "loss": 2.9825,
      "theoretical_loss": 3.9496814269380143,
      "tokens_seen": 460798976
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043452357071213643,
      "loss": 2.7215,
      "theoretical_loss": 3.9496200749824864,
      "tokens_seen": 460864512
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004345135406218656,
      "loss": 3.2174,
      "theoretical_loss": 3.949558734193162,
      "tokens_seen": 460930048
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004345035105315948,
      "loss": 3.0671,
      "theoretical_loss": 3.9494974045664213,
      "tokens_seen": 460995584
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.000434493480441324,
      "loss": 2.8494,
      "theoretical_loss": 3.949436086098646,
      "tokens_seen": 461061120
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004344834503510532,
      "loss": 3.0576,
      "theoretical_loss": 3.9493747787862206,
      "tokens_seen": 461126656
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043447342026078234,
      "loss": 2.9423,
      "theoretical_loss": 3.9493134826255303,
      "tokens_seen": 461192192
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043446339017051157,
      "loss": 3.0723,
      "theoretical_loss": 3.9492521976129624,
      "tokens_seen": 461257728
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004344533600802407,
      "loss": 2.9127,
      "theoretical_loss": 3.9491909237449065,
      "tokens_seen": 461323264
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043444332998996993,
      "loss": 2.8746,
      "theoretical_loss": 3.9491296610177526,
      "tokens_seen": 461388800
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004344332998996991,
      "loss": 3.0022,
      "theoretical_loss": 3.949068409427893,
      "tokens_seen": 461454336
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004344232698094283,
      "loss": 2.8165,
      "theoretical_loss": 3.949007168971722,
      "tokens_seen": 461519872
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004344132397191575,
      "loss": 3.2636,
      "theoretical_loss": 3.9489459396456343,
      "tokens_seen": 461585408
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004344032096288867,
      "loss": 3.1145,
      "theoretical_loss": 3.9488847214460288,
      "tokens_seen": 461650944
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043439317953861584,
      "loss": 2.9202,
      "theoretical_loss": 3.9488235143693036,
      "tokens_seen": 461716480
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004343831494483451,
      "loss": 2.8524,
      "theoretical_loss": 3.94876231841186,
      "tokens_seen": 461782016
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004343731193580742,
      "loss": 3.0581,
      "theoretical_loss": 3.948701133570099,
      "tokens_seen": 461847552
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043436308926780344,
      "loss": 3.0472,
      "theoretical_loss": 3.9486399598404263,
      "tokens_seen": 461913088
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004343530591775326,
      "loss": 2.8361,
      "theoretical_loss": 3.948578797219247,
      "tokens_seen": 461978624
    },
    {
      "epoch": 1.05,
      "objective/train/docs_used": 552255,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.0387556552886963,
      "objective/train/theoretical_loss": 3.9485329325411502,
      "objective/train/tokens_used": 482487776,
      "theoretical_loss": 3.9485329325411502,
      "tokens_seen": 462027776
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004343430290872618,
      "loss": 2.9454,
      "theoretical_loss": 3.9485176457029683,
      "tokens_seen": 462044160
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.000434332998996991,
      "loss": 2.9201,
      "theoretical_loss": 3.948456505287999,
      "tokens_seen": 462109696
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043432296890672016,
      "loss": 2.7927,
      "theoretical_loss": 3.9483953759707515,
      "tokens_seen": 462175232
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043431293881644934,
      "loss": 3.0389,
      "theoretical_loss": 3.948334257747636,
      "tokens_seen": 462240768
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004343029087261786,
      "loss": 3.0607,
      "theoretical_loss": 3.948273150615068,
      "tokens_seen": 462306304
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004342928786359077,
      "loss": 2.8376,
      "theoretical_loss": 3.9482120545694626,
      "tokens_seen": 462371840
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043428284854563694,
      "loss": 3.1191,
      "theoretical_loss": 3.948150969607237,
      "tokens_seen": 462437376
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043427281845536607,
      "loss": 3.0403,
      "theoretical_loss": 3.9480898957248116,
      "tokens_seen": 462502912
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004342627883650953,
      "loss": 3.2992,
      "theoretical_loss": 3.948028832918606,
      "tokens_seen": 462568448
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004342527582748245,
      "loss": 2.7688,
      "theoretical_loss": 3.9479677811850427,
      "tokens_seen": 462633984
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043424272818455367,
      "loss": 2.912,
      "theoretical_loss": 3.947906740520546,
      "tokens_seen": 462699520
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043423269809428285,
      "loss": 2.9298,
      "theoretical_loss": 3.9478457109215412,
      "tokens_seen": 462765056
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004342226680040121,
      "loss": 2.8303,
      "theoretical_loss": 3.947784692384457,
      "tokens_seen": 462830592
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004342126379137412,
      "loss": 3.0642,
      "theoretical_loss": 3.9477236849057205,
      "tokens_seen": 462896128
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043420260782347044,
      "loss": 3.0432,
      "theoretical_loss": 3.9476626884817634,
      "tokens_seen": 462961664
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043419257773319957,
      "loss": 2.8215,
      "theoretical_loss": 3.9476017031090187,
      "tokens_seen": 463027200
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004341825476429288,
      "loss": 2.9692,
      "theoretical_loss": 3.9475407287839195,
      "tokens_seen": 463092736
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.000434172517552658,
      "loss": 3.0146,
      "theoretical_loss": 3.947479765502902,
      "tokens_seen": 463158272
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043416248746238717,
      "loss": 3.0465,
      "theoretical_loss": 3.947418813262403,
      "tokens_seen": 463223808
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043415245737211635,
      "loss": 3.0851,
      "theoretical_loss": 3.947357872058862,
      "tokens_seen": 463289344
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043414242728184553,
      "loss": 2.7837,
      "theoretical_loss": 3.9472969418887196,
      "tokens_seen": 463354880
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004341323971915747,
      "loss": 3.2644,
      "theoretical_loss": 3.9472360227484176,
      "tokens_seen": 463420416
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043412236710130395,
      "loss": 3.228,
      "theoretical_loss": 3.9471751146344003,
      "tokens_seen": 463485952
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004341123370110331,
      "loss": 2.8354,
      "theoretical_loss": 3.947114217543113,
      "tokens_seen": 463551488
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004341023069207623,
      "loss": 2.8766,
      "theoretical_loss": 3.947053331471003,
      "tokens_seen": 463617024
    },
    {
      "epoch": 1.05,
      "objective/train/docs_used": 552845,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9544570446014404,
      "objective/train/theoretical_loss": 3.94700767414612,
      "objective/train/tokens_used": 484126176,
      "theoretical_loss": 3.94700767414612,
      "tokens_seen": 463666176
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043409227683049144,
      "loss": 3.0834,
      "theoretical_loss": 3.9469924564145202,
      "tokens_seen": 463682560
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043408224674022067,
      "loss": 2.8789,
      "theoretical_loss": 3.9469315923701136,
      "tokens_seen": 463748096
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043407221664994985,
      "loss": 2.9113,
      "theoretical_loss": 3.946870739334236,
      "tokens_seen": 463813632
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043406218655967903,
      "loss": 2.9886,
      "theoretical_loss": 3.946809897303342,
      "tokens_seen": 463879168
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004340521564694082,
      "loss": 2.9726,
      "theoretical_loss": 3.9467490662738856,
      "tokens_seen": 463944704
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043404212637913745,
      "loss": 2.9747,
      "theoretical_loss": 3.9466882462423243,
      "tokens_seen": 464010240
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004340320962888666,
      "loss": 3.2831,
      "theoretical_loss": 3.9466274372051178,
      "tokens_seen": 464075776
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004340220661985958,
      "loss": 2.8867,
      "theoretical_loss": 3.9465666391587257,
      "tokens_seen": 464141312
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043401203610832494,
      "loss": 3.0543,
      "theoretical_loss": 3.94650585209961,
      "tokens_seen": 464206848
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004340020060180542,
      "loss": 2.9371,
      "theoretical_loss": 3.9464450760242347,
      "tokens_seen": 464272384
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043399197592778336,
      "loss": 3.0978,
      "theoretical_loss": 3.946384310929065,
      "tokens_seen": 464337920
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043398194583751254,
      "loss": 2.9277,
      "theoretical_loss": 3.9463235568105675,
      "tokens_seen": 464403456
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004339719157472417,
      "loss": 3.1595,
      "theoretical_loss": 3.946262813665211,
      "tokens_seen": 464468992
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004339618856569709,
      "loss": 2.8856,
      "theoretical_loss": 3.9462020814894663,
      "tokens_seen": 464534528
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004339518555667001,
      "loss": 2.9228,
      "theoretical_loss": 3.946141360279804,
      "tokens_seen": 464600064
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004339418254764293,
      "loss": 2.8936,
      "theoretical_loss": 3.9460806500326986,
      "tokens_seen": 464665600
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004339317953861585,
      "loss": 3.135,
      "theoretical_loss": 3.9460199507446245,
      "tokens_seen": 464731136
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004339217652958877,
      "loss": 3.0263,
      "theoretical_loss": 3.9459592624120594,
      "tokens_seen": 464796672
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004339117352056169,
      "loss": 3.0817,
      "theoretical_loss": 3.94589858503148,
      "tokens_seen": 464862208
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043390170511534604,
      "loss": 2.9917,
      "theoretical_loss": 3.945837918599368,
      "tokens_seen": 464927744
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004338916750250753,
      "loss": 3.0115,
      "theoretical_loss": 3.9457772631122037,
      "tokens_seen": 464993280
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004338816449348044,
      "loss": 2.9553,
      "theoretical_loss": 3.945716618566472,
      "tokens_seen": 465058816
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043387161484453364,
      "loss": 2.8438,
      "theoretical_loss": 3.945655984958656,
      "tokens_seen": 465124352
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004338615847542628,
      "loss": 2.9346,
      "theoretical_loss": 3.9455953622852427,
      "tokens_seen": 465189888
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.000433851554663992,
      "loss": 2.9295,
      "theoretical_loss": 3.9455347505427207,
      "tokens_seen": 465255424
    },
    {
      "epoch": 1.05,
      "objective/train/docs_used": 553513,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9996495246887207,
      "objective/train/theoretical_loss": 3.945489298907115,
      "objective/train/tokens_used": 485764576,
      "theoretical_loss": 3.945489298907115,
      "tokens_seen": 465304576
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004338415245737212,
      "loss": 2.9936,
      "theoretical_loss": 3.9454741497275796,
      "tokens_seen": 465320960
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043383149448345036,
      "loss": 2.875,
      "theoretical_loss": 3.945413559836311,
      "tokens_seen": 465386496
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043382146439317954,
      "loss": 3.1446,
      "theoretical_loss": 3.9453529808654064,
      "tokens_seen": 465452032
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004338114343029088,
      "loss": 3.0625,
      "theoretical_loss": 3.9452924128113622,
      "tokens_seen": 465517568
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004338014042126379,
      "loss": 3.0735,
      "theoretical_loss": 3.945231855670674,
      "tokens_seen": 465583104
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043379137412236714,
      "loss": 3.1854,
      "theoretical_loss": 3.9451713094398393,
      "tokens_seen": 465648640
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043378134403209627,
      "loss": 3.1381,
      "theoretical_loss": 3.9451107741153577,
      "tokens_seen": 465714176
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004337713139418255,
      "loss": 3.0158,
      "theoretical_loss": 3.9450502496937307,
      "tokens_seen": 465779712
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004337612838515547,
      "loss": 2.8256,
      "theoretical_loss": 3.9449897361714603,
      "tokens_seen": 465845248
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043375125376128387,
      "loss": 2.8059,
      "theoretical_loss": 3.944929233545051,
      "tokens_seen": 465910784
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043374122367101305,
      "loss": 2.9526,
      "theoretical_loss": 3.9448687418110095,
      "tokens_seen": 465976320
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004337311935807423,
      "loss": 3.1095,
      "theoretical_loss": 3.944808260965842,
      "tokens_seen": 466041856
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004337211634904714,
      "loss": 3.2617,
      "theoretical_loss": 3.9447477910060584,
      "tokens_seen": 466107392
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043371113340020064,
      "loss": 3.0492,
      "theoretical_loss": 3.94468733192817,
      "tokens_seen": 466172928
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043370110330992977,
      "loss": 3.0652,
      "theoretical_loss": 3.944626883728688,
      "tokens_seen": 466238464
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.000433691073219659,
      "loss": 3.1735,
      "theoretical_loss": 3.9445664464041266,
      "tokens_seen": 466304000
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004336810431293882,
      "loss": 2.9657,
      "theoretical_loss": 3.944506019951002,
      "tokens_seen": 466369536
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043367101303911737,
      "loss": 2.8674,
      "theoretical_loss": 3.9444456043658302,
      "tokens_seen": 466435072
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043366098294884655,
      "loss": 3.032,
      "theoretical_loss": 3.944385199645132,
      "tokens_seen": 466500608
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043365095285857573,
      "loss": 2.8132,
      "theoretical_loss": 3.9443248057854254,
      "tokens_seen": 466566144
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004336409227683049,
      "loss": 2.8774,
      "theoretical_loss": 3.9442644227832337,
      "tokens_seen": 466631680
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043363089267803415,
      "loss": 2.8377,
      "theoretical_loss": 3.9442040506350806,
      "tokens_seen": 466697216
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004336208625877633,
      "loss": 3.2088,
      "theoretical_loss": 3.9441436893374906,
      "tokens_seen": 466762752
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004336108324974925,
      "loss": 2.8835,
      "theoretical_loss": 3.9440833388869914,
      "tokens_seen": 466828288
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043360080240722164,
      "loss": 2.9217,
      "theoretical_loss": 3.9440229992801106,
      "tokens_seen": 466893824
    },
    {
      "epoch": 1.05,
      "objective/train/docs_used": 555014,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.009627103805542,
      "objective/train/theoretical_loss": 3.943977751688987,
      "objective/train/tokens_used": 487402976,
      "theoretical_loss": 3.943977751688987,
      "tokens_seen": 466942976
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043359077231695087,
      "loss": 2.8843,
      "theoretical_loss": 3.943962670513378,
      "tokens_seen": 466959360
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043358074222668005,
      "loss": 2.7812,
      "theoretical_loss": 3.9439023525833257,
      "tokens_seen": 467024896
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043357071213640923,
      "loss": 3.0704,
      "theoretical_loss": 3.9438420454864875,
      "tokens_seen": 467090432
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004335606820461384,
      "loss": 3.2732,
      "theoretical_loss": 3.9437817492193963,
      "tokens_seen": 467155968
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043355065195586765,
      "loss": 2.9273,
      "theoretical_loss": 3.9437214637785902,
      "tokens_seen": 467221504
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004335406218655968,
      "loss": 3.0148,
      "theoretical_loss": 3.943661189160607,
      "tokens_seen": 467287040
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.000433530591775326,
      "loss": 3.1349,
      "theoretical_loss": 3.9436009253619853,
      "tokens_seen": 467352576
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043352056168505514,
      "loss": 2.9502,
      "theoretical_loss": 3.9435406723792665,
      "tokens_seen": 467418112
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004335105315947844,
      "loss": 2.8156,
      "theoretical_loss": 3.9434804302089934,
      "tokens_seen": 467483648
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043350050150451356,
      "loss": 2.8378,
      "theoretical_loss": 3.9434201988477113,
      "tokens_seen": 467549184
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043349047141424274,
      "loss": 3.1404,
      "theoretical_loss": 3.943359978291965,
      "tokens_seen": 467614720
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004334804413239719,
      "loss": 2.8589,
      "theoretical_loss": 3.943299768538302,
      "tokens_seen": 467680256
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004334704112337011,
      "loss": 3.0278,
      "theoretical_loss": 3.943239569583272,
      "tokens_seen": 467745792
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004334603811434303,
      "loss": 3.1748,
      "theoretical_loss": 3.943179381423425,
      "tokens_seen": 467811328
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004334503510531595,
      "loss": 2.818,
      "theoretical_loss": 3.9431192040553142,
      "tokens_seen": 467876864
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043344032096288864,
      "loss": 2.8783,
      "theoretical_loss": 3.9430590374754924,
      "tokens_seen": 467942400
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004334302908726179,
      "loss": 3.0593,
      "theoretical_loss": 3.9429988816805155,
      "tokens_seen": 468007936
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.000433420260782347,
      "loss": 2.9382,
      "theoretical_loss": 3.9429387366669406,
      "tokens_seen": 468073472
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043341023069207624,
      "loss": 3.0675,
      "theoretical_loss": 3.9428786024313256,
      "tokens_seen": 468139008
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004334002006018054,
      "loss": 2.8618,
      "theoretical_loss": 3.9428184789702323,
      "tokens_seen": 468204544
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004333901705115346,
      "loss": 2.9609,
      "theoretical_loss": 3.9427583662802204,
      "tokens_seen": 468270080
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004333801404212638,
      "loss": 2.8331,
      "theoretical_loss": 3.9426982643578548,
      "tokens_seen": 468335616
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.000433370110330993,
      "loss": 2.8806,
      "theoretical_loss": 3.9426381731996996,
      "tokens_seen": 468401152
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043336008024072215,
      "loss": 2.8969,
      "theoretical_loss": 3.9425780928023215,
      "tokens_seen": 468466688
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004333500501504514,
      "loss": 2.82,
      "theoretical_loss": 3.9425180231622883,
      "tokens_seen": 468532224
    },
    {
      "epoch": 1.05,
      "objective/train/docs_used": 555695,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4344749450683594,
      "objective/train/theoretical_loss": 3.9424729779897074,
      "objective/train/tokens_used": 489041376,
      "theoretical_loss": 3.9424729779897074,
      "tokens_seen": 468581376
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004333400200601805,
      "loss": 2.8374,
      "theoretical_loss": 3.94245796427617,
      "tokens_seen": 468597760
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043332998996990974,
      "loss": 2.963,
      "theoretical_loss": 3.942397916140538,
      "tokens_seen": 468663296
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004333199598796389,
      "loss": 2.9556,
      "theoretical_loss": 3.942337878751964,
      "tokens_seen": 468728832
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004333099297893681,
      "loss": 2.7379,
      "theoretical_loss": 3.942277852107024,
      "tokens_seen": 468794368
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004332998996990973,
      "loss": 2.9639,
      "theoretical_loss": 3.9422178362022917,
      "tokens_seen": 468859904
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043328986960882647,
      "loss": 3.1148,
      "theoretical_loss": 3.9421578310343466,
      "tokens_seen": 468925440
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043327983951855565,
      "loss": 3.0438,
      "theoretical_loss": 3.9420978365997668,
      "tokens_seen": 468990976
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004332698094282849,
      "loss": 2.9268,
      "theoretical_loss": 3.942037852895133,
      "tokens_seen": 469056512
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.000433259779338014,
      "loss": 2.9954,
      "theoretical_loss": 3.9419778799170286,
      "tokens_seen": 469122048
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043324974924774325,
      "loss": 2.9419,
      "theoretical_loss": 3.941917917662035,
      "tokens_seen": 469187584
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043323971915747243,
      "loss": 3.0639,
      "theoretical_loss": 3.941857966126739,
      "tokens_seen": 469253120
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004332296890672016,
      "loss": 2.9824,
      "theoretical_loss": 3.941798025307728,
      "tokens_seen": 469318656
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004332196589769308,
      "loss": 2.8335,
      "theoretical_loss": 3.941738095201589,
      "tokens_seen": 469384192
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043320962888665997,
      "loss": 2.9499,
      "theoretical_loss": 3.941678175804913,
      "tokens_seen": 469449728
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043319959879638915,
      "loss": 2.9987,
      "theoretical_loss": 3.9416182671142908,
      "tokens_seen": 469515264
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004331895687061184,
      "loss": 3.0375,
      "theoretical_loss": 3.9415583691263167,
      "tokens_seen": 469580800
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043317953861584757,
      "loss": 2.9472,
      "theoretical_loss": 3.9414984818375842,
      "tokens_seen": 469646336
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043316950852557675,
      "loss": 3.2323,
      "theoretical_loss": 3.9414386052446906,
      "tokens_seen": 469711872
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043315947843530593,
      "loss": 3.0547,
      "theoretical_loss": 3.9413787393442328,
      "tokens_seen": 469777408
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004331494483450351,
      "loss": 2.8593,
      "theoretical_loss": 3.9413188841328104,
      "tokens_seen": 469842944
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043313941825476435,
      "loss": 2.9321,
      "theoretical_loss": 3.941259039607025,
      "tokens_seen": 469908480
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004331293881644935,
      "loss": 2.831,
      "theoretical_loss": 3.9411992057634775,
      "tokens_seen": 469974016
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004331193580742227,
      "loss": 3.0224,
      "theoretical_loss": 3.9411393825987737,
      "tokens_seen": 470039552
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043310932798395184,
      "loss": 2.9007,
      "theoretical_loss": 3.941079570109518,
      "tokens_seen": 470105088
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043309929789368107,
      "loss": 2.9678,
      "theoretical_loss": 3.9410197682923185,
      "tokens_seen": 470170624
    },
    {
      "epoch": 1.05,
      "objective/train/docs_used": 556746,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.731928586959839,
      "objective/train/theoretical_loss": 3.9409749239309146,
      "objective/train/tokens_used": 490679776,
      "theoretical_loss": 3.9409749239309146,
      "tokens_seen": 470219776
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043308926780341025,
      "loss": 2.95,
      "theoretical_loss": 3.940959977143783,
      "tokens_seen": 470236160
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043307923771313943,
      "loss": 2.7916,
      "theoretical_loss": 3.9409001966605217,
      "tokens_seen": 470301696
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004330692076228686,
      "loss": 3.0253,
      "theoretical_loss": 3.9408404268391473,
      "tokens_seen": 470367232
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043305917753259785,
      "loss": 2.9868,
      "theoretical_loss": 3.940780667676272,
      "tokens_seen": 470432768
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.000433049147442327,
      "loss": 3.1419,
      "theoretical_loss": 3.9407209191685117,
      "tokens_seen": 470498304
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004330391173520562,
      "loss": 2.9542,
      "theoretical_loss": 3.940661181312482,
      "tokens_seen": 470563840
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043302908726178534,
      "loss": 3.0518,
      "theoretical_loss": 3.9406014541048013,
      "tokens_seen": 470629376
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004330190571715146,
      "loss": 2.8218,
      "theoretical_loss": 3.9405417375420893,
      "tokens_seen": 470694912
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043300902708124376,
      "loss": 3.094,
      "theoretical_loss": 3.9404820316209666,
      "tokens_seen": 470760448
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043299899699097294,
      "loss": 2.9356,
      "theoretical_loss": 3.940422336338056,
      "tokens_seen": 470825984
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004329889669007021,
      "loss": 3.0012,
      "theoretical_loss": 3.940362651689982,
      "tokens_seen": 470891520
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004329789368104313,
      "loss": 3.1362,
      "theoretical_loss": 3.940302977673369,
      "tokens_seen": 470957056
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004329689067201605,
      "loss": 3.0018,
      "theoretical_loss": 3.940243314284846,
      "tokens_seen": 471022592
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004329588766298897,
      "loss": 2.8662,
      "theoretical_loss": 3.940183661521041,
      "tokens_seen": 471088128
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043294884653961884,
      "loss": 3.1666,
      "theoretical_loss": 3.940124019378583,
      "tokens_seen": 471153664
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004329388164493481,
      "loss": 3.1978,
      "theoretical_loss": 3.940064387854106,
      "tokens_seen": 471219200
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004329287863590772,
      "loss": 2.7946,
      "theoretical_loss": 3.9400047669442424,
      "tokens_seen": 471284736
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043291875626880644,
      "loss": 2.9157,
      "theoretical_loss": 3.939945156645627,
      "tokens_seen": 471350272
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004329087261785356,
      "loss": 2.9845,
      "theoretical_loss": 3.9398855569548963,
      "tokens_seen": 471415808
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004328986960882648,
      "loss": 2.9426,
      "theoretical_loss": 3.939825967868688,
      "tokens_seen": 471481344
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.000432888665997994,
      "loss": 2.7147,
      "theoretical_loss": 3.9397663893836428,
      "tokens_seen": 471546880
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004328786359077232,
      "loss": 3.2359,
      "theoretical_loss": 3.9397068214964,
      "tokens_seen": 471612416
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043286860581745235,
      "loss": 3.0412,
      "theoretical_loss": 3.939647264203604,
      "tokens_seen": 471677952
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004328585757271816,
      "loss": 2.9235,
      "theoretical_loss": 3.9395877175018965,
      "tokens_seen": 471743488
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004328485456369107,
      "loss": 3.1163,
      "theoretical_loss": 3.9395281813879257,
      "tokens_seen": 471809024
    },
    {
      "epoch": 1.05,
      "objective/train/docs_used": 557491,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.751948595046997,
      "objective/train/theoretical_loss": 3.939483536248632,
      "objective/train/tokens_used": 492318176,
      "theoretical_loss": 3.939483536248632,
      "tokens_seen": 471858176
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043283851554663994,
      "loss": 2.8767,
      "theoretical_loss": 3.9394686558583376,
      "tokens_seen": 471874560
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004328284854563691,
      "loss": 2.8283,
      "theoretical_loss": 3.9394091409097807,
      "tokens_seen": 471940096
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004328184553660983,
      "loss": 2.7199,
      "theoretical_loss": 3.939349636538905,
      "tokens_seen": 472005632
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004328084252758275,
      "loss": 3.1735,
      "theoretical_loss": 3.9392901427423634,
      "tokens_seen": 472071168
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043279839518555667,
      "loss": 2.9966,
      "theoretical_loss": 3.939230659516808,
      "tokens_seen": 472136704
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043278836509528585,
      "loss": 2.6941,
      "theoretical_loss": 3.9391711868588946,
      "tokens_seen": 472202240
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004327783350050151,
      "loss": 3.0288,
      "theoretical_loss": 3.9391117247652785,
      "tokens_seen": 472267776
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004327683049147442,
      "loss": 3.1079,
      "theoretical_loss": 3.939052273232618,
      "tokens_seen": 472333312
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043275827482447345,
      "loss": 3.0203,
      "theoretical_loss": 3.938992832257572,
      "tokens_seen": 472398848
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043274824473420263,
      "loss": 2.8275,
      "theoretical_loss": 3.9389334018368025,
      "tokens_seen": 472464384
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004327382146439318,
      "loss": 3.2009,
      "theoretical_loss": 3.938873981966971,
      "tokens_seen": 472529920
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.000432728184553661,
      "loss": 2.8613,
      "theoretical_loss": 3.938814572644741,
      "tokens_seen": 472595456
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043271815446339017,
      "loss": 2.9294,
      "theoretical_loss": 3.938755173866779,
      "tokens_seen": 472660992
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043270812437311935,
      "loss": 3.0615,
      "theoretical_loss": 3.9386957856297515,
      "tokens_seen": 472726528
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004326980942828486,
      "loss": 3.0866,
      "theoretical_loss": 3.9386364079303267,
      "tokens_seen": 472792064
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004326880641925777,
      "loss": 2.9678,
      "theoretical_loss": 3.938577040765175,
      "tokens_seen": 472857600
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043267803410230695,
      "loss": 3.0434,
      "theoretical_loss": 3.9385176841309675,
      "tokens_seen": 472923136
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004326680040120361,
      "loss": 3.0813,
      "theoretical_loss": 3.938458338024377,
      "tokens_seen": 472988672
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004326579739217653,
      "loss": 3.0351,
      "theoretical_loss": 3.9383990024420794,
      "tokens_seen": 473054208
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004326479438314945,
      "loss": 3.0305,
      "theoretical_loss": 3.938339677380749,
      "tokens_seen": 473119744
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004326379137412237,
      "loss": 3.0702,
      "theoretical_loss": 3.938280362837064,
      "tokens_seen": 473185280
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043262788365095286,
      "loss": 2.8885,
      "theoretical_loss": 3.938221058807704,
      "tokens_seen": 473250816
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043261785356068204,
      "loss": 3.0995,
      "theoretical_loss": 3.9381617652893484,
      "tokens_seen": 473316352
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004326078234704112,
      "loss": 3.1313,
      "theoretical_loss": 3.93810248227868,
      "tokens_seen": 473381888
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043259779338014045,
      "loss": 2.9592,
      "theoretical_loss": 3.938043209772382,
      "tokens_seen": 473447424
    },
    {
      "epoch": 1.05,
      "objective/train/docs_used": 558752,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.13558030128479,
      "objective/train/theoretical_loss": 3.937998762284158,
      "objective/train/tokens_used": 493956576,
      "theoretical_loss": 3.937998762284158,
      "tokens_seen": 473496576
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004325877632898696,
      "loss": 2.9736,
      "theoretical_loss": 3.9379839477671403,
      "tokens_seen": 473512960
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004325777331995988,
      "loss": 3.0554,
      "theoretical_loss": 3.937924696259641,
      "tokens_seen": 473578496
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.000432567703109328,
      "loss": 3.0045,
      "theoretical_loss": 3.9378654552465715,
      "tokens_seen": 473644032
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004325576730190572,
      "loss": 3.0406,
      "theoretical_loss": 3.937806224724622,
      "tokens_seen": 473709568
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043254764292878636,
      "loss": 2.9355,
      "theoretical_loss": 3.9377470046904834,
      "tokens_seen": 473775104
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043253761283851554,
      "loss": 2.8813,
      "theoretical_loss": 3.9376877951408487,
      "tokens_seen": 473840640
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004325275827482447,
      "loss": 2.823,
      "theoretical_loss": 3.937628596072412,
      "tokens_seen": 473906176
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043251755265797396,
      "loss": 2.9766,
      "theoretical_loss": 3.9375694074818677,
      "tokens_seen": 473971712
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004325075225677031,
      "loss": 2.8709,
      "theoretical_loss": 3.9375102293659143,
      "tokens_seen": 474037248
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004324974924774323,
      "loss": 3.0939,
      "theoretical_loss": 3.9374510617212497,
      "tokens_seen": 474102784
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043248746238716145,
      "loss": 3.0702,
      "theoretical_loss": 3.937391904544574,
      "tokens_seen": 474168320
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004324774322968907,
      "loss": 3.1023,
      "theoretical_loss": 3.937332757832589,
      "tokens_seen": 474233856
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043246740220661986,
      "loss": 2.914,
      "theoretical_loss": 3.9372736215819977,
      "tokens_seen": 474299392
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043245737211634904,
      "loss": 2.7269,
      "theoretical_loss": 3.9372144957895046,
      "tokens_seen": 474364928
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004324473420260782,
      "loss": 3.1261,
      "theoretical_loss": 3.9371553804518156,
      "tokens_seen": 474430464
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004324373119358074,
      "loss": 3.0787,
      "theoretical_loss": 3.9370962755656382,
      "tokens_seen": 474496000
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043242728184553664,
      "loss": 3.2062,
      "theoretical_loss": 3.937037181127682,
      "tokens_seen": 474561536
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004324172517552658,
      "loss": 3.1947,
      "theoretical_loss": 3.936978097134657,
      "tokens_seen": 474627072
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.000432407221664995,
      "loss": 3.0234,
      "theoretical_loss": 3.9369190235832754,
      "tokens_seen": 474692608
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004323971915747242,
      "loss": 2.9252,
      "theoretical_loss": 3.936859960470251,
      "tokens_seen": 474758144
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004323871614844534,
      "loss": 2.8916,
      "theoretical_loss": 3.936800907792298,
      "tokens_seen": 474823680
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043237713139418255,
      "loss": 3.1763,
      "theoretical_loss": 3.9367418655461335,
      "tokens_seen": 474889216
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004323671013039118,
      "loss": 3.1412,
      "theoretical_loss": 3.936682833728476,
      "tokens_seen": 474954752
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004323570712136409,
      "loss": 3.0111,
      "theoretical_loss": 3.936623812336043,
      "tokens_seen": 475020288
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043234704112337014,
      "loss": 3.1496,
      "theoretical_loss": 3.9365648013655576,
      "tokens_seen": 475085824
    },
    {
      "epoch": 1.05,
      "objective/train/docs_used": 559487,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8737614154815674,
      "objective/train/theoretical_loss": 3.936520549975124,
      "objective/train/tokens_used": 495594976,
      "theoretical_loss": 3.936520549975124,
      "tokens_seen": 475134976
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004323370110330993,
      "loss": 3.1738,
      "theoretical_loss": 3.936505800813741,
      "tokens_seen": 475151360
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004323269809428285,
      "loss": 2.7831,
      "theoretical_loss": 3.9364468106773174,
      "tokens_seen": 475216896
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004323169508525577,
      "loss": 2.8794,
      "theoretical_loss": 3.936387830953013,
      "tokens_seen": 475282432
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043230692076228687,
      "loss": 3.0701,
      "theoretical_loss": 3.936328861637553,
      "tokens_seen": 475347968
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043229689067201605,
      "loss": 2.9024,
      "theoretical_loss": 3.936269902727667,
      "tokens_seen": 475413504
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004322868605817453,
      "loss": 3.3167,
      "theoretical_loss": 3.9362109542200843,
      "tokens_seen": 475479040
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004322768304914744,
      "loss": 3.2355,
      "theoretical_loss": 3.9361520161115364,
      "tokens_seen": 475544576
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043226680040120365,
      "loss": 3.1841,
      "theoretical_loss": 3.9360930883987564,
      "tokens_seen": 475610112
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043225677031093283,
      "loss": 3.0284,
      "theoretical_loss": 3.936034171078478,
      "tokens_seen": 475675648
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.000432246740220662,
      "loss": 3.0582,
      "theoretical_loss": 3.9359752641474373,
      "tokens_seen": 475741184
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004322367101303912,
      "loss": 3.0325,
      "theoretical_loss": 3.9359163676023705,
      "tokens_seen": 475806720
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043222668004012037,
      "loss": 3.1134,
      "theoretical_loss": 3.935857481440018,
      "tokens_seen": 475872256
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043221664994984955,
      "loss": 3.2472,
      "theoretical_loss": 3.935798605657119,
      "tokens_seen": 475937792
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004322066198595788,
      "loss": 3.1808,
      "theoretical_loss": 3.9357397402504146,
      "tokens_seen": 476003328
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004321965897693079,
      "loss": 3.0204,
      "theoretical_loss": 3.935680885216649,
      "tokens_seen": 476068864
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043218655967903715,
      "loss": 3.1083,
      "theoretical_loss": 3.9356220405525666,
      "tokens_seen": 476134400
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004321765295887663,
      "loss": 2.8384,
      "theoretical_loss": 3.9355632062549124,
      "tokens_seen": 476199936
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004321664994984955,
      "loss": 3.0837,
      "theoretical_loss": 3.9355043823204348,
      "tokens_seen": 476265472
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004321564694082247,
      "loss": 3.0754,
      "theoretical_loss": 3.935445568745883,
      "tokens_seen": 476331008
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004321464393179539,
      "loss": 3.1409,
      "theoretical_loss": 3.9353867655280066,
      "tokens_seen": 476396544
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043213640922768306,
      "loss": 3.2543,
      "theoretical_loss": 3.935327972663558,
      "tokens_seen": 476462080
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043212637913741224,
      "loss": 2.7809,
      "theoretical_loss": 3.935269190149291,
      "tokens_seen": 476527616
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004321163490471414,
      "loss": 2.9419,
      "theoretical_loss": 3.93521041798196,
      "tokens_seen": 476593152
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043210631895687065,
      "loss": 2.8718,
      "theoretical_loss": 3.935151656158321,
      "tokens_seen": 476658688
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004320962888665998,
      "loss": 3.0566,
      "theoretical_loss": 3.9350929046751313,
      "tokens_seen": 476724224
    },
    {
      "epoch": 1.05,
      "objective/train/docs_used": 560707,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.231940984725952,
      "objective/train/theoretical_loss": 3.9350488478467103,
      "objective/train/tokens_used": 497233376,
      "theoretical_loss": 3.9350488478467103,
      "tokens_seen": 476773376
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.000432086258776329,
      "loss": 3.0351,
      "theoretical_loss": 3.9350341635291515,
      "tokens_seen": 476789760
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004320762286860582,
      "loss": 3.0426,
      "theoretical_loss": 3.934975432717142,
      "tokens_seen": 476855296
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004320661985957874,
      "loss": 2.9293,
      "theoretical_loss": 3.9349167122358644,
      "tokens_seen": 476920832
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043205616850551656,
      "loss": 3.0121,
      "theoretical_loss": 3.934858002082082,
      "tokens_seen": 476986368
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043204613841524574,
      "loss": 3.1352,
      "theoretical_loss": 3.9347993022525607,
      "tokens_seen": 477051904
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004320361083249749,
      "loss": 3.0302,
      "theoretical_loss": 3.9347406127440663,
      "tokens_seen": 477117440
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043202607823470416,
      "loss": 2.9891,
      "theoretical_loss": 3.9346819335533674,
      "tokens_seen": 477182976
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004320160481444333,
      "loss": 3.0205,
      "theoretical_loss": 3.9346232646772332,
      "tokens_seen": 477248512
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004320060180541625,
      "loss": 2.7596,
      "theoretical_loss": 3.934564606112435,
      "tokens_seen": 477314048
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043199598796389165,
      "loss": 3.0225,
      "theoretical_loss": 3.934505957855744,
      "tokens_seen": 477379584
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004319859578736209,
      "loss": 2.9677,
      "theoretical_loss": 3.934447319903935,
      "tokens_seen": 477445120
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043197592778335006,
      "loss": 2.7368,
      "theoretical_loss": 3.934388692253783,
      "tokens_seen": 477510656
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043196589769307924,
      "loss": 3.0439,
      "theoretical_loss": 3.934330074902065,
      "tokens_seen": 477576192
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004319558676028084,
      "loss": 2.9094,
      "theoretical_loss": 3.9342714678455586,
      "tokens_seen": 477641728
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004319458375125376,
      "loss": 2.786,
      "theoretical_loss": 3.9342128710810433,
      "tokens_seen": 477707264
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004319358074222668,
      "loss": 3.1414,
      "theoretical_loss": 3.9341542846053006,
      "tokens_seen": 477772800
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.000431925777331996,
      "loss": 2.915,
      "theoretical_loss": 3.934095708415113,
      "tokens_seen": 477838336
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043191574724172515,
      "loss": 2.9227,
      "theoretical_loss": 3.934037142507264,
      "tokens_seen": 477903872
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004319057171514544,
      "loss": 2.7711,
      "theoretical_loss": 3.9339785868785397,
      "tokens_seen": 477969408
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043189568706118357,
      "loss": 2.8507,
      "theoretical_loss": 3.9339200415257265,
      "tokens_seen": 478034944
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043188565697091275,
      "loss": 3.0298,
      "theoretical_loss": 3.933861506445613,
      "tokens_seen": 478100480
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00043187562688064193,
      "loss": 2.8273,
      "theoretical_loss": 3.933802981634988,
      "tokens_seen": 478166016
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0004318655967903711,
      "loss": 3.1108,
      "theoretical_loss": 3.9337444670906434,
      "tokens_seen": 478231552
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004318555667001003,
      "loss": 2.9331,
      "theoretical_loss": 3.9336859628093723,
      "tokens_seen": 478297088
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004318455366098295,
      "loss": 3.0098,
      "theoretical_loss": 3.933627468787968,
      "tokens_seen": 478362624
    },
    {
      "epoch": 1.06,
      "objective/train/docs_used": 561422,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4702494144439697,
      "objective/train/theoretical_loss": 3.933583605003024,
      "objective/train/tokens_used": 498871776,
      "theoretical_loss": 3.933583605003024,
      "tokens_seen": 478411776
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00043183550651955865,
      "loss": 2.9758,
      "theoretical_loss": 3.933568985023226,
      "tokens_seen": 478428160
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004318254764292879,
      "loss": 3.136,
      "theoretical_loss": 3.933510511511943,
      "tokens_seen": 478493696
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.000431815446339017,
      "loss": 3.188,
      "theoretical_loss": 3.9334520482509183,
      "tokens_seen": 478559232
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00043180541624874625,
      "loss": 3.0016,
      "theoretical_loss": 3.933393595236951,
      "tokens_seen": 478624768
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00043179538615847543,
      "loss": 2.8216,
      "theoretical_loss": 3.9333351524668427,
      "tokens_seen": 478690304
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004317853560682046,
      "loss": 2.9534,
      "theoretical_loss": 3.9332767199373957,
      "tokens_seen": 478755840
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004317753259779338,
      "loss": 3.0022,
      "theoretical_loss": 3.9332182976454146,
      "tokens_seen": 478821376
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00043176529588766303,
      "loss": 3.1134,
      "theoretical_loss": 3.9331598855877044,
      "tokens_seen": 478886912
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00043175526579739215,
      "loss": 2.8389,
      "theoretical_loss": 3.9331014837610727,
      "tokens_seen": 478952448
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004317452357071214,
      "loss": 2.9801,
      "theoretical_loss": 3.933043092162327,
      "tokens_seen": 479017984
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004317352056168505,
      "loss": 2.9627,
      "theoretical_loss": 3.9329847107882787,
      "tokens_seen": 479083520
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00043172517552657975,
      "loss": 2.7128,
      "theoretical_loss": 3.9329263396357375,
      "tokens_seen": 479149056
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00043171514543630893,
      "loss": 2.8601,
      "theoretical_loss": 3.932867978701517,
      "tokens_seen": 479214592
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004317051153460381,
      "loss": 3.0035,
      "theoretical_loss": 3.932809627982431,
      "tokens_seen": 479280128
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004316950852557673,
      "loss": 3.0842,
      "theoretical_loss": 3.9327512874752952,
      "tokens_seen": 479345664
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004316850551654965,
      "loss": 3.1308,
      "theoretical_loss": 3.9326929571769265,
      "tokens_seen": 479411200
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004316750250752257,
      "loss": 3.1854,
      "theoretical_loss": 3.9326346370841434,
      "tokens_seen": 479476736
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004316649949849549,
      "loss": 2.9725,
      "theoretical_loss": 3.932576327193766,
      "tokens_seen": 479542272
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004316549648946841,
      "loss": 2.7423,
      "theoretical_loss": 3.9325180275026153,
      "tokens_seen": 479607808
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00043164493480441326,
      "loss": 3.0698,
      "theoretical_loss": 3.9324597380075135,
      "tokens_seen": 479673344
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00043163490471414244,
      "loss": 2.9938,
      "theoretical_loss": 3.932401458705286,
      "tokens_seen": 479738880
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004316248746238716,
      "loss": 3.025,
      "theoretical_loss": 3.932343189592757,
      "tokens_seen": 479804416
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00043161484453360085,
      "loss": 3.0979,
      "theoretical_loss": 3.9322849306667536,
      "tokens_seen": 479869952
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00043160481444333,
      "loss": 2.8941,
      "theoretical_loss": 3.9322266819241056,
      "tokens_seen": 479935488
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004315947843530592,
      "loss": 3.0009,
      "theoretical_loss": 3.9321684433616415,
      "tokens_seen": 480001024
    },
    {
      "epoch": 1.06,
      "objective/train/docs_used": 562507,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.448958158493042,
      "objective/train/theoretical_loss": 3.932124771118633,
      "objective/train/tokens_used": 500510176,
      "theoretical_loss": 3.932124771118633,
      "tokens_seen": 480050176
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004315847542627884,
      "loss": 2.906,
      "theoretical_loss": 3.932110214976193,
      "tokens_seen": 480066560
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004315747241725176,
      "loss": 2.8742,
      "theoretical_loss": 3.9320519967645926,
      "tokens_seen": 480132096
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00043156469408224676,
      "loss": 3.1216,
      "theoretical_loss": 3.931993788723674,
      "tokens_seen": 480197632
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00043155466399197594,
      "loss": 2.9462,
      "theoretical_loss": 3.9319355908502738,
      "tokens_seen": 480263168
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004315446339017051,
      "loss": 2.9604,
      "theoretical_loss": 3.931877403141228,
      "tokens_seen": 480328704
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00043153460381143436,
      "loss": 3.1821,
      "theoretical_loss": 3.9318192255933755,
      "tokens_seen": 480394240
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004315245737211635,
      "loss": 3.1779,
      "theoretical_loss": 3.9317610582035556,
      "tokens_seen": 480459776
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004315145436308927,
      "loss": 3.0144,
      "theoretical_loss": 3.9317029009686086,
      "tokens_seen": 480525312
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00043150451354062185,
      "loss": 3.1989,
      "theoretical_loss": 3.9316447538853794,
      "tokens_seen": 480590848
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004314944834503511,
      "loss": 2.9107,
      "theoretical_loss": 3.9315866169507094,
      "tokens_seen": 480656384
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00043148445336008026,
      "loss": 2.7994,
      "theoretical_loss": 3.931528490161446,
      "tokens_seen": 480721920
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00043147442326980944,
      "loss": 2.9174,
      "theoretical_loss": 3.9314703735144345,
      "tokens_seen": 480787456
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004314643931795386,
      "loss": 2.7913,
      "theoretical_loss": 3.931412267006524,
      "tokens_seen": 480852992
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004314543630892678,
      "loss": 3.0289,
      "theoretical_loss": 3.931354170634563,
      "tokens_seen": 480918528
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.000431444332998997,
      "loss": 3.0211,
      "theoretical_loss": 3.9312960843954046,
      "tokens_seen": 480984064
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004314343029087262,
      "loss": 2.9529,
      "theoretical_loss": 3.9312380082858995,
      "tokens_seen": 481049600
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00043142427281845535,
      "loss": 3.0352,
      "theoretical_loss": 3.931179942302902,
      "tokens_seen": 481115136
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004314142427281846,
      "loss": 2.9488,
      "theoretical_loss": 3.9311218864432673,
      "tokens_seen": 481180672
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00043140421263791377,
      "loss": 2.9354,
      "theoretical_loss": 3.9310638407038523,
      "tokens_seen": 481246208
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00043139418254764295,
      "loss": 2.8764,
      "theoretical_loss": 3.931005805081515,
      "tokens_seen": 481311744
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00043138415245737213,
      "loss": 2.8272,
      "theoretical_loss": 3.9309477795731143,
      "tokens_seen": 481377280
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004313741223671013,
      "loss": 2.9896,
      "theoretical_loss": 3.930889764175512,
      "tokens_seen": 481442816
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004313640922768305,
      "loss": 2.9828,
      "theoretical_loss": 3.930831758885569,
      "tokens_seen": 481508352
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004313540621865597,
      "loss": 3.0283,
      "theoretical_loss": 3.9307737637001505,
      "tokens_seen": 481573888
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00043134403209628885,
      "loss": 2.915,
      "theoretical_loss": 3.93071577861612,
      "tokens_seen": 481639424
    },
    {
      "epoch": 1.06,
      "objective/train/docs_used": 563271,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.949291229248047,
      "objective/train/theoretical_loss": 3.9306722964302496,
      "objective/train/tokens_used": 502148576,
      "theoretical_loss": 3.9306722964302496,
      "tokens_seen": 481688576
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004313340020060181,
      "loss": 3.0166,
      "theoretical_loss": 3.930657803630346,
      "tokens_seen": 481704960
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004313239719157472,
      "loss": 2.8268,
      "theoretical_loss": 3.9305998387396945,
      "tokens_seen": 481770496
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00043131394182547645,
      "loss": 3.1203,
      "theoretical_loss": 3.9305418839410353,
      "tokens_seen": 481836032
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00043130391173520563,
      "loss": 3.1673,
      "theoretical_loss": 3.930483939231239,
      "tokens_seen": 481901568
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004312938816449348,
      "loss": 2.9154,
      "theoretical_loss": 3.930426004607178,
      "tokens_seen": 481967104
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.000431283851554664,
      "loss": 2.7331,
      "theoretical_loss": 3.930368080065726,
      "tokens_seen": 482032640
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00043127382146439323,
      "loss": 2.9744,
      "theoretical_loss": 3.930310165603757,
      "tokens_seen": 482098176
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00043126379137412236,
      "loss": 2.959,
      "theoretical_loss": 3.930252261218147,
      "tokens_seen": 482163712
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004312537612838516,
      "loss": 2.9013,
      "theoretical_loss": 3.9301943669057744,
      "tokens_seen": 482229248
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004312437311935807,
      "loss": 2.9071,
      "theoretical_loss": 3.9301364826635186,
      "tokens_seen": 482294784
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00043123370110330995,
      "loss": 2.7957,
      "theoretical_loss": 3.930078608488259,
      "tokens_seen": 482360320
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00043122367101303913,
      "loss": 2.822,
      "theoretical_loss": 3.9300207443768773,
      "tokens_seen": 482425856
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004312136409227683,
      "loss": 3.0268,
      "theoretical_loss": 3.929962890326258,
      "tokens_seen": 482491392
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004312036108324975,
      "loss": 3.1638,
      "theoretical_loss": 3.9299050463332845,
      "tokens_seen": 482556928
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004311935807422267,
      "loss": 3.1264,
      "theoretical_loss": 3.929847212394843,
      "tokens_seen": 482622464
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00043118355065195586,
      "loss": 2.8405,
      "theoretical_loss": 3.929789388507821,
      "tokens_seen": 482688000
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004311735205616851,
      "loss": 2.9131,
      "theoretical_loss": 3.9297315746691073,
      "tokens_seen": 482753536
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004311634904714142,
      "loss": 2.8616,
      "theoretical_loss": 3.9296737708755916,
      "tokens_seen": 482819072
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00043115346038114346,
      "loss": 3.1964,
      "theoretical_loss": 3.9296159771241657,
      "tokens_seen": 482884608
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004311434302908726,
      "loss": 3.1841,
      "theoretical_loss": 3.9295581934117223,
      "tokens_seen": 482950144
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004311334002006018,
      "loss": 2.7733,
      "theoretical_loss": 3.9295004197351564,
      "tokens_seen": 483015680
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.000431123370110331,
      "loss": 2.8998,
      "theoretical_loss": 3.9294426560913625,
      "tokens_seen": 483081216
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004311133400200602,
      "loss": 3.0662,
      "theoretical_loss": 3.9293849024772385,
      "tokens_seen": 483146752
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00043110330992978936,
      "loss": 3.1417,
      "theoretical_loss": 3.9293271588896825,
      "tokens_seen": 483212288
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004310932798395186,
      "loss": 3.0213,
      "theoretical_loss": 3.929269425325594,
      "tokens_seen": 483277824
    },
    {
      "epoch": 1.06,
      "objective/train/docs_used": 567309,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9936575889587402,
      "objective/train/theoretical_loss": 3.929226131728565,
      "objective/train/tokens_used": 503786976,
      "theoretical_loss": 3.929226131728565,
      "tokens_seen": 483326976
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004310832497492477,
      "loss": 2.9944,
      "theoretical_loss": 3.929211701781875,
      "tokens_seen": 483343360
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00043107321965897696,
      "loss": 2.9511,
      "theoretical_loss": 3.929153988255427,
      "tokens_seen": 483408896
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004310631895687061,
      "loss": 3.0354,
      "theoretical_loss": 3.9290962847431548,
      "tokens_seen": 483474432
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004310531594784353,
      "loss": 2.9623,
      "theoretical_loss": 3.9290385912419628,
      "tokens_seen": 483539968
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004310431293881645,
      "loss": 2.8225,
      "theoretical_loss": 3.9289809077487585,
      "tokens_seen": 483605504
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004310330992978937,
      "loss": 3.083,
      "theoretical_loss": 3.9289232342604494,
      "tokens_seen": 483671040
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00043102306920762286,
      "loss": 3.022,
      "theoretical_loss": 3.9288655707739455,
      "tokens_seen": 483736576
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00043101303911735205,
      "loss": 2.8435,
      "theoretical_loss": 3.928807917286157,
      "tokens_seen": 483802112
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004310030090270812,
      "loss": 2.8934,
      "theoretical_loss": 3.928750273793997,
      "tokens_seen": 483867648
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00043099297893681046,
      "loss": 3.0615,
      "theoretical_loss": 3.9286926402943774,
      "tokens_seen": 483933184
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004309829488465396,
      "loss": 2.8151,
      "theoretical_loss": 3.928635016784215,
      "tokens_seen": 483998720
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004309729187562688,
      "loss": 3.0473,
      "theoretical_loss": 3.9285774032604244,
      "tokens_seen": 484064256
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00043096288866599795,
      "loss": 3.172,
      "theoretical_loss": 3.9285197997199246,
      "tokens_seen": 484129792
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004309528585757272,
      "loss": 2.8463,
      "theoretical_loss": 3.928462206159634,
      "tokens_seen": 484195328
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00043094282848545637,
      "loss": 2.9347,
      "theoretical_loss": 3.928404622576472,
      "tokens_seen": 484260864
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00043093279839518555,
      "loss": 2.9781,
      "theoretical_loss": 3.928347048967362,
      "tokens_seen": 484326400
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004309227683049148,
      "loss": 3.0882,
      "theoretical_loss": 3.928289485329227,
      "tokens_seen": 484391936
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00043091273821464397,
      "loss": 2.8604,
      "theoretical_loss": 3.92823193165899,
      "tokens_seen": 484457472
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00043090270812437315,
      "loss": 2.9441,
      "theoretical_loss": 3.928174387953579,
      "tokens_seen": 484523008
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00043089267803410233,
      "loss": 2.9715,
      "theoretical_loss": 3.9281168542099194,
      "tokens_seen": 484588544
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004308826479438315,
      "loss": 3.1032,
      "theoretical_loss": 3.9280593304249405,
      "tokens_seen": 484654080
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004308726178535607,
      "loss": 2.9442,
      "theoretical_loss": 3.928001816595572,
      "tokens_seen": 484719616
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004308625877632899,
      "loss": 3.163,
      "theoretical_loss": 3.9279443127187452,
      "tokens_seen": 484785152
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00043085255767301905,
      "loss": 3.1574,
      "theoretical_loss": 3.9278868187913933,
      "tokens_seen": 484850688
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004308425275827483,
      "loss": 2.7887,
      "theoretical_loss": 3.9278293348104496,
      "tokens_seen": 484916224
    },
    {
      "epoch": 1.06,
      "objective/train/docs_used": 572265,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.1022114753723145,
      "objective/train/theoretical_loss": 3.927786228350229,
      "objective/train/tokens_used": 505425376,
      "theoretical_loss": 3.927786228350229,
      "tokens_seen": 484965376
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004308324974924774,
      "loss": 3.0504,
      "theoretical_loss": 3.9277718607728502,
      "tokens_seen": 484981760
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00043082246740220665,
      "loss": 2.8934,
      "theoretical_loss": 3.9277143966755315,
      "tokens_seen": 485047296
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00043081243731193583,
      "loss": 2.9485,
      "theoretical_loss": 3.927656942515431,
      "tokens_seen": 485112832
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.000430802407221665,
      "loss": 3.1336,
      "theoretical_loss": 3.927599498289489,
      "tokens_seen": 485178368
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004307923771313942,
      "loss": 2.9035,
      "theoretical_loss": 3.927542063994646,
      "tokens_seen": 485243904
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00043078234704112343,
      "loss": 3.1617,
      "theoretical_loss": 3.927484639627844,
      "tokens_seen": 485309440
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00043077231695085256,
      "loss": 2.9164,
      "theoretical_loss": 3.927427225186027,
      "tokens_seen": 485374976
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004307622868605818,
      "loss": 3.2708,
      "theoretical_loss": 3.927369820666139,
      "tokens_seen": 485440512
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004307522567703109,
      "loss": 3.16,
      "theoretical_loss": 3.9273124260651278,
      "tokens_seen": 485506048
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00043074222668004015,
      "loss": 3.1773,
      "theoretical_loss": 3.927255041379939,
      "tokens_seen": 485571584
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00043073219658976933,
      "loss": 2.9499,
      "theoretical_loss": 3.9271976666075226,
      "tokens_seen": 485637120
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004307221664994985,
      "loss": 2.79,
      "theoretical_loss": 3.9271403017448288,
      "tokens_seen": 485702656
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004307121364092277,
      "loss": 3.0427,
      "theoretical_loss": 3.9270829467888095,
      "tokens_seen": 485768192
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004307021063189569,
      "loss": 3.1896,
      "theoretical_loss": 3.9270256017364167,
      "tokens_seen": 485833728
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00043069207622868606,
      "loss": 3.217,
      "theoretical_loss": 3.9269682665846064,
      "tokens_seen": 485899264
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004306820461384153,
      "loss": 3.0031,
      "theoretical_loss": 3.9269109413303322,
      "tokens_seen": 485964800
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004306720160481444,
      "loss": 3.174,
      "theoretical_loss": 3.9268536259705527,
      "tokens_seen": 486030336
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00043066198595787366,
      "loss": 2.8777,
      "theoretical_loss": 3.926796320502225,
      "tokens_seen": 486095872
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004306519558676028,
      "loss": 3.1536,
      "theoretical_loss": 3.926739024922311,
      "tokens_seen": 486161408
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.000430641925777332,
      "loss": 2.9334,
      "theoretical_loss": 3.926681739227769,
      "tokens_seen": 486226944
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004306318956870612,
      "loss": 2.8147,
      "theoretical_loss": 3.926624463415563,
      "tokens_seen": 486292480
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004306218655967904,
      "loss": 2.9436,
      "theoretical_loss": 3.9265671974826564,
      "tokens_seen": 486358016
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00043061183550651956,
      "loss": 2.9908,
      "theoretical_loss": 3.9265099414260147,
      "tokens_seen": 486423552
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004306018054162488,
      "loss": 3.0091,
      "theoretical_loss": 3.9264526952426033,
      "tokens_seen": 486489088
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004305917753259779,
      "loss": 2.9953,
      "theoretical_loss": 3.926395458929391,
      "tokens_seen": 486554624
    },
    {
      "epoch": 1.06,
      "objective/train/docs_used": 577157,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9272384643554688,
      "objective/train/theoretical_loss": 3.9263525381699758,
      "objective/train/tokens_used": 507063776,
      "theoretical_loss": 3.9263525381699758,
      "tokens_seen": 486603776
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00043058174523570716,
      "loss": 3.0174,
      "theoretical_loss": 3.926338232483346,
      "tokens_seen": 486620160
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004305717151454363,
      "loss": 3.1603,
      "theoretical_loss": 3.9262810159014396,
      "tokens_seen": 486685696
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004305616850551655,
      "loss": 3.1405,
      "theoretical_loss": 3.926223809180643,
      "tokens_seen": 486751232
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004305516549648947,
      "loss": 2.9448,
      "theoretical_loss": 3.9261666123179295,
      "tokens_seen": 486816768
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004305416248746239,
      "loss": 3.2668,
      "theoretical_loss": 3.9261094253102735,
      "tokens_seen": 486882304
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00043053159478435306,
      "loss": 2.8592,
      "theoretical_loss": 3.926052248154651,
      "tokens_seen": 486947840
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00043052156469408225,
      "loss": 2.8944,
      "theoretical_loss": 3.925995080848039,
      "tokens_seen": 487013376
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004305115346038114,
      "loss": 2.9189,
      "theoretical_loss": 3.9259379233874157,
      "tokens_seen": 487078912
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00043050150451354066,
      "loss": 2.8615,
      "theoretical_loss": 3.9258807757697607,
      "tokens_seen": 487144448
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004304914744232698,
      "loss": 3.1765,
      "theoretical_loss": 3.9258236379920564,
      "tokens_seen": 487209984
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.000430481444332999,
      "loss": 2.7061,
      "theoretical_loss": 3.925766510051284,
      "tokens_seen": 487275520
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00043047141424272815,
      "loss": 2.9585,
      "theoretical_loss": 3.9257093919444275,
      "tokens_seen": 487341056
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004304613841524574,
      "loss": 3.07,
      "theoretical_loss": 3.925652283668472,
      "tokens_seen": 487406592
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00043045135406218657,
      "loss": 3.2781,
      "theoretical_loss": 3.9255951852204047,
      "tokens_seen": 487472128
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00043044132397191575,
      "loss": 2.9378,
      "theoretical_loss": 3.9255380965972124,
      "tokens_seen": 487537664
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00043043129388164493,
      "loss": 3.1473,
      "theoretical_loss": 3.9254810177958848,
      "tokens_seen": 487603200
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00043042126379137417,
      "loss": 2.8296,
      "theoretical_loss": 3.9254239488134117,
      "tokens_seen": 487668736
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004304112337011033,
      "loss": 3.0285,
      "theoretical_loss": 3.925366889646786,
      "tokens_seen": 487734272
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00043040120361083253,
      "loss": 2.8856,
      "theoretical_loss": 3.925309840292999,
      "tokens_seen": 487799808
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00043039117352056165,
      "loss": 2.8328,
      "theoretical_loss": 3.925252800749047,
      "tokens_seen": 487865344
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004303811434302909,
      "loss": 2.9618,
      "theoretical_loss": 3.925195771011924,
      "tokens_seen": 487930880
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00043037111334002007,
      "loss": 3.0661,
      "theoretical_loss": 3.925138751078629,
      "tokens_seen": 487996416
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00043036108324974925,
      "loss": 3.176,
      "theoretical_loss": 3.925081740946159,
      "tokens_seen": 488061952
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00043035105315947843,
      "loss": 3.0429,
      "theoretical_loss": 3.9250247406115135,
      "tokens_seen": 488127488
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004303410230692076,
      "loss": 3.0888,
      "theoretical_loss": 3.924967750071694,
      "tokens_seen": 488193024
    },
    {
      "epoch": 1.06,
      "objective/train/docs_used": 582212,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.2292799949645996,
      "objective/train/theoretical_loss": 3.9249250135928815,
      "objective/train/tokens_used": 508702176,
      "theoretical_loss": 3.9249250135928815,
      "tokens_seen": 488242176
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004303309929789368,
      "loss": 2.9761,
      "theoretical_loss": 3.9249107693237035,
      "tokens_seen": 488258560
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00043032096288866603,
      "loss": 2.9192,
      "theoretical_loss": 3.924853798364545,
      "tokens_seen": 488324096
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00043031093279839516,
      "loss": 3.129,
      "theoretical_loss": 3.9247968371912227,
      "tokens_seen": 488389632
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004303009027081244,
      "loss": 3.078,
      "theoretical_loss": 3.9247398858007445,
      "tokens_seen": 488455168
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004302908726178535,
      "loss": 3.0365,
      "theoretical_loss": 3.924682944190117,
      "tokens_seen": 488520704
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00043028084252758276,
      "loss": 2.8859,
      "theoretical_loss": 3.924626012356349,
      "tokens_seen": 488586240
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00043027081243731194,
      "loss": 3.0156,
      "theoretical_loss": 3.924569090296451,
      "tokens_seen": 488651776
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004302607823470411,
      "loss": 3.0717,
      "theoretical_loss": 3.9245121780074355,
      "tokens_seen": 488717312
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004302507522567703,
      "loss": 2.8483,
      "theoretical_loss": 3.924455275486314,
      "tokens_seen": 488782848
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00043024072216649953,
      "loss": 2.7496,
      "theoretical_loss": 3.9243983827301006,
      "tokens_seen": 488848384
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00043023069207622866,
      "loss": 3.1034,
      "theoretical_loss": 3.924341499735812,
      "tokens_seen": 488913920
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004302206619859579,
      "loss": 2.8161,
      "theoretical_loss": 3.924284626500464,
      "tokens_seen": 488979456
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.000430210631895687,
      "loss": 3.0165,
      "theoretical_loss": 3.9242277630210753,
      "tokens_seen": 489044992
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00043020060180541626,
      "loss": 3.1262,
      "theoretical_loss": 3.9241709092946655,
      "tokens_seen": 489110528
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00043019057171514544,
      "loss": 3.0549,
      "theoretical_loss": 3.924114065318255,
      "tokens_seen": 489176064
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004301805416248746,
      "loss": 3.0324,
      "theoretical_loss": 3.9240572310888657,
      "tokens_seen": 489241600
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00043017051153460386,
      "loss": 2.9433,
      "theoretical_loss": 3.924000406603521,
      "tokens_seen": 489307136
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.000430160481444333,
      "loss": 2.8329,
      "theoretical_loss": 3.9239435918592456,
      "tokens_seen": 489372672
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004301504513540622,
      "loss": 3.0254,
      "theoretical_loss": 3.923886786853066,
      "tokens_seen": 489438208
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004301404212637914,
      "loss": 2.874,
      "theoretical_loss": 3.9238299915820085,
      "tokens_seen": 489503744
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004301303911735206,
      "loss": 2.8958,
      "theoretical_loss": 3.9237732060431023,
      "tokens_seen": 489569280
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00043012036108324976,
      "loss": 2.9936,
      "theoretical_loss": 3.9237164302333776,
      "tokens_seen": 489634816
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.000430110330992979,
      "loss": 3.0332,
      "theoretical_loss": 3.923659664149865,
      "tokens_seen": 489700352
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004301003009027081,
      "loss": 3.1645,
      "theoretical_loss": 3.923602907789597,
      "tokens_seen": 489765888
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00043009027081243736,
      "loss": 2.7748,
      "theoretical_loss": 3.9235461611496083,
      "tokens_seen": 489831424
    },
    {
      "epoch": 1.06,
      "objective/train/docs_used": 587295,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.1512694358825684,
      "objective/train/theoretical_loss": 3.923503607546766,
      "objective/train/tokens_used": 510340576,
      "theoretical_loss": 3.923503607546766,
      "tokens_seen": 489880576
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004300802407221665,
      "loss": 2.8325,
      "theoretical_loss": 3.9234894242269327,
      "tokens_seen": 489896960
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004300702106318957,
      "loss": 2.8625,
      "theoretical_loss": 3.9234326970186073,
      "tokens_seen": 489962496
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004300601805416249,
      "loss": 3.0172,
      "theoretical_loss": 3.92337597952167,
      "tokens_seen": 490028032
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004300501504513541,
      "loss": 3.0533,
      "theoretical_loss": 3.9233192717331598,
      "tokens_seen": 490093568
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00043004012036108326,
      "loss": 2.9677,
      "theoretical_loss": 3.923262573650116,
      "tokens_seen": 490159104
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00043003009027081245,
      "loss": 2.8643,
      "theoretical_loss": 3.9232058852695815,
      "tokens_seen": 490224640
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00043002006018054163,
      "loss": 3.1406,
      "theoretical_loss": 3.9231492065885982,
      "tokens_seen": 490290176
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00043001003009027086,
      "loss": 3.2728,
      "theoretical_loss": 3.9230925376042114,
      "tokens_seen": 490355712
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00043,
      "loss": 3.1938,
      "theoretical_loss": 3.923035878313465,
      "tokens_seen": 490421248
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004299899699097292,
      "loss": 2.9093,
      "theoretical_loss": 3.9229792287134075,
      "tokens_seen": 490486784
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042997993981945835,
      "loss": 2.8083,
      "theoretical_loss": 3.922922588801086,
      "tokens_seen": 490552320
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004299699097291876,
      "loss": 3.3187,
      "theoretical_loss": 3.9228659585735497,
      "tokens_seen": 490617856
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042995987963891677,
      "loss": 2.8643,
      "theoretical_loss": 3.92280933802785,
      "tokens_seen": 490683392
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042994984954864595,
      "loss": 2.8438,
      "theoretical_loss": 3.922752727161038,
      "tokens_seen": 490748928
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042993981945837513,
      "loss": 3.1228,
      "theoretical_loss": 3.922696125970168,
      "tokens_seen": 490814464
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042992978936810437,
      "loss": 2.8619,
      "theoretical_loss": 3.9226395344522933,
      "tokens_seen": 490880000
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004299197592778335,
      "loss": 3.0386,
      "theoretical_loss": 3.9225829526044707,
      "tokens_seen": 490945536
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042990972918756273,
      "loss": 2.8399,
      "theoretical_loss": 3.922526380423757,
      "tokens_seen": 491011072
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042989969909729185,
      "loss": 2.8468,
      "theoretical_loss": 3.922469817907211,
      "tokens_seen": 491076608
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004298896690070211,
      "loss": 2.9284,
      "theoretical_loss": 3.9224132650518913,
      "tokens_seen": 491142144
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042987963891675027,
      "loss": 2.9865,
      "theoretical_loss": 3.9223567218548596,
      "tokens_seen": 491207680
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042986960882647945,
      "loss": 2.9417,
      "theoretical_loss": 3.9223001883131783,
      "tokens_seen": 491273216
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042985957873620863,
      "loss": 3.009,
      "theoretical_loss": 3.9222436644239114,
      "tokens_seen": 491338752
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004298495486459378,
      "loss": 2.8589,
      "theoretical_loss": 3.922187150184122,
      "tokens_seen": 491404288
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.000429839518555667,
      "loss": 2.9139,
      "theoretical_loss": 3.9221306455908778,
      "tokens_seen": 491469824
    },
    {
      "debugging/Self-BLEU-5": 0.5154921509631192,
      "debugging/distinct-1-grams": 0.734842616282129,
      "debugging/distinct-2-grams": 0.8846155258692667,
      "debugging/entropy-1-grams": 6.056339442041576,
      "debugging/entropy-2-grams": 7.145308721973038,
      "debugging/length": 484.85,
      "debugging/num_segments": 20,
      "debugging/score": 0.003003109095205246,
      "debugging/score_std": 0.004943536934620138,
      "epoch": 1.06,
      "objective/train/docs_used": 590236,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.2846457958221436,
      "objective/train/theoretical_loss": 3.922088273474726,
      "objective/train/tokens_used": 511978976,
      "theoretical_loss": 3.922088273474726,
      "tokens_seen": 491518976
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042982948846539623,
      "loss": 3.0361,
      "theoretical_loss": 3.9220741506412464,
      "tokens_seen": 491535360
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042981945837512536,
      "loss": 2.8973,
      "theoretical_loss": 3.9220176653322953,
      "tokens_seen": 491600896
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004298094282848546,
      "loss": 2.7277,
      "theoretical_loss": 3.9219611896610944,
      "tokens_seen": 491666432
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004297993981945837,
      "loss": 2.8928,
      "theoretical_loss": 3.921904723624716,
      "tokens_seen": 491731968
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042978936810431296,
      "loss": 2.8518,
      "theoretical_loss": 3.9218482672202324,
      "tokens_seen": 491797504
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042977933801404214,
      "loss": 3.1745,
      "theoretical_loss": 3.9217918204447173,
      "tokens_seen": 491863040
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004297693079237713,
      "loss": 2.4701,
      "theoretical_loss": 3.9217353832952453,
      "tokens_seen": 491928576
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004297592778335005,
      "loss": 3.0984,
      "theoretical_loss": 3.921678955768893,
      "tokens_seen": 491994112
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042974924774322973,
      "loss": 3.0063,
      "theoretical_loss": 3.9216225378627385,
      "tokens_seen": 492059648
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042973921765295886,
      "loss": 2.8113,
      "theoretical_loss": 3.92156612957386,
      "tokens_seen": 492125184
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004297291875626881,
      "loss": 3.1234,
      "theoretical_loss": 3.921509730899338,
      "tokens_seen": 492190720
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004297191574724172,
      "loss": 2.9805,
      "theoretical_loss": 3.9214533418362545,
      "tokens_seen": 492256256
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042970912738214646,
      "loss": 2.7724,
      "theoretical_loss": 3.9213969623816913,
      "tokens_seen": 492321792
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042969909729187564,
      "loss": 2.9657,
      "theoretical_loss": 3.921340592532733,
      "tokens_seen": 492387328
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004296890672016048,
      "loss": 3.0349,
      "theoretical_loss": 3.921284232286465,
      "tokens_seen": 492452864
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.000429679037111334,
      "loss": 3.1185,
      "theoretical_loss": 3.9212278816399735,
      "tokens_seen": 492518400
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004296690070210632,
      "loss": 3.1046,
      "theoretical_loss": 3.9211715405903464,
      "tokens_seen": 492583936
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042965897693079236,
      "loss": 2.7548,
      "theoretical_loss": 3.9211152091346735,
      "tokens_seen": 492649472
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004296489468405216,
      "loss": 2.9028,
      "theoretical_loss": 3.9210588872700436,
      "tokens_seen": 492715008
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004296389167502507,
      "loss": 2.9545,
      "theoretical_loss": 3.9210025749935493,
      "tokens_seen": 492780544
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042962888665997996,
      "loss": 2.8856,
      "theoretical_loss": 3.9209462723022837,
      "tokens_seen": 492846080
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004296188565697091,
      "loss": 2.8481,
      "theoretical_loss": 3.9208899791933414,
      "tokens_seen": 492911616
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004296088264794383,
      "loss": 3.013,
      "theoretical_loss": 3.9208336956638163,
      "tokens_seen": 492977152
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004295987963891675,
      "loss": 2.8045,
      "theoretical_loss": 3.920777421710807,
      "tokens_seen": 493042688
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004295887662988967,
      "loss": 2.8337,
      "theoretical_loss": 3.9207211573314096,
      "tokens_seen": 493108224
    },
    {
      "epoch": 1.06,
      "objective/train/docs_used": 590917,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5917956829071045,
      "objective/train/theoretical_loss": 3.9206789653278,
      "objective/train/tokens_used": 513617376,
      "theoretical_loss": 3.9206789653278,
      "tokens_seen": 493157376
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042957873620862587,
      "loss": 2.8503,
      "theoretical_loss": 3.9206649025227245,
      "tokens_seen": 493173760
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004295687061183551,
      "loss": 3.2481,
      "theoretical_loss": 3.9206086572818517,
      "tokens_seen": 493239296
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042955867602808423,
      "loss": 3.1467,
      "theoretical_loss": 3.920552421605894,
      "tokens_seen": 493304832
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042954864593781347,
      "loss": 2.8639,
      "theoretical_loss": 3.920496195491953,
      "tokens_seen": 493370368
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004295386158475426,
      "loss": 3.1696,
      "theoretical_loss": 3.920439978937134,
      "tokens_seen": 493435904
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042952858575727183,
      "loss": 3.0995,
      "theoretical_loss": 3.9203837719385417,
      "tokens_seen": 493501440
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.000429518555667001,
      "loss": 2.6516,
      "theoretical_loss": 3.920327574493284,
      "tokens_seen": 493566976
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004295085255767302,
      "loss": 3.0273,
      "theoretical_loss": 3.920271386598468,
      "tokens_seen": 493632512
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042949849548645937,
      "loss": 2.7616,
      "theoretical_loss": 3.9202152082512036,
      "tokens_seen": 493698048
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042948846539618855,
      "loss": 2.9871,
      "theoretical_loss": 3.9201590394486012,
      "tokens_seen": 493763584
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042947843530591773,
      "loss": 2.8426,
      "theoretical_loss": 3.9201028801877724,
      "tokens_seen": 493829120
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042946840521564697,
      "loss": 2.9274,
      "theoretical_loss": 3.920046730465831,
      "tokens_seen": 493894656
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004294583751253761,
      "loss": 2.8587,
      "theoretical_loss": 3.919990590279891,
      "tokens_seen": 493960192
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042944834503510533,
      "loss": 2.7168,
      "theoretical_loss": 3.9199344596270675,
      "tokens_seen": 494025728
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004294383149448345,
      "loss": 2.8623,
      "theoretical_loss": 3.9198783385044784,
      "tokens_seen": 494091264
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004294282848545637,
      "loss": 2.8781,
      "theoretical_loss": 3.919822226909241,
      "tokens_seen": 494156800
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042941825476429293,
      "loss": 3.0756,
      "theoretical_loss": 3.919766124838475,
      "tokens_seen": 494222336
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042940822467402206,
      "loss": 2.9693,
      "theoretical_loss": 3.9197100322893013,
      "tokens_seen": 494287872
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004293981945837513,
      "loss": 2.8422,
      "theoretical_loss": 3.919653949258841,
      "tokens_seen": 494353408
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042938816449348047,
      "loss": 2.9434,
      "theoretical_loss": 3.919597875744218,
      "tokens_seen": 494418944
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042937813440320965,
      "loss": 2.9088,
      "theoretical_loss": 3.9195418117425564,
      "tokens_seen": 494484480
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042936810431293883,
      "loss": 2.8325,
      "theoretical_loss": 3.919485757250982,
      "tokens_seen": 494550016
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.000429358074222668,
      "loss": 2.8604,
      "theoretical_loss": 3.9194297122666213,
      "tokens_seen": 494615552
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004293480441323972,
      "loss": 2.9683,
      "theoretical_loss": 3.919373676786603,
      "tokens_seen": 494681088
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042933801404212643,
      "loss": 3.1226,
      "theoretical_loss": 3.919317650808056,
      "tokens_seen": 494746624
    },
    {
      "epoch": 1.06,
      "objective/train/docs_used": 592366,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.896491527557373,
      "objective/train/theoretical_loss": 3.91927563755776,
      "objective/train/tokens_used": 515255776,
      "theoretical_loss": 3.91927563755776,
      "tokens_seen": 494795776
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042932798395185556,
      "loss": 3.0733,
      "theoretical_loss": 3.9192616343281106,
      "tokens_seen": 494812160
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004293179538615848,
      "loss": 3.0851,
      "theoretical_loss": 3.9192056273438998,
      "tokens_seen": 494877696
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004293079237713139,
      "loss": 3.1179,
      "theoretical_loss": 3.919149629852556,
      "tokens_seen": 494943232
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042929789368104316,
      "loss": 3.011,
      "theoretical_loss": 3.919093641851214,
      "tokens_seen": 495008768
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042928786359077234,
      "loss": 2.7185,
      "theoretical_loss": 3.9190376633370088,
      "tokens_seen": 495074304
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004292778335005015,
      "loss": 3.0177,
      "theoretical_loss": 3.918981694307077,
      "tokens_seen": 495139840
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004292678034102307,
      "loss": 2.9071,
      "theoretical_loss": 3.9189257347585578,
      "tokens_seen": 495205376
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042925777331995993,
      "loss": 3.0857,
      "theoretical_loss": 3.9188697846885905,
      "tokens_seen": 495270912
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042924774322968906,
      "loss": 3.1196,
      "theoretical_loss": 3.9188138440943145,
      "tokens_seen": 495336448
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004292377131394183,
      "loss": 2.9395,
      "theoretical_loss": 3.9187579129728727,
      "tokens_seen": 495401984
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004292276830491474,
      "loss": 3.0981,
      "theoretical_loss": 3.9187019913214076,
      "tokens_seen": 495467520
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042921765295887666,
      "loss": 2.8836,
      "theoretical_loss": 3.918646079137064,
      "tokens_seen": 495533056
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042920762286860584,
      "loss": 3.0769,
      "theoretical_loss": 3.918590176416987,
      "tokens_seen": 495598592
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.000429197592778335,
      "loss": 2.9373,
      "theoretical_loss": 3.918534283158323,
      "tokens_seen": 495664128
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004291875626880642,
      "loss": 2.8688,
      "theoretical_loss": 3.9184783993582215,
      "tokens_seen": 495729664
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004291775325977934,
      "loss": 2.866,
      "theoretical_loss": 3.91842252501383,
      "tokens_seen": 495795200
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042916750250752256,
      "loss": 3.0718,
      "theoretical_loss": 3.9183666601222997,
      "tokens_seen": 495860736
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004291574724172518,
      "loss": 2.7853,
      "theoretical_loss": 3.9183108046807833,
      "tokens_seen": 495926272
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004291474423269809,
      "loss": 2.8375,
      "theoretical_loss": 3.918254958686432,
      "tokens_seen": 495991808
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042913741223671016,
      "loss": 3.17,
      "theoretical_loss": 3.918199122136401,
      "tokens_seen": 496057344
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004291273821464393,
      "loss": 2.9484,
      "theoretical_loss": 3.9181432950278463,
      "tokens_seen": 496122880
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004291173520561685,
      "loss": 2.966,
      "theoretical_loss": 3.9180874773579233,
      "tokens_seen": 496188416
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004291073219658977,
      "loss": 2.9595,
      "theoretical_loss": 3.91803166912379,
      "tokens_seen": 496253952
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004290972918756269,
      "loss": 2.6767,
      "theoretical_loss": 3.917975870322607,
      "tokens_seen": 496319488
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042908726178535607,
      "loss": 2.8961,
      "theoretical_loss": 3.917920080951533,
      "tokens_seen": 496385024
    },
    {
      "epoch": 1.06,
      "objective/train/docs_used": 593160,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7651126384735107,
      "objective/train/theoretical_loss": 3.9178782451100296,
      "objective/train/tokens_used": 516894176,
      "theoretical_loss": 3.9178782451100296,
      "tokens_seen": 496434176
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004290772316950853,
      "loss": 3.037,
      "theoretical_loss": 3.9178643010077305,
      "tokens_seen": 496450560
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042906720160481443,
      "loss": 2.8813,
      "theoretical_loss": 3.9178085304883616,
      "tokens_seen": 496516096
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042905717151454367,
      "loss": 2.8188,
      "theoretical_loss": 3.9177527693905914,
      "tokens_seen": 496581632
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004290471414242728,
      "loss": 3.0604,
      "theoretical_loss": 3.9176970177115837,
      "tokens_seen": 496647168
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042903711133400203,
      "loss": 2.8691,
      "theoretical_loss": 3.9176412754485064,
      "tokens_seen": 496712704
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004290270812437312,
      "loss": 2.9651,
      "theoretical_loss": 3.917585542598527,
      "tokens_seen": 496778240
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004290170511534604,
      "loss": 3.1923,
      "theoretical_loss": 3.9175298191588137,
      "tokens_seen": 496843776
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042900702106318957,
      "loss": 2.8673,
      "theoretical_loss": 3.9174741051265367,
      "tokens_seen": 496909312
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042899699097291875,
      "loss": 2.9436,
      "theoretical_loss": 3.9174184004988684,
      "tokens_seen": 496974848
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042898696088264793,
      "loss": 3.2724,
      "theoretical_loss": 3.9173627052729803,
      "tokens_seen": 497040384
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042897693079237717,
      "loss": 3.1074,
      "theoretical_loss": 3.9173070194460466,
      "tokens_seen": 497105920
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004289669007021063,
      "loss": 2.9554,
      "theoretical_loss": 3.9172513430152427,
      "tokens_seen": 497171456
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042895687061183553,
      "loss": 2.8133,
      "theoretical_loss": 3.9171956759777453,
      "tokens_seen": 497236992
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004289468405215647,
      "loss": 2.9769,
      "theoretical_loss": 3.9171400183307306,
      "tokens_seen": 497302528
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004289368104312939,
      "loss": 3.1201,
      "theoretical_loss": 3.917084370071378,
      "tokens_seen": 497368064
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004289267803410231,
      "loss": 2.8267,
      "theoretical_loss": 3.9170287311968677,
      "tokens_seen": 497433600
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042891675025075226,
      "loss": 2.8601,
      "theoretical_loss": 3.9169731017043805,
      "tokens_seen": 497499136
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042890672016048144,
      "loss": 3.0935,
      "theoretical_loss": 3.9169174815910983,
      "tokens_seen": 497564672
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042889669007021067,
      "loss": 2.785,
      "theoretical_loss": 3.916861870854206,
      "tokens_seen": 497630208
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004288866599799398,
      "loss": 3.1246,
      "theoretical_loss": 3.9168062694908876,
      "tokens_seen": 497695744
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042887662988966903,
      "loss": 3.036,
      "theoretical_loss": 3.916750677498329,
      "tokens_seen": 497761280
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042886659979939816,
      "loss": 3.0575,
      "theoretical_loss": 3.916695094873718,
      "tokens_seen": 497826816
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004288565697091274,
      "loss": 2.9982,
      "theoretical_loss": 3.9166395216142424,
      "tokens_seen": 497892352
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004288465396188566,
      "loss": 3.0071,
      "theoretical_loss": 3.9165839577170924,
      "tokens_seen": 497957888
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042883650952858576,
      "loss": 2.8608,
      "theoretical_loss": 3.916528403179459,
      "tokens_seen": 498023424
    },
    {
      "epoch": 1.06,
      "objective/train/docs_used": 593907,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9000375270843506,
      "objective/train/theoretical_loss": 3.916486743416727,
      "objective/train/tokens_used": 518532576,
      "theoretical_loss": 3.916486743416727,
      "tokens_seen": 498072576
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042882647943831494,
      "loss": 2.9289,
      "theoretical_loss": 3.9164728579985333,
      "tokens_seen": 498088960
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004288164493480441,
      "loss": 2.7877,
      "theoretical_loss": 3.91641732217151,
      "tokens_seen": 498154496
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004288064192577733,
      "loss": 2.9093,
      "theoretical_loss": 3.9163617956955825,
      "tokens_seen": 498220032
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042879638916750254,
      "loss": 2.9958,
      "theoretical_loss": 3.916306278567947,
      "tokens_seen": 498285568
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042878635907723166,
      "loss": 2.9177,
      "theoretical_loss": 3.9162507707858003,
      "tokens_seen": 498351104
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004287763289869609,
      "loss": 2.8919,
      "theoretical_loss": 3.916195272346341,
      "tokens_seen": 498416640
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004287662988966901,
      "loss": 2.8702,
      "theoretical_loss": 3.916139783246768,
      "tokens_seen": 498482176
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042875626880641926,
      "loss": 2.7663,
      "theoretical_loss": 3.916084303484282,
      "tokens_seen": 498547712
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042874623871614844,
      "loss": 2.8986,
      "theoretical_loss": 3.9160288330560844,
      "tokens_seen": 498613248
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004287362086258776,
      "loss": 2.9018,
      "theoretical_loss": 3.9159733719593786,
      "tokens_seen": 498678784
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004287261785356068,
      "loss": 2.907,
      "theoretical_loss": 3.9159179201913688,
      "tokens_seen": 498744320
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042871614844533604,
      "loss": 2.955,
      "theoretical_loss": 3.9158624777492603,
      "tokens_seen": 498809856
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042870611835506517,
      "loss": 3.0871,
      "theoretical_loss": 3.9158070446302595,
      "tokens_seen": 498875392
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004286960882647944,
      "loss": 2.8137,
      "theoretical_loss": 3.915751620831575,
      "tokens_seen": 498940928
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042868605817452353,
      "loss": 2.955,
      "theoretical_loss": 3.9156962063504146,
      "tokens_seen": 499006464
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042867602808425276,
      "loss": 2.7694,
      "theoretical_loss": 3.915640801183989,
      "tokens_seen": 499072000
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.000428665997993982,
      "loss": 3.0629,
      "theoretical_loss": 3.9155854053295105,
      "tokens_seen": 499137536
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004286559679037111,
      "loss": 2.9166,
      "theoretical_loss": 3.91553001878419,
      "tokens_seen": 499203072
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042864593781344036,
      "loss": 2.9496,
      "theoretical_loss": 3.915474641545243,
      "tokens_seen": 499268608
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004286359077231695,
      "loss": 2.9411,
      "theoretical_loss": 3.915419273609883,
      "tokens_seen": 499334144
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004286258776328987,
      "loss": 2.7409,
      "theoretical_loss": 3.9153639149753277,
      "tokens_seen": 499399680
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004286158475426279,
      "loss": 2.7517,
      "theoretical_loss": 3.915308565638793,
      "tokens_seen": 499465216
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004286058174523571,
      "loss": 2.8606,
      "theoretical_loss": 3.9152532255974983,
      "tokens_seen": 499530752
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042859578736208627,
      "loss": 3.058,
      "theoretical_loss": 3.9151978948486637,
      "tokens_seen": 499596288
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004285857572718155,
      "loss": 3.086,
      "theoretical_loss": 3.9151425733895096,
      "tokens_seen": 499661824
    },
    {
      "epoch": 1.06,
      "objective/train/docs_used": 595290,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.1176106929779053,
      "objective/train/theoretical_loss": 3.915101088389826,
      "objective/train/tokens_used": 520170976,
      "theoretical_loss": 3.915101088389826,
      "tokens_seen": 499710976
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042857572718154463,
      "loss": 3.1203,
      "theoretical_loss": 3.9150872612172583,
      "tokens_seen": 499727360
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042856569709127387,
      "loss": 2.8166,
      "theoretical_loss": 3.9150319583291333,
      "tokens_seen": 499792896
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.000428555667001003,
      "loss": 3.111,
      "theoretical_loss": 3.914976664722359,
      "tokens_seen": 499858432
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042854563691073223,
      "loss": 3.1595,
      "theoretical_loss": 3.914921380394162,
      "tokens_seen": 499923968
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004285356068204614,
      "loss": 2.9206,
      "theoretical_loss": 3.914866105341768,
      "tokens_seen": 499989504
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004285255767301906,
      "loss": 2.7353,
      "theoretical_loss": 3.914810839562406,
      "tokens_seen": 500055040
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042851554663991977,
      "loss": 3.1139,
      "theoretical_loss": 3.9147555830533047,
      "tokens_seen": 500120576
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042850551654964895,
      "loss": 2.973,
      "theoretical_loss": 3.9147003358116956,
      "tokens_seen": 500186112
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042849548645937813,
      "loss": 2.8314,
      "theoretical_loss": 3.9146450978348093,
      "tokens_seen": 500251648
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042848545636910737,
      "loss": 2.9281,
      "theoretical_loss": 3.9145898691198795,
      "tokens_seen": 500317184
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004284754262788365,
      "loss": 2.805,
      "theoretical_loss": 3.91453464966414,
      "tokens_seen": 500382720
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042846539618856573,
      "loss": 2.773,
      "theoretical_loss": 3.9144794394648263,
      "tokens_seen": 500448256
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004284553660982949,
      "loss": 2.9605,
      "theoretical_loss": 3.9144242385191745,
      "tokens_seen": 500513792
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004284453360080241,
      "loss": 2.9448,
      "theoretical_loss": 3.9143690468244223,
      "tokens_seen": 500579328
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004284353059177533,
      "loss": 3.1425,
      "theoretical_loss": 3.9143138643778093,
      "tokens_seen": 500644864
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042842527582748246,
      "loss": 2.8884,
      "theoretical_loss": 3.914258691176575,
      "tokens_seen": 500710400
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042841524573721164,
      "loss": 3.1058,
      "theoretical_loss": 3.9142035272179605,
      "tokens_seen": 500775936
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042840521564694087,
      "loss": 2.7515,
      "theoretical_loss": 3.9141483724992083,
      "tokens_seen": 500841472
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042839518555667,
      "loss": 3.033,
      "theoretical_loss": 3.914093227017562,
      "tokens_seen": 500907008
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042838515546639923,
      "loss": 2.963,
      "theoretical_loss": 3.9140380907702665,
      "tokens_seen": 500972544
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042837512537612836,
      "loss": 3.0991,
      "theoretical_loss": 3.9139829637545676,
      "tokens_seen": 501038080
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004283650952858576,
      "loss": 2.987,
      "theoretical_loss": 3.913927845967712,
      "tokens_seen": 501103616
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004283550651955868,
      "loss": 3.0178,
      "theoretical_loss": 3.9138727374069493,
      "tokens_seen": 501169152
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042834503510531596,
      "loss": 3.0842,
      "theoretical_loss": 3.9138176380695278,
      "tokens_seen": 501234688
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042833500501504514,
      "loss": 3.0508,
      "theoretical_loss": 3.9137625479526985,
      "tokens_seen": 501300224
    },
    {
      "epoch": 1.06,
      "objective/train/docs_used": 595812,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.282951831817627,
      "objective/train/theoretical_loss": 3.913721236414437,
      "objective/train/tokens_used": 521809376,
      "theoretical_loss": 3.913721236414437,
      "tokens_seen": 501349376
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004283249749247743,
      "loss": 3.1562,
      "theoretical_loss": 3.9137074670537135,
      "tokens_seen": 501365760
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004283149448345035,
      "loss": 3.1112,
      "theoretical_loss": 3.9136523953698257,
      "tokens_seen": 501431296
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042830491474423274,
      "loss": 2.9509,
      "theoretical_loss": 3.913597332898289,
      "tokens_seen": 501496832
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042829488465396186,
      "loss": 2.7248,
      "theoretical_loss": 3.9135422796363595,
      "tokens_seen": 501562368
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004282848545636911,
      "loss": 3.0696,
      "theoretical_loss": 3.913487235581293,
      "tokens_seen": 501627904
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004282748244734203,
      "loss": 2.7844,
      "theoretical_loss": 3.913432200730348,
      "tokens_seen": 501693440
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042826479438314946,
      "loss": 2.8522,
      "theoretical_loss": 3.913377175080783,
      "tokens_seen": 501758976
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042825476429287864,
      "loss": 2.7476,
      "theoretical_loss": 3.9133221586298577,
      "tokens_seen": 501824512
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004282447342026078,
      "loss": 2.8357,
      "theoretical_loss": 3.913267151374834,
      "tokens_seen": 501890048
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.000428234704112337,
      "loss": 2.9961,
      "theoretical_loss": 3.913212153312974,
      "tokens_seen": 501955584
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042822467402206624,
      "loss": 3.0573,
      "theoretical_loss": 3.9131571644415413,
      "tokens_seen": 502021120
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042821464393179537,
      "loss": 2.9952,
      "theoretical_loss": 3.913102184757801,
      "tokens_seen": 502086656
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004282046138415246,
      "loss": 2.7377,
      "theoretical_loss": 3.9130472142590187,
      "tokens_seen": 502152192
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042819458375125373,
      "loss": 2.9165,
      "theoretical_loss": 3.9129922529424617,
      "tokens_seen": 502217728
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042818455366098296,
      "loss": 2.8041,
      "theoretical_loss": 3.9129373008053983,
      "tokens_seen": 502283264
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042817452357071215,
      "loss": 2.9187,
      "theoretical_loss": 3.9128823578450977,
      "tokens_seen": 502348800
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004281644934804413,
      "loss": 2.8599,
      "theoretical_loss": 3.9128274240588308,
      "tokens_seen": 502414336
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004281544633901705,
      "loss": 3.1846,
      "theoretical_loss": 3.9127724994438693,
      "tokens_seen": 502479872
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004281444332998997,
      "loss": 2.9606,
      "theoretical_loss": 3.9127175839974866,
      "tokens_seen": 502545408
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042813440320962887,
      "loss": 2.8654,
      "theoretical_loss": 3.912662677716956,
      "tokens_seen": 502610944
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004281243731193581,
      "loss": 3.0389,
      "theoretical_loss": 3.9126077805995534,
      "tokens_seen": 502676480
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042811434302908723,
      "loss": 3.0599,
      "theoretical_loss": 3.912552892642555,
      "tokens_seen": 502742016
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042810431293881647,
      "loss": 2.8393,
      "theoretical_loss": 3.912498013843238,
      "tokens_seen": 502807552
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042809428284854565,
      "loss": 2.9673,
      "theoretical_loss": 3.9124431441988823,
      "tokens_seen": 502873088
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042808425275827483,
      "loss": 2.9235,
      "theoretical_loss": 3.9123882837067674,
      "tokens_seen": 502938624
    },
    {
      "epoch": 1.06,
      "objective/train/docs_used": 597072,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4846396446228027,
      "objective/train/theoretical_loss": 3.9123471443422035,
      "objective/train/tokens_used": 523447776,
      "theoretical_loss": 3.9123471443422035,
      "tokens_seen": 502987776
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.000428074222668004,
      "loss": 2.9092,
      "theoretical_loss": 3.912333432364174,
      "tokens_seen": 503004160
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004280641925777332,
      "loss": 2.7859,
      "theoretical_loss": 3.912278590168385,
      "tokens_seen": 503069696
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004280541624874624,
      "loss": 2.85,
      "theoretical_loss": 3.9122237571166827,
      "tokens_seen": 503135232
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004280441323971916,
      "loss": 2.9421,
      "theoretical_loss": 3.912168933206353,
      "tokens_seen": 503200768
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042803410230692074,
      "loss": 3.1732,
      "theoretical_loss": 3.9121141184346806,
      "tokens_seen": 503266304
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042802407221664997,
      "loss": 3.1852,
      "theoretical_loss": 3.912059312798953,
      "tokens_seen": 503331840
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004280140421263791,
      "loss": 3.0069,
      "theoretical_loss": 3.9120045162964594,
      "tokens_seen": 503397376
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042800401203610833,
      "loss": 3.0059,
      "theoretical_loss": 3.911949728924487,
      "tokens_seen": 503462912
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004279939819458375,
      "loss": 3.1234,
      "theoretical_loss": 3.9118949506803267,
      "tokens_seen": 503528448
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004279839518555667,
      "loss": 2.9553,
      "theoretical_loss": 3.9118401815612707,
      "tokens_seen": 503593984
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004279739217652959,
      "loss": 2.8518,
      "theoretical_loss": 3.9117854215646117,
      "tokens_seen": 503659520
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004279638916750251,
      "loss": 3.1847,
      "theoretical_loss": 3.9117306706876427,
      "tokens_seen": 503725056
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042795386158475424,
      "loss": 2.9503,
      "theoretical_loss": 3.9116759289276595,
      "tokens_seen": 503790592
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004279438314944835,
      "loss": 2.8428,
      "theoretical_loss": 3.911621196281958,
      "tokens_seen": 503856128
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004279338014042126,
      "loss": 2.9084,
      "theoretical_loss": 3.9115664727478356,
      "tokens_seen": 503921664
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042792377131394184,
      "loss": 3.0378,
      "theoretical_loss": 3.9115117583225905,
      "tokens_seen": 503987200
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042791374122367107,
      "loss": 3.0881,
      "theoretical_loss": 3.911457053003523,
      "tokens_seen": 504052736
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004279037111334002,
      "loss": 2.8481,
      "theoretical_loss": 3.9114023567879332,
      "tokens_seen": 504118272
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042789368104312943,
      "loss": 2.9654,
      "theoretical_loss": 3.9113476696731233,
      "tokens_seen": 504183808
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042788365095285856,
      "loss": 3.1006,
      "theoretical_loss": 3.911292991656396,
      "tokens_seen": 504249344
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004278736208625878,
      "loss": 3.0438,
      "theoretical_loss": 3.9112383227350564,
      "tokens_seen": 504314880
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.000427863590772317,
      "loss": 3.0496,
      "theoretical_loss": 3.911183662906409,
      "tokens_seen": 504380416
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042785356068204616,
      "loss": 2.8254,
      "theoretical_loss": 3.911129012167761,
      "tokens_seen": 504445952
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042784353059177534,
      "loss": 2.6538,
      "theoretical_loss": 3.911074370516419,
      "tokens_seen": 504511488
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004278335005015045,
      "loss": 2.9649,
      "theoretical_loss": 3.911019737949693,
      "tokens_seen": 504577024
    },
    {
      "epoch": 1.06,
      "objective/train/docs_used": 597784,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.1036291122436523,
      "objective/train/theoretical_loss": 3.9109787694848093,
      "objective/train/tokens_used": 525086176,
      "theoretical_loss": 3.9109787694848093,
      "tokens_seen": 504626176
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004278234704112337,
      "loss": 3.0501,
      "theoretical_loss": 3.9109651144648927,
      "tokens_seen": 504642560
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042781344032096294,
      "loss": 3.2442,
      "theoretical_loss": 3.910910500059329,
      "tokens_seen": 504708096
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042780341023069206,
      "loss": 3.1424,
      "theoretical_loss": 3.910855894730314,
      "tokens_seen": 504773632
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004277933801404213,
      "loss": 2.6353,
      "theoretical_loss": 3.9108012984751612,
      "tokens_seen": 504839168
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004277833500501505,
      "loss": 3.0854,
      "theoretical_loss": 3.9107467112911856,
      "tokens_seen": 504904704
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042777331995987966,
      "loss": 3.162,
      "theoretical_loss": 3.9106921331757016,
      "tokens_seen": 504970240
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042776328986960884,
      "loss": 3.0313,
      "theoretical_loss": 3.910637564126028,
      "tokens_seen": 505035776
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.000427753259779338,
      "loss": 2.8933,
      "theoretical_loss": 3.9105830041394816,
      "tokens_seen": 505101312
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004277432296890672,
      "loss": 2.9269,
      "theoretical_loss": 3.910528453213381,
      "tokens_seen": 505166848
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042773319959879644,
      "loss": 2.9692,
      "theoretical_loss": 3.910473911345048,
      "tokens_seen": 505232384
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042772316950852557,
      "loss": 2.987,
      "theoretical_loss": 3.9104193785318024,
      "tokens_seen": 505297920
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004277131394182548,
      "loss": 2.7983,
      "theoretical_loss": 3.9103648547709673,
      "tokens_seen": 505363456
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042770310932798393,
      "loss": 2.9446,
      "theoretical_loss": 3.910310340059867,
      "tokens_seen": 505428992
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042769307923771316,
      "loss": 2.9765,
      "theoretical_loss": 3.9102558343958256,
      "tokens_seen": 505494528
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042768304914744235,
      "loss": 2.9479,
      "theoretical_loss": 3.9102013377761695,
      "tokens_seen": 505560064
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042767301905717153,
      "loss": 2.8593,
      "theoretical_loss": 3.910146850198225,
      "tokens_seen": 505625600
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004276629889669007,
      "loss": 3.1793,
      "theoretical_loss": 3.910092371659321,
      "tokens_seen": 505691136
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004276529588766299,
      "loss": 3.1078,
      "theoretical_loss": 3.9100379021567875,
      "tokens_seen": 505756672
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042764292878635907,
      "loss": 2.8893,
      "theoretical_loss": 3.9099834416879533,
      "tokens_seen": 505822208
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004276328986960883,
      "loss": 3.161,
      "theoretical_loss": 3.909928990250151,
      "tokens_seen": 505887744
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042762286860581743,
      "loss": 2.8981,
      "theoretical_loss": 3.9098745478407135,
      "tokens_seen": 505953280
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042761283851554667,
      "loss": 3.054,
      "theoretical_loss": 3.909820114456975,
      "tokens_seen": 506018816
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042760280842527585,
      "loss": 2.8977,
      "theoretical_loss": 3.909765690096269,
      "tokens_seen": 506084352
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042759277833500503,
      "loss": 2.7607,
      "theoretical_loss": 3.9097112747559337,
      "tokens_seen": 506149888
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004275827482447342,
      "loss": 2.8732,
      "theoretical_loss": 3.909656868433305,
      "tokens_seen": 506215424
    },
    {
      "epoch": 1.06,
      "objective/train/docs_used": 599198,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.1902496814727783,
      "objective/train/theoretical_loss": 3.9096160696076026,
      "objective/train/tokens_used": 526724576,
      "theoretical_loss": 3.9096160696076026,
      "tokens_seen": 506264576
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004275727181544634,
      "loss": 2.8656,
      "theoretical_loss": 3.9096024711257216,
      "tokens_seen": 506280960
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004275626880641926,
      "loss": 2.7504,
      "theoretical_loss": 3.909548082830523,
      "tokens_seen": 506346496
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004275526579739218,
      "loss": 3.0306,
      "theoretical_loss": 3.90949370354505,
      "tokens_seen": 506412032
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042754262788365094,
      "loss": 3.0496,
      "theoretical_loss": 3.909439333266645,
      "tokens_seen": 506477568
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042753259779338017,
      "loss": 2.7544,
      "theoretical_loss": 3.90938497199265,
      "tokens_seen": 506543104
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004275225677031093,
      "loss": 2.8873,
      "theoretical_loss": 3.909330619720409,
      "tokens_seen": 506608640
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042751253761283853,
      "loss": 3.0007,
      "theoretical_loss": 3.909276276447268,
      "tokens_seen": 506674176
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004275025075225677,
      "loss": 2.8218,
      "theoretical_loss": 3.909221942170573,
      "tokens_seen": 506739712
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004274924774322969,
      "loss": 2.8335,
      "theoretical_loss": 3.9091676168876717,
      "tokens_seen": 506805248
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004274824473420261,
      "loss": 2.826,
      "theoretical_loss": 3.9091133005959118,
      "tokens_seen": 506870784
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004274724172517553,
      "loss": 2.9372,
      "theoretical_loss": 3.9090589932926436,
      "tokens_seen": 506936320
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042746238716148444,
      "loss": 2.9096,
      "theoretical_loss": 3.909004694975218,
      "tokens_seen": 507001856
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004274523570712137,
      "loss": 2.9741,
      "theoretical_loss": 3.908950405640987,
      "tokens_seen": 507067392
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004274423269809428,
      "loss": 2.9571,
      "theoretical_loss": 3.908896125287303,
      "tokens_seen": 507132928
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042743229689067204,
      "loss": 2.7355,
      "theoretical_loss": 3.908841853911521,
      "tokens_seen": 507198464
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004274222668004012,
      "loss": 3.0665,
      "theoretical_loss": 3.9087875915109955,
      "tokens_seen": 507264000
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004274122367101304,
      "loss": 2.9467,
      "theoretical_loss": 3.908733338083084,
      "tokens_seen": 507329536
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004274022066198596,
      "loss": 2.7414,
      "theoretical_loss": 3.9086790936251425,
      "tokens_seen": 507395072
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042739217652958876,
      "loss": 2.9773,
      "theoretical_loss": 3.908624858134531,
      "tokens_seen": 507460608
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042738214643931794,
      "loss": 2.893,
      "theoretical_loss": 3.908570631608609,
      "tokens_seen": 507526144
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004273721163490472,
      "loss": 2.8147,
      "theoretical_loss": 3.9085164140447377,
      "tokens_seen": 507591680
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004273620862587763,
      "loss": 2.8449,
      "theoretical_loss": 3.908462205440278,
      "tokens_seen": 507657216
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042735205616850554,
      "loss": 2.9502,
      "theoretical_loss": 3.9084080057925936,
      "tokens_seen": 507722752
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042734202607823467,
      "loss": 2.9411,
      "theoretical_loss": 3.908353815099049,
      "tokens_seen": 507788288
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004273319959879639,
      "loss": 2.8535,
      "theoretical_loss": 3.90829963335701,
      "tokens_seen": 507853824
    },
    {
      "epoch": 1.06,
      "objective/train/docs_used": 599925,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.022632360458374,
      "objective/train/theoretical_loss": 3.9082590029233213,
      "objective/train/tokens_used": 528362976,
      "theoretical_loss": 3.9082590029233213,
      "tokens_seen": 507902976
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004273219658976931,
      "loss": 3.1551,
      "theoretical_loss": 3.908245460563842,
      "tokens_seen": 507919360
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042731193580742226,
      "loss": 2.7927,
      "theoretical_loss": 3.908191296716913,
      "tokens_seen": 507984896
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042730190571715145,
      "loss": 2.7956,
      "theoretical_loss": 3.908137141813592,
      "tokens_seen": 508050432
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004272918756268807,
      "loss": 2.7378,
      "theoretical_loss": 3.9080829958512484,
      "tokens_seen": 508115968
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004272818455366098,
      "loss": 2.8576,
      "theoretical_loss": 3.9080288588272536,
      "tokens_seen": 508181504
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042727181544633904,
      "loss": 2.6358,
      "theoretical_loss": 3.90797473073898,
      "tokens_seen": 508247040
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042726178535606817,
      "loss": 2.9043,
      "theoretical_loss": 3.9079206115837994,
      "tokens_seen": 508312576
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004272517552657974,
      "loss": 3.0203,
      "theoretical_loss": 3.907866501359087,
      "tokens_seen": 508378112
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004272417251755266,
      "loss": 2.8236,
      "theoretical_loss": 3.907812400062218,
      "tokens_seen": 508443648
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042723169508525577,
      "loss": 2.8062,
      "theoretical_loss": 3.9077583076905684,
      "tokens_seen": 508509184
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042722166499498495,
      "loss": 2.9367,
      "theoretical_loss": 3.907704224241517,
      "tokens_seen": 508574720
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042721163490471413,
      "loss": 2.758,
      "theoretical_loss": 3.907650149712442,
      "tokens_seen": 508640256
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004272016048144433,
      "loss": 2.8615,
      "theoretical_loss": 3.9075960841007227,
      "tokens_seen": 508705792
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042719157472417255,
      "loss": 2.8553,
      "theoretical_loss": 3.9075420274037405,
      "tokens_seen": 508771328
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004271815446339017,
      "loss": 2.7945,
      "theoretical_loss": 3.9074879796188773,
      "tokens_seen": 508836864
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004271715145436309,
      "loss": 2.718,
      "theoretical_loss": 3.907433940743516,
      "tokens_seen": 508902400
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004271614844533601,
      "loss": 3.0734,
      "theoretical_loss": 3.9073799107750413,
      "tokens_seen": 508967936
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042715145436308927,
      "loss": 2.8651,
      "theoretical_loss": 3.9073258897108385,
      "tokens_seen": 509033472
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004271414242728185,
      "loss": 2.684,
      "theoretical_loss": 3.9072718775482933,
      "tokens_seen": 509099008
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042713139418254763,
      "loss": 2.9408,
      "theoretical_loss": 3.907217874284794,
      "tokens_seen": 509164544
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042712136409227687,
      "loss": 2.7014,
      "theoretical_loss": 3.9071638799177295,
      "tokens_seen": 509230080
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042711133400200605,
      "loss": 3.0804,
      "theoretical_loss": 3.907109894444489,
      "tokens_seen": 509295616
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042710130391173523,
      "loss": 2.8745,
      "theoretical_loss": 3.907055917862463,
      "tokens_seen": 509361152
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004270912738214644,
      "loss": 3.088,
      "theoretical_loss": 3.9070019501690445,
      "tokens_seen": 509426688
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004270812437311936,
      "loss": 2.8545,
      "theoretical_loss": 3.9069479913616254,
      "tokens_seen": 509492224
    },
    {
      "epoch": 1.06,
      "objective/train/docs_used": 601205,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.895568370819092,
      "objective/train/theoretical_loss": 3.906907528085932,
      "objective/train/tokens_used": 530001376,
      "theoretical_loss": 3.906907528085932,
      "tokens_seen": 509541376
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004270712136409228,
      "loss": 2.8726,
      "theoretical_loss": 3.906894041437601,
      "tokens_seen": 509557760
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.000427061183550652,
      "loss": 2.6773,
      "theoretical_loss": 3.906840100394366,
      "tokens_seen": 509623296
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042705115346038114,
      "loss": 2.8422,
      "theoretical_loss": 3.9067861682293166,
      "tokens_seen": 509688832
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042704112337011037,
      "loss": 2.8928,
      "theoretical_loss": 3.9067322449398505,
      "tokens_seen": 509754368
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004270310932798395,
      "loss": 3.308,
      "theoretical_loss": 3.9066783305233663,
      "tokens_seen": 509819904
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042702106318956873,
      "loss": 3.0129,
      "theoretical_loss": 3.906624424977263,
      "tokens_seen": 509885440
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004270110330992979,
      "loss": 3.0955,
      "theoretical_loss": 3.9065705282989427,
      "tokens_seen": 509950976
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004270010030090271,
      "loss": 3.2208,
      "theoretical_loss": 3.9065166404858056,
      "tokens_seen": 510016512
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004269909729187563,
      "loss": 2.9995,
      "theoretical_loss": 3.906462761535256,
      "tokens_seen": 510082048
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004269809428284855,
      "loss": 2.6179,
      "theoretical_loss": 3.9064088914446975,
      "tokens_seen": 510147584
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042697091273821464,
      "loss": 2.9016,
      "theoretical_loss": 3.9063550302115346,
      "tokens_seen": 510213120
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004269608826479439,
      "loss": 2.7884,
      "theoretical_loss": 3.9063011778331744,
      "tokens_seen": 510278656
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.000426950852557673,
      "loss": 3.0448,
      "theoretical_loss": 3.9062473343070234,
      "tokens_seen": 510344192
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042694082246740224,
      "loss": 3.076,
      "theoretical_loss": 3.9061934996304912,
      "tokens_seen": 510409728
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004269307923771314,
      "loss": 2.9906,
      "theoretical_loss": 3.906139673800986,
      "tokens_seen": 510475264
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004269207622868606,
      "loss": 3.0174,
      "theoretical_loss": 3.9060858568159196,
      "tokens_seen": 510540800
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004269107321965898,
      "loss": 2.8787,
      "theoretical_loss": 3.906032048672702,
      "tokens_seen": 510606336
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042690070210631896,
      "loss": 2.8605,
      "theoretical_loss": 3.9059782493687476,
      "tokens_seen": 510671872
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042689067201604814,
      "loss": 3.0283,
      "theoretical_loss": 3.9059244589014694,
      "tokens_seen": 510737408
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004268806419257774,
      "loss": 2.945,
      "theoretical_loss": 3.9058706772682825,
      "tokens_seen": 510802944
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004268706118355065,
      "loss": 3.2505,
      "theoretical_loss": 3.905816904466603,
      "tokens_seen": 510868480
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042686058174523574,
      "loss": 3.0026,
      "theoretical_loss": 3.905763140493848,
      "tokens_seen": 510934016
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042685055165496487,
      "loss": 2.9411,
      "theoretical_loss": 3.905709385347436,
      "tokens_seen": 510999552
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004268405215646941,
      "loss": 3.2739,
      "theoretical_loss": 3.9056556390247863,
      "tokens_seen": 511065088
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0004268304914744233,
      "loss": 3.2656,
      "theoretical_loss": 3.905601901523318,
      "tokens_seen": 511130624
    },
    {
      "epoch": 1.06,
      "objective/train/docs_used": 601687,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9587132930755615,
      "objective/train/theoretical_loss": 3.905561604184567,
      "objective/train/tokens_used": 531639776,
      "theoretical_loss": 3.905561604184567,
      "tokens_seen": 511179776
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042682046138415246,
      "loss": 3.1384,
      "theoretical_loss": 3.9055481728404544,
      "tokens_seen": 511196160
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00042681043129388165,
      "loss": 3.1338,
      "theoretical_loss": 3.905494452973617,
      "tokens_seen": 511261696
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004268004012036109,
      "loss": 2.854,
      "theoretical_loss": 3.905440741920229,
      "tokens_seen": 511327232
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042679037111334,
      "loss": 2.9444,
      "theoretical_loss": 3.9053870396777164,
      "tokens_seen": 511392768
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042678034102306924,
      "loss": 2.9435,
      "theoretical_loss": 3.905333346243504,
      "tokens_seen": 511458304
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042677031093279837,
      "loss": 3.0561,
      "theoretical_loss": 3.905279661615019,
      "tokens_seen": 511523840
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004267602808425276,
      "loss": 2.9755,
      "theoretical_loss": 3.9052259857896887,
      "tokens_seen": 511589376
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004267502507522568,
      "loss": 2.9857,
      "theoretical_loss": 3.9051723187649436,
      "tokens_seen": 511654912
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042674022066198597,
      "loss": 3.0649,
      "theoretical_loss": 3.905118660538213,
      "tokens_seen": 511720448
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042673019057171515,
      "loss": 2.9349,
      "theoretical_loss": 3.905065011106928,
      "tokens_seen": 511785984
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042672016048144433,
      "loss": 3.0933,
      "theoretical_loss": 3.9050113704685208,
      "tokens_seen": 511851520
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004267101303911735,
      "loss": 3.0626,
      "theoretical_loss": 3.904957738620425,
      "tokens_seen": 511917056
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042670010030090275,
      "loss": 3.0902,
      "theoretical_loss": 3.9049041155600746,
      "tokens_seen": 511982592
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004266900702106319,
      "loss": 2.8052,
      "theoretical_loss": 3.9048505012849053,
      "tokens_seen": 512048128
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004266800401203611,
      "loss": 2.9866,
      "theoretical_loss": 3.9047968957923542,
      "tokens_seen": 512113664
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042667001003009024,
      "loss": 3.2107,
      "theoretical_loss": 3.9047432990798585,
      "tokens_seen": 512179200
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042665997993981947,
      "loss": 2.9144,
      "theoretical_loss": 3.9046897111448575,
      "tokens_seen": 512244736
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042664994984954865,
      "loss": 2.8282,
      "theoretical_loss": 3.90463613198479,
      "tokens_seen": 512310272
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042663991975927783,
      "loss": 3.0063,
      "theoretical_loss": 3.9045825615970973,
      "tokens_seen": 512375808
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.000426629889669007,
      "loss": 2.9466,
      "theoretical_loss": 3.904528999979222,
      "tokens_seen": 512441344
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042661985957873625,
      "loss": 2.9841,
      "theoretical_loss": 3.9044754471286063,
      "tokens_seen": 512506880
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004266098294884654,
      "loss": 2.7521,
      "theoretical_loss": 3.9044219030426945,
      "tokens_seen": 512572416
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004265997993981946,
      "loss": 2.8133,
      "theoretical_loss": 3.904368367718932,
      "tokens_seen": 512637952
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042658976930792374,
      "loss": 2.9629,
      "theoretical_loss": 3.904314841154765,
      "tokens_seen": 512703488
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.000426579739217653,
      "loss": 3.0293,
      "theoretical_loss": 3.9042613233476406,
      "tokens_seen": 512769024
    },
    {
      "epoch": 1.07,
      "objective/train/docs_used": 602985,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.070340633392334,
      "objective/train/theoretical_loss": 3.904221190737572,
      "objective/train/tokens_used": 533278176,
      "theoretical_loss": 3.904221190737572,
      "tokens_seen": 512818176
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042656970912738216,
      "loss": 2.8697,
      "theoretical_loss": 3.9042078142950074,
      "tokens_seen": 512834560
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042655967903711134,
      "loss": 2.8872,
      "theoretical_loss": 3.9041543139943147,
      "tokens_seen": 512900096
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004265496489468405,
      "loss": 3.0673,
      "theoretical_loss": 3.9041008224430134,
      "tokens_seen": 512965632
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004265396188565697,
      "loss": 2.8721,
      "theoretical_loss": 3.9040473396385544,
      "tokens_seen": 513031168
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004265295887662989,
      "loss": 2.8403,
      "theoretical_loss": 3.903993865578391,
      "tokens_seen": 513096704
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004265195586760281,
      "loss": 2.8301,
      "theoretical_loss": 3.9039404002599767,
      "tokens_seen": 513162240
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042650952858575724,
      "loss": 2.9424,
      "theoretical_loss": 3.903886943680766,
      "tokens_seen": 513227776
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004264994984954865,
      "loss": 3.0853,
      "theoretical_loss": 3.903833495838215,
      "tokens_seen": 513293312
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004264894684052156,
      "loss": 3.214,
      "theoretical_loss": 3.903780056729781,
      "tokens_seen": 513358848
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042647943831494484,
      "loss": 3.1121,
      "theoretical_loss": 3.9037266263529213,
      "tokens_seen": 513424384
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.000426469408224674,
      "loss": 3.105,
      "theoretical_loss": 3.903673204705096,
      "tokens_seen": 513489920
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004264593781344032,
      "loss": 2.6414,
      "theoretical_loss": 3.903619791783764,
      "tokens_seen": 513555456
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004264493480441324,
      "loss": 3.0392,
      "theoretical_loss": 3.903566387586387,
      "tokens_seen": 513620992
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004264393179538616,
      "loss": 2.8988,
      "theoretical_loss": 3.9035129921104277,
      "tokens_seen": 513686528
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042642928786359075,
      "loss": 2.9126,
      "theoretical_loss": 3.9034596053533486,
      "tokens_seen": 513752064
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042641925777332,
      "loss": 3.1472,
      "theoretical_loss": 3.9034062273126144,
      "tokens_seen": 513817600
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042640922768304916,
      "loss": 2.9302,
      "theoretical_loss": 3.9033528579856904,
      "tokens_seen": 513883136
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042639919759277834,
      "loss": 3.1095,
      "theoretical_loss": 3.9032994973700434,
      "tokens_seen": 513948672
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004263891675025076,
      "loss": 3.0708,
      "theoretical_loss": 3.9032461454631404,
      "tokens_seen": 514014208
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004263791374122367,
      "loss": 3.0222,
      "theoretical_loss": 3.9031928022624514,
      "tokens_seen": 514079744
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042636910732196594,
      "loss": 3.1103,
      "theoretical_loss": 3.9031394677654445,
      "tokens_seen": 514145280
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042635907723169507,
      "loss": 2.7347,
      "theoretical_loss": 3.903086141969591,
      "tokens_seen": 514210816
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004263490471414243,
      "loss": 2.9696,
      "theoretical_loss": 3.903032824872362,
      "tokens_seen": 514276352
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004263390170511535,
      "loss": 3.09,
      "theoretical_loss": 3.902979516471232,
      "tokens_seen": 514341888
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042632898696088266,
      "loss": 2.9927,
      "theoretical_loss": 3.9029262167636736,
      "tokens_seen": 514407424
    },
    {
      "epoch": 1.07,
      "objective/train/docs_used": 603651,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.3740158081054688,
      "objective/train/theoretical_loss": 3.902886247686642,
      "objective/train/tokens_used": 534916576,
      "theoretical_loss": 3.902886247686642,
      "tokens_seen": 514456576
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042631895687061185,
      "loss": 2.8779,
      "theoretical_loss": 3.9028729257471615,
      "tokens_seen": 514472960
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004263089267803411,
      "loss": 3.0081,
      "theoretical_loss": 3.902819643419173,
      "tokens_seen": 514538496
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004262988966900702,
      "loss": 3.0532,
      "theoretical_loss": 3.9027663697771837,
      "tokens_seen": 514604032
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042628886659979944,
      "loss": 3.1043,
      "theoretical_loss": 3.9027131048186727,
      "tokens_seen": 514669568
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042627883650952857,
      "loss": 2.8976,
      "theoretical_loss": 3.902659848541119,
      "tokens_seen": 514735104
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004262688064192578,
      "loss": 3.0731,
      "theoretical_loss": 3.9026066009420024,
      "tokens_seen": 514800640
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.000426258776328987,
      "loss": 2.867,
      "theoretical_loss": 3.902553362018805,
      "tokens_seen": 514866176
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042624874623871617,
      "loss": 2.911,
      "theoretical_loss": 3.902500131769008,
      "tokens_seen": 514931712
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042623871614844535,
      "loss": 3.2478,
      "theoretical_loss": 3.9024469101900956,
      "tokens_seen": 514997248
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042622868605817453,
      "loss": 2.9697,
      "theoretical_loss": 3.902393697279552,
      "tokens_seen": 515062784
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004262186559679037,
      "loss": 2.874,
      "theoretical_loss": 3.9023404930348624,
      "tokens_seen": 515128320
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042620862587763295,
      "loss": 2.9714,
      "theoretical_loss": 3.9022872974535137,
      "tokens_seen": 515193856
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004261985957873621,
      "loss": 3.0392,
      "theoretical_loss": 3.9022341105329934,
      "tokens_seen": 515259392
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004261885656970913,
      "loss": 2.9288,
      "theoretical_loss": 3.90218093227079,
      "tokens_seen": 515324928
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042617853560682044,
      "loss": 2.958,
      "theoretical_loss": 3.902127762664393,
      "tokens_seen": 515390464
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042616850551654967,
      "loss": 2.8225,
      "theoretical_loss": 3.902074601711294,
      "tokens_seen": 515456000
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042615847542627885,
      "loss": 2.7655,
      "theoretical_loss": 3.9020214494089833,
      "tokens_seen": 515521536
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042614844533600803,
      "loss": 2.8536,
      "theoretical_loss": 3.901968305754955,
      "tokens_seen": 515587072
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004261384152457372,
      "loss": 2.8864,
      "theoretical_loss": 3.901915170746702,
      "tokens_seen": 515652608
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042612838515546645,
      "loss": 2.9468,
      "theoretical_loss": 3.90186204438172,
      "tokens_seen": 515718144
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004261183550651956,
      "loss": 2.9706,
      "theoretical_loss": 3.9018089266575045,
      "tokens_seen": 515783680
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004261083249749248,
      "loss": 2.9346,
      "theoretical_loss": 3.9017558175715523,
      "tokens_seen": 515849216
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042609829488465394,
      "loss": 2.8477,
      "theoretical_loss": 3.9017027171213616,
      "tokens_seen": 515914752
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004260882647943832,
      "loss": 2.8322,
      "theoretical_loss": 3.9016496253044317,
      "tokens_seen": 515980288
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042607823470411236,
      "loss": 2.764,
      "theoretical_loss": 3.901596542118263,
      "tokens_seen": 516045824
    },
    {
      "epoch": 1.07,
      "objective/train/docs_used": 604730,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.402284622192383,
      "objective/train/theoretical_loss": 3.901556735391069,
      "objective/train/tokens_used": 536554976,
      "theoretical_loss": 3.901556735391069,
      "tokens_seen": 516094976
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042606820461384154,
      "loss": 3.1128,
      "theoretical_loss": 3.901543467560355,
      "tokens_seen": 516111360
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004260581745235707,
      "loss": 2.903,
      "theoretical_loss": 3.9014904016282124,
      "tokens_seen": 516176896
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004260481444332999,
      "loss": 3.158,
      "theoretical_loss": 3.901437344319336,
      "tokens_seen": 516242432
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004260381143430291,
      "loss": 2.6293,
      "theoretical_loss": 3.9013842956312317,
      "tokens_seen": 516307968
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004260280842527583,
      "loss": 2.7853,
      "theoretical_loss": 3.901331255561405,
      "tokens_seen": 516373504
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042601805416248744,
      "loss": 2.7866,
      "theoretical_loss": 3.9012782241073602,
      "tokens_seen": 516439040
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004260080240722167,
      "loss": 3.2774,
      "theoretical_loss": 3.9012252012666067,
      "tokens_seen": 516504576
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004259979939819458,
      "loss": 2.8639,
      "theoretical_loss": 3.901172187036652,
      "tokens_seen": 516570112
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042598796389167504,
      "loss": 2.9217,
      "theoretical_loss": 3.901119181415006,
      "tokens_seen": 516635648
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004259779338014042,
      "loss": 2.8692,
      "theoretical_loss": 3.901066184399179,
      "tokens_seen": 516701184
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004259679037111334,
      "loss": 2.8358,
      "theoretical_loss": 3.9010131959866827,
      "tokens_seen": 516766720
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004259578736208626,
      "loss": 2.9191,
      "theoretical_loss": 3.9009602161750294,
      "tokens_seen": 516832256
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004259478435305918,
      "loss": 2.8169,
      "theoretical_loss": 3.9009072449617324,
      "tokens_seen": 516897792
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042593781344032095,
      "loss": 3.0129,
      "theoretical_loss": 3.900854282344307,
      "tokens_seen": 516963328
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004259277833500502,
      "loss": 3.0345,
      "theoretical_loss": 3.9008013283202683,
      "tokens_seen": 517028864
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004259177532597793,
      "loss": 2.9059,
      "theoretical_loss": 3.9007483828871337,
      "tokens_seen": 517094400
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042590772316950854,
      "loss": 3.049,
      "theoretical_loss": 3.9006954460424206,
      "tokens_seen": 517159936
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004258976930792377,
      "loss": 2.6156,
      "theoretical_loss": 3.9006425177836475,
      "tokens_seen": 517225472
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004258876629889669,
      "loss": 2.9233,
      "theoretical_loss": 3.9005895981083345,
      "tokens_seen": 517291008
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004258776328986961,
      "loss": 2.7387,
      "theoretical_loss": 3.900536687014002,
      "tokens_seen": 517356544
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042586760280842527,
      "loss": 2.8999,
      "theoretical_loss": 3.9004837844981726,
      "tokens_seen": 517422080
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042585757271815445,
      "loss": 2.5944,
      "theoretical_loss": 3.9004308905583684,
      "tokens_seen": 517487616
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004258475426278837,
      "loss": 3.1158,
      "theoretical_loss": 3.900378005192114,
      "tokens_seen": 517553152
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004258375125376128,
      "loss": 2.851,
      "theoretical_loss": 3.900325128396934,
      "tokens_seen": 517618688
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042582748244734205,
      "loss": 2.6032,
      "theoretical_loss": 3.900272260170354,
      "tokens_seen": 517684224
    },
    {
      "epoch": 1.07,
      "objective/train/docs_used": 605280,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.3959805965423584,
      "objective/train/theoretical_loss": 3.900232614622076,
      "objective/train/tokens_used": 538193376,
      "theoretical_loss": 3.900232614622076,
      "tokens_seen": 517733376
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004258174523570712,
      "loss": 3.1391,
      "theoretical_loss": 3.9002194005099016,
      "tokens_seen": 517749760
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004258074222668004,
      "loss": 2.8197,
      "theoretical_loss": 3.900166549413105,
      "tokens_seen": 517815296
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004257973921765296,
      "loss": 3.117,
      "theoretical_loss": 3.900113706877492,
      "tokens_seen": 517880832
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042578736208625877,
      "loss": 2.9555,
      "theoretical_loss": 3.9000608729005943,
      "tokens_seen": 517946368
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042577733199598795,
      "loss": 2.7419,
      "theoretical_loss": 3.9000080474799415,
      "tokens_seen": 518011904
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004257673019057172,
      "loss": 2.6787,
      "theoretical_loss": 3.8999552306130667,
      "tokens_seen": 518077440
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004257572718154463,
      "loss": 2.8847,
      "theoretical_loss": 3.8999024222975036,
      "tokens_seen": 518142976
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042574724172517555,
      "loss": 2.898,
      "theoretical_loss": 3.899849622530785,
      "tokens_seen": 518208512
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004257372116349047,
      "loss": 3.0798,
      "theoretical_loss": 3.899796831310446,
      "tokens_seen": 518274048
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004257271815446339,
      "loss": 2.8054,
      "theoretical_loss": 3.899744048634024,
      "tokens_seen": 518339584
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004257171514543631,
      "loss": 2.847,
      "theoretical_loss": 3.899691274499056,
      "tokens_seen": 518405120
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004257071213640923,
      "loss": 2.9221,
      "theoretical_loss": 3.8996385089030796,
      "tokens_seen": 518470656
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042569709127382145,
      "loss": 2.92,
      "theoretical_loss": 3.899585751843635,
      "tokens_seen": 518536192
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042568706118355064,
      "loss": 2.9806,
      "theoretical_loss": 3.899533003318261,
      "tokens_seen": 518601728
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004256770310932798,
      "loss": 2.9909,
      "theoretical_loss": 3.8994802633245014,
      "tokens_seen": 518667264
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042566700100300905,
      "loss": 2.8064,
      "theoretical_loss": 3.899427531859896,
      "tokens_seen": 518732800
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042565697091273823,
      "loss": 2.6601,
      "theoretical_loss": 3.899374808921989,
      "tokens_seen": 518798336
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004256469408224674,
      "loss": 2.8185,
      "theoretical_loss": 3.8993220945083253,
      "tokens_seen": 518863872
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042563691073219665,
      "loss": 2.9848,
      "theoretical_loss": 3.8992693886164504,
      "tokens_seen": 518929408
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004256268806419258,
      "loss": 3.0078,
      "theoretical_loss": 3.8992166912439092,
      "tokens_seen": 518994944
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.000425616850551655,
      "loss": 2.8503,
      "theoretical_loss": 3.8991640023882512,
      "tokens_seen": 519060480
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042560682046138414,
      "loss": 3.1139,
      "theoretical_loss": 3.8991113220470233,
      "tokens_seen": 519126016
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004255967903711134,
      "loss": 2.8546,
      "theoretical_loss": 3.8990586502177758,
      "tokens_seen": 519191552
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042558676028084256,
      "loss": 2.8532,
      "theoretical_loss": 3.8990059868980587,
      "tokens_seen": 519257088
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042557673019057174,
      "loss": 2.891,
      "theoretical_loss": 3.8989533320854237,
      "tokens_seen": 519322624
    },
    {
      "epoch": 1.07,
      "objective/train/docs_used": 606732,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.3824806213378906,
      "objective/train/theoretical_loss": 3.8989138465572477,
      "objective/train/tokens_used": 539831776,
      "theoretical_loss": 3.8989138465572477,
      "tokens_seen": 519371776
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004255667001003009,
      "loss": 2.845,
      "theoretical_loss": 3.898900685777423,
      "tokens_seen": 519388160
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004255566700100301,
      "loss": 2.9526,
      "theoretical_loss": 3.8988480479716108,
      "tokens_seen": 519453696
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004255466399197593,
      "loss": 3.05,
      "theoretical_loss": 3.8987954186655402,
      "tokens_seen": 519519232
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004255366098294885,
      "loss": 2.9402,
      "theoretical_loss": 3.8987427978567686,
      "tokens_seen": 519584768
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042552657973921764,
      "loss": 2.8061,
      "theoretical_loss": 3.898690185542852,
      "tokens_seen": 519650304
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004255165496489469,
      "loss": 2.8778,
      "theoretical_loss": 3.8986375817213466,
      "tokens_seen": 519715840
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.000425506519558676,
      "loss": 2.7645,
      "theoretical_loss": 3.8985849863898125,
      "tokens_seen": 519781376
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042549648946840524,
      "loss": 3.1105,
      "theoretical_loss": 3.8985323995458083,
      "tokens_seen": 519846912
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004254864593781344,
      "loss": 2.9125,
      "theoretical_loss": 3.8984798211868963,
      "tokens_seen": 519912448
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004254764292878636,
      "loss": 3.1081,
      "theoretical_loss": 3.8984272513106357,
      "tokens_seen": 519977984
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004254663991975928,
      "loss": 3.0043,
      "theoretical_loss": 3.89837468991459,
      "tokens_seen": 520043520
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.000425456369107322,
      "loss": 2.9944,
      "theoretical_loss": 3.898322136996324,
      "tokens_seen": 520109056
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042544633901705115,
      "loss": 2.826,
      "theoretical_loss": 3.898269592553401,
      "tokens_seen": 520174592
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004254363089267804,
      "loss": 2.5198,
      "theoretical_loss": 3.8982170565833867,
      "tokens_seen": 520240128
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004254262788365095,
      "loss": 2.9318,
      "theoretical_loss": 3.8981645290838483,
      "tokens_seen": 520305664
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042541624874623874,
      "loss": 2.7986,
      "theoretical_loss": 3.898112010052353,
      "tokens_seen": 520371200
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004254062186559679,
      "loss": 3.0719,
      "theoretical_loss": 3.898059499486469,
      "tokens_seen": 520436736
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004253961885656971,
      "loss": 2.9799,
      "theoretical_loss": 3.8980069973837677,
      "tokens_seen": 520502272
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004253861584754263,
      "loss": 2.9913,
      "theoretical_loss": 3.8979545037418175,
      "tokens_seen": 520567808
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042537612838515547,
      "loss": 2.8971,
      "theoretical_loss": 3.897902018558192,
      "tokens_seen": 520633344
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042536609829488465,
      "loss": 2.9909,
      "theoretical_loss": 3.8978495418304626,
      "tokens_seen": 520698880
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004253560682046139,
      "loss": 3.0641,
      "theoretical_loss": 3.8977970735562035,
      "tokens_seen": 520764416
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.000425346038114343,
      "loss": 3.1792,
      "theoretical_loss": 3.8977446137329887,
      "tokens_seen": 520829952
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042533600802407225,
      "loss": 3.1222,
      "theoretical_loss": 3.8976921623583944,
      "tokens_seen": 520895488
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004253259779338014,
      "loss": 2.9062,
      "theoretical_loss": 3.8976397194299977,
      "tokens_seen": 520961024
    },
    {
      "epoch": 1.07,
      "objective/train/docs_used": 607442,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.450615167617798,
      "objective/train/theoretical_loss": 3.8976003927750593,
      "objective/train/tokens_used": 541470176,
      "theoretical_loss": 3.8976003927750593,
      "tokens_seen": 521010176
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004253159478435306,
      "loss": 2.8771,
      "theoretical_loss": 3.8975872849453754,
      "tokens_seen": 521026560
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004253059177532598,
      "loss": 2.9847,
      "theoretical_loss": 3.897534858902107,
      "tokens_seen": 521092096
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042529588766298897,
      "loss": 2.7599,
      "theoretical_loss": 3.897482441297771,
      "tokens_seen": 521157632
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042528585757271815,
      "loss": 2.8196,
      "theoretical_loss": 3.8974300321299484,
      "tokens_seen": 521223168
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004252758274824474,
      "loss": 2.7218,
      "theoretical_loss": 3.897377631396222,
      "tokens_seen": 521288704
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004252657973921765,
      "loss": 2.7832,
      "theoretical_loss": 3.8973252390941733,
      "tokens_seen": 521354240
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042525576730190575,
      "loss": 3.1878,
      "theoretical_loss": 3.8972728552213862,
      "tokens_seen": 521419776
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004252457372116349,
      "loss": 2.9531,
      "theoretical_loss": 3.897220479775446,
      "tokens_seen": 521485312
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004252357071213641,
      "loss": 2.8667,
      "theoretical_loss": 3.8971681127539366,
      "tokens_seen": 521550848
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004252256770310933,
      "loss": 2.84,
      "theoretical_loss": 3.8971157541544468,
      "tokens_seen": 521616384
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004252156469408225,
      "loss": 2.7896,
      "theoretical_loss": 3.8970634039745624,
      "tokens_seen": 521681920
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042520561685055165,
      "loss": 3.1027,
      "theoretical_loss": 3.897011062211873,
      "tokens_seen": 521747456
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042519558676028084,
      "loss": 2.7281,
      "theoretical_loss": 3.896958728863968,
      "tokens_seen": 521812992
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042518555667001,
      "loss": 2.9021,
      "theoretical_loss": 3.8969064039284387,
      "tokens_seen": 521878528
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042517552657973925,
      "loss": 2.8982,
      "theoretical_loss": 3.8968540874028754,
      "tokens_seen": 521944064
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004251654964894684,
      "loss": 3.0018,
      "theoretical_loss": 3.8968017792848713,
      "tokens_seen": 522009600
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004251554663991976,
      "loss": 3.1168,
      "theoretical_loss": 3.8967494795720197,
      "tokens_seen": 522075136
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004251454363089268,
      "loss": 2.9336,
      "theoretical_loss": 3.896697188261916,
      "tokens_seen": 522140672
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.000425135406218656,
      "loss": 2.8327,
      "theoretical_loss": 3.8966449053521552,
      "tokens_seen": 522206208
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042512537612838516,
      "loss": 2.992,
      "theoretical_loss": 3.896592630840334,
      "tokens_seen": 522271744
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042511534603811434,
      "loss": 2.643,
      "theoretical_loss": 3.896540364724049,
      "tokens_seen": 522337280
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004251053159478435,
      "loss": 2.9115,
      "theoretical_loss": 3.8964881070009003,
      "tokens_seen": 522402816
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042509528585757276,
      "loss": 2.8193,
      "theoretical_loss": 3.896435857668486,
      "tokens_seen": 522468352
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004250852557673019,
      "loss": 2.856,
      "theoretical_loss": 3.8963836167244077,
      "tokens_seen": 522533888
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004250752256770311,
      "loss": 2.9309,
      "theoretical_loss": 3.896331384166267,
      "tokens_seen": 522599424
    },
    {
      "epoch": 1.07,
      "objective/train/docs_used": 608749,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.052105665206909,
      "objective/train/theoretical_loss": 3.896292215249489,
      "objective/train/tokens_used": 543108576,
      "theoretical_loss": 3.896292215249489,
      "tokens_seen": 522648576
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042506519558676024,
      "loss": 2.8582,
      "theoretical_loss": 3.896279159991664,
      "tokens_seen": 522664960
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004250551654964895,
      "loss": 2.9196,
      "theoretical_loss": 3.8962269441982063,
      "tokens_seen": 522730496
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042504513540621866,
      "loss": 2.9007,
      "theoretical_loss": 3.8961747367834945,
      "tokens_seen": 522796032
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042503510531594784,
      "loss": 3.1226,
      "theoretical_loss": 3.896122537745136,
      "tokens_seen": 522861568
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.000425025075225677,
      "loss": 3.0625,
      "theoretical_loss": 3.896070347080737,
      "tokens_seen": 522927104
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004250150451354062,
      "loss": 2.948,
      "theoretical_loss": 3.896018164787905,
      "tokens_seen": 522992640
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004250050150451354,
      "loss": 2.7946,
      "theoretical_loss": 3.895965990864247,
      "tokens_seen": 523058176
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004249949849548646,
      "loss": 2.8184,
      "theoretical_loss": 3.895913825307374,
      "tokens_seen": 523123712
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042498495486459375,
      "loss": 2.6486,
      "theoretical_loss": 3.895861668114896,
      "tokens_seen": 523189248
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.000424974924774323,
      "loss": 2.7146,
      "theoretical_loss": 3.895809519284424,
      "tokens_seen": 523254784
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042496489468405216,
      "loss": 3.0432,
      "theoretical_loss": 3.8957573788135704,
      "tokens_seen": 523320320
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042495486459378135,
      "loss": 2.756,
      "theoretical_loss": 3.895705246699949,
      "tokens_seen": 523385856
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004249448345035105,
      "loss": 2.9122,
      "theoretical_loss": 3.895653122941173,
      "tokens_seen": 523451392
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004249348044132397,
      "loss": 3.0265,
      "theoretical_loss": 3.895601007534858,
      "tokens_seen": 523516928
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042492477432296894,
      "loss": 3.0803,
      "theoretical_loss": 3.8955489004786212,
      "tokens_seen": 523582464
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004249147442326981,
      "loss": 2.9926,
      "theoretical_loss": 3.8954968017700793,
      "tokens_seen": 523648000
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004249047141424273,
      "loss": 2.8496,
      "theoretical_loss": 3.89544471140685,
      "tokens_seen": 523713536
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004248946840521565,
      "loss": 2.8785,
      "theoretical_loss": 3.8953926293865524,
      "tokens_seen": 523779072
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042488465396188567,
      "loss": 3.0031,
      "theoretical_loss": 3.8953405557068077,
      "tokens_seen": 523844608
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042487462387161485,
      "loss": 2.9991,
      "theoretical_loss": 3.895288490365236,
      "tokens_seen": 523910144
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004248645937813441,
      "loss": 2.8258,
      "theoretical_loss": 3.8952364333594596,
      "tokens_seen": 523975680
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004248545636910732,
      "loss": 2.7998,
      "theoretical_loss": 3.8951843846871013,
      "tokens_seen": 524041216
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042484453360080245,
      "loss": 2.7548,
      "theoretical_loss": 3.895132344345787,
      "tokens_seen": 524106752
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004248345035105316,
      "loss": 3.0234,
      "theoretical_loss": 3.895080312333139,
      "tokens_seen": 524172288
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004248244734202608,
      "loss": 2.9753,
      "theoretical_loss": 3.8950282886467846,
      "tokens_seen": 524237824
    },
    {
      "epoch": 1.07,
      "objective/train/docs_used": 609143,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.270958662033081,
      "objective/train/theoretical_loss": 3.894989276344722,
      "objective/train/tokens_used": 544746976,
      "theoretical_loss": 3.894989276344722,
      "tokens_seen": 524286976
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042481444332999,
      "loss": 3.1384,
      "theoretical_loss": 3.8949762732843514,
      "tokens_seen": 524303360
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042480441323971917,
      "loss": 2.5421,
      "theoretical_loss": 3.8949242662434664,
      "tokens_seen": 524368896
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042479438314944835,
      "loss": 3.0288,
      "theoretical_loss": 3.894872267521759,
      "tokens_seen": 524434432
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004247843530591776,
      "loss": 3.0604,
      "theoretical_loss": 3.8948202771168585,
      "tokens_seen": 524499968
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004247743229689067,
      "loss": 2.8957,
      "theoretical_loss": 3.894768295026396,
      "tokens_seen": 524565504
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042476429287863595,
      "loss": 2.7608,
      "theoretical_loss": 3.894716321248004,
      "tokens_seen": 524631040
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004247542627883651,
      "loss": 2.9537,
      "theoretical_loss": 3.8946643557793146,
      "tokens_seen": 524696576
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004247442326980943,
      "loss": 2.9661,
      "theoretical_loss": 3.8946123986179613,
      "tokens_seen": 524762112
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004247342026078235,
      "loss": 2.9881,
      "theoretical_loss": 3.8945604497615802,
      "tokens_seen": 524827648
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004247241725175527,
      "loss": 2.8578,
      "theoretical_loss": 3.8945085092078053,
      "tokens_seen": 524893184
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042471414242728186,
      "loss": 2.9858,
      "theoretical_loss": 3.8944565769542745,
      "tokens_seen": 524958720
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042470411233701104,
      "loss": 2.8689,
      "theoretical_loss": 3.8944046529986247,
      "tokens_seen": 525024256
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004246940822467402,
      "loss": 2.9482,
      "theoretical_loss": 3.8943527373384947,
      "tokens_seen": 525089792
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042468405215646945,
      "loss": 3.0764,
      "theoretical_loss": 3.8943008299715247,
      "tokens_seen": 525155328
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004246740220661986,
      "loss": 3.077,
      "theoretical_loss": 3.8942489308953547,
      "tokens_seen": 525220864
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004246639919759278,
      "loss": 2.7081,
      "theoretical_loss": 3.894197040107626,
      "tokens_seen": 525286400
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.000424653961885657,
      "loss": 2.9005,
      "theoretical_loss": 3.8941451576059807,
      "tokens_seen": 525351936
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004246439317953862,
      "loss": 2.7815,
      "theoretical_loss": 3.8940932833880635,
      "tokens_seen": 525417472
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042463390170511536,
      "loss": 2.698,
      "theoretical_loss": 3.8940414174515183,
      "tokens_seen": 525483008
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042462387161484454,
      "loss": 2.7497,
      "theoretical_loss": 3.8939895597939898,
      "tokens_seen": 525548544
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004246138415245737,
      "loss": 2.8141,
      "theoretical_loss": 3.893937710413125,
      "tokens_seen": 525614080
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042460381143430296,
      "loss": 2.9974,
      "theoretical_loss": 3.893885869306571,
      "tokens_seen": 525679616
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004245937813440321,
      "loss": 2.6887,
      "theoretical_loss": 3.8938340364719766,
      "tokens_seen": 525745152
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004245837512537613,
      "loss": 2.852,
      "theoretical_loss": 3.89378221190699,
      "tokens_seen": 525810688
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042457372116349044,
      "loss": 2.9421,
      "theoretical_loss": 3.8937303956092615,
      "tokens_seen": 525876224
    },
    {
      "epoch": 1.07,
      "objective/train/docs_used": 610436,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9925901889801025,
      "objective/train/theoretical_loss": 3.893691538809941,
      "objective/train/tokens_used": 546385376,
      "theoretical_loss": 3.893691538809941,
      "tokens_seen": 525925376
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004245636910732197,
      "loss": 2.7874,
      "theoretical_loss": 3.893678587576443,
      "tokens_seen": 525941760
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042455366098294886,
      "loss": 2.9986,
      "theoretical_loss": 3.8936267878061868,
      "tokens_seen": 526007296
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042454363089267804,
      "loss": 2.7701,
      "theoretical_loss": 3.8935749962961443,
      "tokens_seen": 526072832
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004245336008024072,
      "loss": 2.8532,
      "theoretical_loss": 3.8935232130439714,
      "tokens_seen": 526138368
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004245235707121364,
      "loss": 2.8331,
      "theoretical_loss": 3.893471438047322,
      "tokens_seen": 526203904
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004245135406218656,
      "loss": 3.127,
      "theoretical_loss": 3.8934196713038522,
      "tokens_seen": 526269440
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004245035105315948,
      "loss": 3.0505,
      "theoretical_loss": 3.8933679128112186,
      "tokens_seen": 526334976
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042449348044132395,
      "loss": 3.023,
      "theoretical_loss": 3.89331616256708,
      "tokens_seen": 526400512
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004244834503510532,
      "loss": 3.0169,
      "theoretical_loss": 3.893264420569094,
      "tokens_seen": 526466048
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042447342026078236,
      "loss": 2.9502,
      "theoretical_loss": 3.893212686814921,
      "tokens_seen": 526531584
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042446339017051155,
      "loss": 2.9393,
      "theoretical_loss": 3.893160961302222,
      "tokens_seen": 526597120
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004244533600802407,
      "loss": 2.765,
      "theoretical_loss": 3.893109244028658,
      "tokens_seen": 526662656
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004244433299899699,
      "loss": 3.2206,
      "theoretical_loss": 3.8930575349918923,
      "tokens_seen": 526728192
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004244332998996991,
      "loss": 2.9727,
      "theoretical_loss": 3.893005834189588,
      "tokens_seen": 526793728
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004244232698094283,
      "loss": 3.0495,
      "theoretical_loss": 3.8929541416194096,
      "tokens_seen": 526859264
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042441323971915745,
      "loss": 3.11,
      "theoretical_loss": 3.8929024572790225,
      "tokens_seen": 526924800
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004244032096288867,
      "loss": 2.7881,
      "theoretical_loss": 3.8928507811660937,
      "tokens_seen": 526990336
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004243931795386158,
      "loss": 3.0553,
      "theoretical_loss": 3.8927991132782904,
      "tokens_seen": 527055872
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042438314944834505,
      "loss": 3.0207,
      "theoretical_loss": 3.8927474536132802,
      "tokens_seen": 527121408
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042437311935807423,
      "loss": 2.9144,
      "theoretical_loss": 3.8926958021687335,
      "tokens_seen": 527186944
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004243630892678034,
      "loss": 2.8823,
      "theoretical_loss": 3.8926441589423195,
      "tokens_seen": 527252480
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004243530591775326,
      "loss": 3.1453,
      "theoretical_loss": 3.89259252393171,
      "tokens_seen": 527318016
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004243430290872618,
      "loss": 2.7057,
      "theoretical_loss": 3.8925408971345776,
      "tokens_seen": 527383552
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042433299899699095,
      "loss": 2.645,
      "theoretical_loss": 3.8924892785485943,
      "tokens_seen": 527449088
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004243229689067202,
      "loss": 2.9985,
      "theoretical_loss": 3.8924376681714348,
      "tokens_seen": 527514624
    },
    {
      "epoch": 1.07,
      "objective/train/docs_used": 611010,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8362579345703125,
      "objective/train/theoretical_loss": 3.8923989657742073,
      "objective/train/tokens_used": 548023776,
      "theoretical_loss": 3.8923989657742073,
      "tokens_seen": 527563776
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004243129388164493,
      "loss": 2.9011,
      "theoretical_loss": 3.892386066000774,
      "tokens_seen": 527580160
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042430290872617855,
      "loss": 2.876,
      "theoretical_loss": 3.892334472034287,
      "tokens_seen": 527645696
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042429287863590773,
      "loss": 3.0023,
      "theoretical_loss": 3.8922828862696526,
      "tokens_seen": 527711232
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004242828485456369,
      "loss": 2.8032,
      "theoretical_loss": 3.8922313087045466,
      "tokens_seen": 527776768
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004242728184553661,
      "loss": 2.8308,
      "theoretical_loss": 3.892179739336649,
      "tokens_seen": 527842304
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004242627883650953,
      "loss": 2.8076,
      "theoretical_loss": 3.892128178163639,
      "tokens_seen": 527907840
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042425275827482446,
      "loss": 3.0146,
      "theoretical_loss": 3.892076625183197,
      "tokens_seen": 527973376
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004242427281845537,
      "loss": 2.9239,
      "theoretical_loss": 3.8920250803930054,
      "tokens_seen": 528038912
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004242326980942828,
      "loss": 2.7142,
      "theoretical_loss": 3.891973543790746,
      "tokens_seen": 528104448
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042422266800401206,
      "loss": 2.8914,
      "theoretical_loss": 3.891922015374103,
      "tokens_seen": 528169984
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004242126379137412,
      "loss": 2.6705,
      "theoretical_loss": 3.8918704951407603,
      "tokens_seen": 528235520
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004242026078234704,
      "loss": 2.8523,
      "theoretical_loss": 3.8918189830884033,
      "tokens_seen": 528301056
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004241925777331996,
      "loss": 2.9579,
      "theoretical_loss": 3.8917674792147183,
      "tokens_seen": 528366592
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004241825476429288,
      "loss": 2.8322,
      "theoretical_loss": 3.8917159835173933,
      "tokens_seen": 528432128
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.000424172517552658,
      "loss": 2.8149,
      "theoretical_loss": 3.891664495994115,
      "tokens_seen": 528497664
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004241624874623872,
      "loss": 3.0571,
      "theoretical_loss": 3.891613016642574,
      "tokens_seen": 528563200
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004241524573721164,
      "loss": 3.1115,
      "theoretical_loss": 3.8915615454604593,
      "tokens_seen": 528628736
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042414242728184556,
      "loss": 2.8828,
      "theoretical_loss": 3.8915100824454627,
      "tokens_seen": 528694272
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042413239719157474,
      "loss": 3.0419,
      "theoretical_loss": 3.8914586275952763,
      "tokens_seen": 528759808
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004241223671013039,
      "loss": 2.9529,
      "theoretical_loss": 3.8914071809075916,
      "tokens_seen": 528825344
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042411233701103316,
      "loss": 2.8425,
      "theoretical_loss": 3.891355742380104,
      "tokens_seen": 528890880
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004241023069207623,
      "loss": 2.7841,
      "theoretical_loss": 3.891304312010507,
      "tokens_seen": 528956416
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004240922768304915,
      "loss": 2.9775,
      "theoretical_loss": 3.8912528897964975,
      "tokens_seen": 529021952
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042408224674022065,
      "loss": 2.9298,
      "theoretical_loss": 3.891201475735772,
      "tokens_seen": 529087488
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004240722166499499,
      "loss": 2.8611,
      "theoretical_loss": 3.891150069826027,
      "tokens_seen": 529153024
    },
    {
      "epoch": 1.07,
      "objective/train/docs_used": 612097,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.09621000289917,
      "objective/train/theoretical_loss": 3.8911115207414153,
      "objective/train/tokens_used": 549662176,
      "theoretical_loss": 3.8911115207414153,
      "tokens_seen": 529202176
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042406218655967906,
      "loss": 3.0195,
      "theoretical_loss": 3.8910986720649614,
      "tokens_seen": 529218560
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042405215646940824,
      "loss": 3.0257,
      "theoretical_loss": 3.8910472824502755,
      "tokens_seen": 529284096
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004240421263791374,
      "loss": 3.2495,
      "theoretical_loss": 3.890995900979669,
      "tokens_seen": 529349632
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004240320962888666,
      "loss": 2.8085,
      "theoretical_loss": 3.890944527650843,
      "tokens_seen": 529415168
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004240220661985958,
      "loss": 2.9752,
      "theoretical_loss": 3.8908931624615004,
      "tokens_seen": 529480704
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.000424012036108325,
      "loss": 2.9726,
      "theoretical_loss": 3.8908418054093437,
      "tokens_seen": 529546240
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042400200601805415,
      "loss": 2.9715,
      "theoretical_loss": 3.890790456492078,
      "tokens_seen": 529611776
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004239919759277834,
      "loss": 2.7099,
      "theoretical_loss": 3.890739115707407,
      "tokens_seen": 529677312
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042398194583751256,
      "loss": 2.9256,
      "theoretical_loss": 3.8906877830530373,
      "tokens_seen": 529742848
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042397191574724175,
      "loss": 2.8714,
      "theoretical_loss": 3.8906364585266764,
      "tokens_seen": 529808384
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004239618856569709,
      "loss": 2.8072,
      "theoretical_loss": 3.8905851421260316,
      "tokens_seen": 529873920
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004239518555667001,
      "loss": 2.8951,
      "theoretical_loss": 3.890533833848811,
      "tokens_seen": 529939456
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004239418254764293,
      "loss": 2.8725,
      "theoretical_loss": 3.8904825336927256,
      "tokens_seen": 530004992
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004239317953861585,
      "loss": 3.0388,
      "theoretical_loss": 3.890431241655485,
      "tokens_seen": 530070528
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042392176529588765,
      "loss": 2.9665,
      "theoretical_loss": 3.890379957734801,
      "tokens_seen": 530136064
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004239117352056169,
      "loss": 2.9384,
      "theoretical_loss": 3.890328681928387,
      "tokens_seen": 530201600
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.000423901705115346,
      "loss": 3.113,
      "theoretical_loss": 3.890277414233955,
      "tokens_seen": 530267136
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042389167502507525,
      "loss": 2.801,
      "theoretical_loss": 3.89022615464922,
      "tokens_seen": 530332672
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042388164493480443,
      "loss": 2.8593,
      "theoretical_loss": 3.8901749031718973,
      "tokens_seen": 530398208
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004238716148445336,
      "loss": 3.1283,
      "theoretical_loss": 3.890123659799703,
      "tokens_seen": 530463744
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004238615847542628,
      "loss": 2.9641,
      "theoretical_loss": 3.890072424530354,
      "tokens_seen": 530529280
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.000423851554663992,
      "loss": 3.1211,
      "theoretical_loss": 3.890021197361569,
      "tokens_seen": 530594816
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042384152457372115,
      "loss": 3.0983,
      "theoretical_loss": 3.889969978291066,
      "tokens_seen": 530660352
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004238314944834504,
      "loss": 3.0237,
      "theoretical_loss": 3.8899187673165656,
      "tokens_seen": 530725888
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004238214643931795,
      "loss": 2.7202,
      "theoretical_loss": 3.8898675644357885,
      "tokens_seen": 530791424
    },
    {
      "epoch": 1.07,
      "objective/train/docs_used": 612816,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.779944658279419,
      "objective/train/theoretical_loss": 3.889829167585341,
      "objective/train/tokens_used": 551300576,
      "theoretical_loss": 3.889829167585341,
      "tokens_seen": 530840576
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042381143430290875,
      "loss": 2.7002,
      "theoretical_loss": 3.8898163696464563,
      "tokens_seen": 530856960
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042380140421263793,
      "loss": 2.8166,
      "theoretical_loss": 3.889765182946292,
      "tokens_seen": 530922496
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004237913741223671,
      "loss": 3.0376,
      "theoretical_loss": 3.8897140043330185,
      "tokens_seen": 530988032
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004237813440320963,
      "loss": 3.2238,
      "theoretical_loss": 3.889662833804361,
      "tokens_seen": 531053568
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004237713139418255,
      "loss": 2.808,
      "theoretical_loss": 3.889611671358044,
      "tokens_seen": 531119104
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042376128385155466,
      "loss": 3.0963,
      "theoretical_loss": 3.8895605169917955,
      "tokens_seen": 531184640
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004237512537612839,
      "loss": 2.9041,
      "theoretical_loss": 3.8895093707033404,
      "tokens_seen": 531250176
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.000423741223671013,
      "loss": 3.0183,
      "theoretical_loss": 3.8894582324904086,
      "tokens_seen": 531315712
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042373119358074226,
      "loss": 2.8208,
      "theoretical_loss": 3.889407102350729,
      "tokens_seen": 531381248
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004237211634904714,
      "loss": 2.8103,
      "theoretical_loss": 3.8893559802820317,
      "tokens_seen": 531446784
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004237111334002006,
      "loss": 2.868,
      "theoretical_loss": 3.8893048662820466,
      "tokens_seen": 531512320
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004237011033099298,
      "loss": 2.7113,
      "theoretical_loss": 3.8892537603485065,
      "tokens_seen": 531577856
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.000423691073219659,
      "loss": 2.7521,
      "theoretical_loss": 3.889202662479144,
      "tokens_seen": 531643392
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042368104312938816,
      "loss": 2.9089,
      "theoretical_loss": 3.8891515726716923,
      "tokens_seen": 531708928
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004236710130391174,
      "loss": 2.9575,
      "theoretical_loss": 3.8891004909238873,
      "tokens_seen": 531774464
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004236609829488465,
      "loss": 2.8788,
      "theoretical_loss": 3.889049417233463,
      "tokens_seen": 531840000
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042365095285857576,
      "loss": 2.7697,
      "theoretical_loss": 3.8889983515981568,
      "tokens_seen": 531905536
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004236409227683049,
      "loss": 3.0722,
      "theoretical_loss": 3.888947294015705,
      "tokens_seen": 531971072
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004236308926780341,
      "loss": 3.1524,
      "theoretical_loss": 3.8888962444838473,
      "tokens_seen": 532036608
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004236208625877633,
      "loss": 3.0272,
      "theoretical_loss": 3.888845203000322,
      "tokens_seen": 532102144
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004236108324974925,
      "loss": 2.8684,
      "theoretical_loss": 3.8887941695628694,
      "tokens_seen": 532167680
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042360080240722166,
      "loss": 2.9734,
      "theoretical_loss": 3.88874314416923,
      "tokens_seen": 532233216
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042359077231695085,
      "loss": 2.8621,
      "theoretical_loss": 3.888692126817147,
      "tokens_seen": 532298752
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042358074222668,
      "loss": 2.6702,
      "theoretical_loss": 3.8886411175043616,
      "tokens_seen": 532364288
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042357071213640926,
      "loss": 2.8799,
      "theoretical_loss": 3.8885901162286185,
      "tokens_seen": 532429824
    },
    {
      "epoch": 1.07,
      "objective/train/docs_used": 614210,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.1336987018585205,
      "objective/train/theoretical_loss": 3.888551870544763,
      "objective/train/tokens_used": 552938976,
      "theoretical_loss": 3.888551870544763,
      "tokens_seen": 532478976
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004235606820461384,
      "loss": 3.1439,
      "theoretical_loss": 3.8885391229876625,
      "tokens_seen": 532495360
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004235506519558676,
      "loss": 2.9976,
      "theoretical_loss": 3.8884881377792384,
      "tokens_seen": 532560896
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042354062186559675,
      "loss": 3.0269,
      "theoretical_loss": 3.888437160601093,
      "tokens_seen": 532626432
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.000423530591775326,
      "loss": 2.83,
      "theoretical_loss": 3.8883861914509743,
      "tokens_seen": 532691968
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042352056168505517,
      "loss": 2.9997,
      "theoretical_loss": 3.88833523032663,
      "tokens_seen": 532757504
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042351053159478435,
      "loss": 3.0225,
      "theoretical_loss": 3.888284277225809,
      "tokens_seen": 532823040
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042350050150451353,
      "loss": 2.8851,
      "theoretical_loss": 3.888233332146262,
      "tokens_seen": 532888576
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042349047141424276,
      "loss": 2.8663,
      "theoretical_loss": 3.8881823950857397,
      "tokens_seen": 532954112
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004234804413239719,
      "loss": 2.9248,
      "theoretical_loss": 3.8881314660419934,
      "tokens_seen": 533019648
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042347041123370113,
      "loss": 2.9781,
      "theoretical_loss": 3.8880805450127776,
      "tokens_seen": 533085184
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042346038114343025,
      "loss": 2.9781,
      "theoretical_loss": 3.8880296319958445,
      "tokens_seen": 533150720
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004234503510531595,
      "loss": 2.7219,
      "theoretical_loss": 3.8879787269889494,
      "tokens_seen": 533216256
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042344032096288867,
      "loss": 3.0108,
      "theoretical_loss": 3.887927829989848,
      "tokens_seen": 533281792
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042343029087261785,
      "loss": 2.8635,
      "theoretical_loss": 3.887876940996296,
      "tokens_seen": 533347328
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004234202607823471,
      "loss": 2.8422,
      "theoretical_loss": 3.8878260600060512,
      "tokens_seen": 533412864
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004234102306920762,
      "loss": 3.0575,
      "theoretical_loss": 3.8877751870168717,
      "tokens_seen": 533478400
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042340020060180545,
      "loss": 2.9074,
      "theoretical_loss": 3.8877243220265174,
      "tokens_seen": 533543936
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042339017051153463,
      "loss": 3.0241,
      "theoretical_loss": 3.887673465032748,
      "tokens_seen": 533609472
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004233801404212638,
      "loss": 2.865,
      "theoretical_loss": 3.8876226160333234,
      "tokens_seen": 533675008
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.000423370110330993,
      "loss": 3.0424,
      "theoretical_loss": 3.8875717750260064,
      "tokens_seen": 533740544
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004233600802407222,
      "loss": 3.1135,
      "theoretical_loss": 3.8875209420085595,
      "tokens_seen": 533806080
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042335005015045135,
      "loss": 3.0429,
      "theoretical_loss": 3.887470116978747,
      "tokens_seen": 533871616
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004233400200601806,
      "loss": 3.0547,
      "theoretical_loss": 3.887419299934333,
      "tokens_seen": 533937152
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004233299899699097,
      "loss": 3.0215,
      "theoretical_loss": 3.887368490873083,
      "tokens_seen": 534002688
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042331995987963895,
      "loss": 3.0887,
      "theoretical_loss": 3.887317689792763,
      "tokens_seen": 534068224
    },
    {
      "epoch": 1.07,
      "objective/train/docs_used": 614920,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9131345748901367,
      "objective/train/theoretical_loss": 3.8872795942186653,
      "objective/train/tokens_used": 554577376,
      "theoretical_loss": 3.8872795942186653,
      "tokens_seen": 534117376
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042330992978936813,
      "loss": 2.9236,
      "theoretical_loss": 3.887266896691141,
      "tokens_seen": 534133760
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004232998996990973,
      "loss": 2.948,
      "theoretical_loss": 3.8872161115659845,
      "tokens_seen": 534199296
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004232898696088265,
      "loss": 2.7217,
      "theoretical_loss": 3.8871653344150623,
      "tokens_seen": 534264832
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004232798395185557,
      "loss": 2.8982,
      "theoretical_loss": 3.8871145652361454,
      "tokens_seen": 534330368
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042326980942828486,
      "loss": 3.1815,
      "theoretical_loss": 3.887063804027004,
      "tokens_seen": 534395904
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004232597793380141,
      "loss": 2.967,
      "theoretical_loss": 3.88701305078541,
      "tokens_seen": 534461440
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004232497492477432,
      "loss": 3.0584,
      "theoretical_loss": 3.886962305509136,
      "tokens_seen": 534526976
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042323971915747246,
      "loss": 3.0766,
      "theoretical_loss": 3.8869115681959556,
      "tokens_seen": 534592512
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004232296890672016,
      "loss": 3.0689,
      "theoretical_loss": 3.886860838843643,
      "tokens_seen": 534658048
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004232196589769308,
      "loss": 2.9005,
      "theoretical_loss": 3.886810117449974,
      "tokens_seen": 534723584
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042320962888666,
      "loss": 2.9064,
      "theoretical_loss": 3.8867594040127242,
      "tokens_seen": 534789120
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004231995987963892,
      "loss": 2.9348,
      "theoretical_loss": 3.8867086985296715,
      "tokens_seen": 534854656
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042318956870611836,
      "loss": 3.1647,
      "theoretical_loss": 3.8866580009985934,
      "tokens_seen": 534920192
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004231795386158476,
      "loss": 2.8449,
      "theoretical_loss": 3.8866073114172686,
      "tokens_seen": 534985728
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004231695085255767,
      "loss": 2.9154,
      "theoretical_loss": 3.8865566297834775,
      "tokens_seen": 535051264
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042315947843530596,
      "loss": 3.0347,
      "theoretical_loss": 3.886505956095,
      "tokens_seen": 535116800
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004231494483450351,
      "loss": 2.9126,
      "theoretical_loss": 3.886455290349619,
      "tokens_seen": 535182336
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004231394182547643,
      "loss": 3.0763,
      "theoretical_loss": 3.8864046325451156,
      "tokens_seen": 535247872
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004231293881644935,
      "loss": 2.9425,
      "theoretical_loss": 3.8863539826792737,
      "tokens_seen": 535313408
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004231193580742227,
      "loss": 2.9192,
      "theoretical_loss": 3.886303340749878,
      "tokens_seen": 535378944
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042310932798395186,
      "loss": 3.0836,
      "theoretical_loss": 3.8862527067547123,
      "tokens_seen": 535444480
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042309929789368105,
      "loss": 2.7298,
      "theoretical_loss": 3.8862020806915645,
      "tokens_seen": 535510016
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004230892678034102,
      "loss": 3.1905,
      "theoretical_loss": 3.88615146255822,
      "tokens_seen": 535575552
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042307923771313946,
      "loss": 2.9487,
      "theoretical_loss": 3.8861008523524676,
      "tokens_seen": 535641088
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004230692076228686,
      "loss": 2.7571,
      "theoretical_loss": 3.8860502500720946,
      "tokens_seen": 535706624
    },
    {
      "epoch": 1.07,
      "objective/train/docs_used": 616369,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9449539184570312,
      "objective/train/theoretical_loss": 3.886012303561517,
      "objective/train/tokens_used": 556215776,
      "theoretical_loss": 3.886012303561517,
      "tokens_seen": 535755776
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004230591775325978,
      "loss": 2.9742,
      "theoretical_loss": 3.885999655714893,
      "tokens_seen": 535772160
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042304914744232695,
      "loss": 3.0579,
      "theoretical_loss": 3.8859490692786505,
      "tokens_seen": 535837696
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004230391173520562,
      "loss": 3.0758,
      "theoretical_loss": 3.8858984907611607,
      "tokens_seen": 535903232
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042302908726178537,
      "loss": 2.8493,
      "theoretical_loss": 3.8858479201602147,
      "tokens_seen": 535968768
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042301905717151455,
      "loss": 3.0233,
      "theoretical_loss": 3.8857973574736056,
      "tokens_seen": 536034304
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042300902708124373,
      "loss": 2.9448,
      "theoretical_loss": 3.8857468026991278,
      "tokens_seen": 536099840
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042299899699097297,
      "loss": 2.9356,
      "theoretical_loss": 3.885696255834576,
      "tokens_seen": 536165376
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004229889669007021,
      "loss": 2.8509,
      "theoretical_loss": 3.885645716877746,
      "tokens_seen": 536230912
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042297893681043133,
      "loss": 2.9117,
      "theoretical_loss": 3.8855951858264346,
      "tokens_seen": 536296448
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042296890672016045,
      "loss": 3.0141,
      "theoretical_loss": 3.8855446626784396,
      "tokens_seen": 536361984
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004229588766298897,
      "loss": 2.8382,
      "theoretical_loss": 3.8854941474315585,
      "tokens_seen": 536427520
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042294884653961887,
      "loss": 2.84,
      "theoretical_loss": 3.885443640083592,
      "tokens_seen": 536493056
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042293881644934805,
      "loss": 3.0141,
      "theoretical_loss": 3.8853931406323383,
      "tokens_seen": 536558592
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042292878635907723,
      "loss": 3.0608,
      "theoretical_loss": 3.8853426490756005,
      "tokens_seen": 536624128
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004229187562688064,
      "loss": 3.0681,
      "theoretical_loss": 3.8852921654111796,
      "tokens_seen": 536689664
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004229087261785356,
      "loss": 3.1048,
      "theoretical_loss": 3.8852416896368784,
      "tokens_seen": 536755200
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042289869608826483,
      "loss": 2.8965,
      "theoretical_loss": 3.8851912217505014,
      "tokens_seen": 536820736
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042288866599799396,
      "loss": 3.0463,
      "theoretical_loss": 3.885140761749852,
      "tokens_seen": 536886272
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004228786359077232,
      "loss": 2.8842,
      "theoretical_loss": 3.8850903096327363,
      "tokens_seen": 536951808
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004228686058174523,
      "loss": 3.0815,
      "theoretical_loss": 3.8850398653969607,
      "tokens_seen": 537017344
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042285857572718155,
      "loss": 3.1825,
      "theoretical_loss": 3.884989429040332,
      "tokens_seen": 537082880
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042284854563691074,
      "loss": 2.6999,
      "theoretical_loss": 3.8849390005606597,
      "tokens_seen": 537148416
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004228385155466399,
      "loss": 3.094,
      "theoretical_loss": 3.884888579955751,
      "tokens_seen": 537213952
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004228284854563691,
      "loss": 2.8036,
      "theoretical_loss": 3.8848381672234167,
      "tokens_seen": 537279488
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042281845536609833,
      "loss": 3.1333,
      "theoretical_loss": 3.8847877623614675,
      "tokens_seen": 537345024
    },
    {
      "epoch": 1.07,
      "objective/train/docs_used": 617122,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5995001792907715,
      "objective/train/theoretical_loss": 3.8847499638786287,
      "objective/train/tokens_used": 557854176,
      "theoretical_loss": 3.8847499638786287,
      "tokens_seen": 537394176
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042280842527582746,
      "loss": 2.8701,
      "theoretical_loss": 3.884737365367714,
      "tokens_seen": 537410560
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004227983951855567,
      "loss": 2.9849,
      "theoretical_loss": 3.8846869762399705,
      "tokens_seen": 537476096
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004227883650952858,
      "loss": 3.0028,
      "theoretical_loss": 3.8846365949760493,
      "tokens_seen": 537541632
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042277833500501506,
      "loss": 2.748,
      "theoretical_loss": 3.8845862215737643,
      "tokens_seen": 537607168
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042276830491474424,
      "loss": 2.9641,
      "theoretical_loss": 3.884535856030932,
      "tokens_seen": 537672704
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004227582748244734,
      "loss": 2.9048,
      "theoretical_loss": 3.8844854983453665,
      "tokens_seen": 537738240
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004227482447342026,
      "loss": 2.9251,
      "theoretical_loss": 3.884435148514886,
      "tokens_seen": 537803776
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004227382146439318,
      "loss": 2.9232,
      "theoretical_loss": 3.884384806537308,
      "tokens_seen": 537869312
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042272818455366096,
      "loss": 3.2003,
      "theoretical_loss": 3.8843344724104507,
      "tokens_seen": 537934848
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004227181544633902,
      "loss": 3.0039,
      "theoretical_loss": 3.884284146132134,
      "tokens_seen": 538000384
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004227081243731193,
      "loss": 2.8426,
      "theoretical_loss": 3.8842338277001778,
      "tokens_seen": 538065920
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042269809428284856,
      "loss": 2.8665,
      "theoretical_loss": 3.884183517112404,
      "tokens_seen": 538131456
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004226880641925777,
      "loss": 2.8744,
      "theoretical_loss": 3.884133214366634,
      "tokens_seen": 538196992
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004226780341023069,
      "loss": 3.132,
      "theoretical_loss": 3.884082919460692,
      "tokens_seen": 538262528
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042266800401203616,
      "loss": 2.7897,
      "theoretical_loss": 3.8840326323923993,
      "tokens_seen": 538328064
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004226579739217653,
      "loss": 2.889,
      "theoretical_loss": 3.883982353159583,
      "tokens_seen": 538393600
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004226479438314945,
      "loss": 2.858,
      "theoretical_loss": 3.8839320817600678,
      "tokens_seen": 538459136
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004226379137412237,
      "loss": 2.9575,
      "theoretical_loss": 3.8838818181916803,
      "tokens_seen": 538524672
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004226278836509529,
      "loss": 3.1264,
      "theoretical_loss": 3.8838315624522473,
      "tokens_seen": 538590208
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042261785356068206,
      "loss": 2.9071,
      "theoretical_loss": 3.883781314539598,
      "tokens_seen": 538655744
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042260782347041125,
      "loss": 2.7484,
      "theoretical_loss": 3.8837310744515605,
      "tokens_seen": 538721280
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004225977933801404,
      "loss": 2.934,
      "theoretical_loss": 3.8836808421859645,
      "tokens_seen": 538786816
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042258776328986966,
      "loss": 3.0486,
      "theoretical_loss": 3.883630617740642,
      "tokens_seen": 538852352
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004225777331995988,
      "loss": 2.9116,
      "theoretical_loss": 3.883580401113423,
      "tokens_seen": 538917888
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.000422567703109328,
      "loss": 2.8901,
      "theoretical_loss": 3.8835301923021417,
      "tokens_seen": 538983424
    },
    {
      "epoch": 1.07,
      "objective/train/docs_used": 618589,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.2161307334899902,
      "objective/train/theoretical_loss": 3.8834925408215852,
      "objective/train/tokens_used": 559492576,
      "theoretical_loss": 3.8834925408215852,
      "tokens_seen": 539032576
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042255767301905715,
      "loss": 2.9765,
      "theoretical_loss": 3.8834799913046303,
      "tokens_seen": 539048960
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004225476429287864,
      "loss": 3.1059,
      "theoretical_loss": 3.883429798118723,
      "tokens_seen": 539114496
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042253761283851557,
      "loss": 2.8609,
      "theoretical_loss": 3.8833796127422557,
      "tokens_seen": 539180032
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042252758274824475,
      "loss": 3.0228,
      "theoretical_loss": 3.8833294351730636,
      "tokens_seen": 539245568
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042251755265797393,
      "loss": 2.9747,
      "theoretical_loss": 3.8832792654089845,
      "tokens_seen": 539311104
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042250752256770317,
      "loss": 3.0639,
      "theoretical_loss": 3.8832291034478548,
      "tokens_seen": 539376640
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004224974924774323,
      "loss": 3.0354,
      "theoretical_loss": 3.8831789492875135,
      "tokens_seen": 539442176
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042248746238716153,
      "loss": 2.953,
      "theoretical_loss": 3.8831288029258006,
      "tokens_seen": 539507712
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042247743229689065,
      "loss": 2.6457,
      "theoretical_loss": 3.8830786643605553,
      "tokens_seen": 539573248
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004224674022066199,
      "loss": 3.2025,
      "theoretical_loss": 3.88302853358962,
      "tokens_seen": 539638784
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042245737211634907,
      "loss": 2.9132,
      "theoretical_loss": 3.8829784106108356,
      "tokens_seen": 539704320
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042244734202607825,
      "loss": 2.6975,
      "theoretical_loss": 3.882928295422045,
      "tokens_seen": 539769856
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042243731193580743,
      "loss": 2.6874,
      "theoretical_loss": 3.8828781880210927,
      "tokens_seen": 539835392
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004224272818455366,
      "loss": 2.9888,
      "theoretical_loss": 3.882828088405823,
      "tokens_seen": 539900928
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004224172517552658,
      "loss": 2.7301,
      "theoretical_loss": 3.8827779965740805,
      "tokens_seen": 539966464
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042240722166499503,
      "loss": 2.7876,
      "theoretical_loss": 3.882727912523712,
      "tokens_seen": 540032000
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042239719157472416,
      "loss": 2.7523,
      "theoretical_loss": 3.882677836252565,
      "tokens_seen": 540097536
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004223871614844534,
      "loss": 3.0484,
      "theoretical_loss": 3.8826277677584873,
      "tokens_seen": 540163072
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004223771313941825,
      "loss": 3.0711,
      "theoretical_loss": 3.882577707039328,
      "tokens_seen": 540228608
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042236710130391176,
      "loss": 2.949,
      "theoretical_loss": 3.8825276540929354,
      "tokens_seen": 540294144
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042235707121364094,
      "loss": 2.7796,
      "theoretical_loss": 3.8824776089171618,
      "tokens_seen": 540359680
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004223470411233701,
      "loss": 2.7581,
      "theoretical_loss": 3.8824275715098575,
      "tokens_seen": 540425216
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004223370110330993,
      "loss": 2.9913,
      "theoretical_loss": 3.882377541868875,
      "tokens_seen": 540490752
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042232698094282853,
      "loss": 3.1344,
      "theoretical_loss": 3.882327519992068,
      "tokens_seen": 540556288
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042231695085255766,
      "loss": 2.8887,
      "theoretical_loss": 3.88227750587729,
      "tokens_seen": 540621824
    },
    {
      "epoch": 1.07,
      "objective/train/docs_used": 619158,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8451883792877197,
      "objective/train/theoretical_loss": 3.8822400003837476,
      "objective/train/tokens_used": 561130976,
      "theoretical_loss": 3.8822400003837476,
      "tokens_seen": 540670976
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004223069207622869,
      "loss": 2.9941,
      "theoretical_loss": 3.8822274995223958,
      "tokens_seen": 540687360
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.000422296890672016,
      "loss": 2.7622,
      "theoretical_loss": 3.8821775009252413,
      "tokens_seen": 540752896
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042228686058174526,
      "loss": 2.8625,
      "theoretical_loss": 3.8821275100836825,
      "tokens_seen": 540818432
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042227683049147444,
      "loss": 2.6665,
      "theoretical_loss": 3.8820775269955776,
      "tokens_seen": 540883968
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004222668004012036,
      "loss": 2.8965,
      "theoretical_loss": 3.8820275516587843,
      "tokens_seen": 540949504
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004222567703109328,
      "loss": 2.7812,
      "theoretical_loss": 3.8819775840711612,
      "tokens_seen": 541015040
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.000422246740220662,
      "loss": 2.7666,
      "theoretical_loss": 3.8819276242305696,
      "tokens_seen": 541080576
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042223671013039116,
      "loss": 2.5857,
      "theoretical_loss": 3.881877672134869,
      "tokens_seen": 541146112
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004222266800401204,
      "loss": 2.9337,
      "theoretical_loss": 3.8818277277819213,
      "tokens_seen": 541211648
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004222166499498495,
      "loss": 2.8628,
      "theoretical_loss": 3.88177779116959,
      "tokens_seen": 541277184
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042220661985957876,
      "loss": 2.5178,
      "theoretical_loss": 3.881727862295737,
      "tokens_seen": 541342720
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004221965897693079,
      "loss": 2.8533,
      "theoretical_loss": 3.8816779411582276,
      "tokens_seen": 541408256
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004221865596790371,
      "loss": 2.8706,
      "theoretical_loss": 3.881628027754926,
      "tokens_seen": 541473792
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004221765295887663,
      "loss": 2.9555,
      "theoretical_loss": 3.8815781220836985,
      "tokens_seen": 541539328
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004221664994984955,
      "loss": 2.957,
      "theoretical_loss": 3.8815282241424116,
      "tokens_seen": 541604864
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042215646940822467,
      "loss": 2.5968,
      "theoretical_loss": 3.881478333928933,
      "tokens_seen": 541670400
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004221464393179539,
      "loss": 2.8318,
      "theoretical_loss": 3.8814284514411312,
      "tokens_seen": 541735936
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042213640922768303,
      "loss": 2.748,
      "theoretical_loss": 3.881378576676876,
      "tokens_seen": 541801472
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042212637913741226,
      "loss": 2.952,
      "theoretical_loss": 3.8813287096340363,
      "tokens_seen": 541867008
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004221163490471414,
      "loss": 2.8788,
      "theoretical_loss": 3.881278850310484,
      "tokens_seen": 541932544
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004221063189568706,
      "loss": 3.0883,
      "theoretical_loss": 3.88122899870409,
      "tokens_seen": 541998080
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004220962888665998,
      "loss": 2.8509,
      "theoretical_loss": 3.8811791548127275,
      "tokens_seen": 542063616
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.000422086258776329,
      "loss": 2.943,
      "theoretical_loss": 3.8811293186342706,
      "tokens_seen": 542129152
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042207622868605817,
      "loss": 2.9274,
      "theoretical_loss": 3.881079490166593,
      "tokens_seen": 542194688
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042206619859578735,
      "loss": 3.0352,
      "theoretical_loss": 3.881029669407569,
      "tokens_seen": 542260224
    },
    {
      "epoch": 1.07,
      "objective/train/docs_used": 620361,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.204775810241699,
      "objective/train/theoretical_loss": 3.8809923088958285,
      "objective/train/tokens_used": 562769376,
      "theoretical_loss": 3.8809923088958285,
      "tokens_seen": 542309376
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042205616850551653,
      "loss": 3.0893,
      "theoretical_loss": 3.8809798563550766,
      "tokens_seen": 542325760
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042204613841524577,
      "loss": 3.1817,
      "theoretical_loss": 3.880930051006991,
      "tokens_seen": 542391296
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004220361083249749,
      "loss": 2.8584,
      "theoretical_loss": 3.88088025336119,
      "tokens_seen": 542456832
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042202607823470413,
      "loss": 2.8222,
      "theoretical_loss": 3.8808304634155535,
      "tokens_seen": 542522368
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004220160481444333,
      "loss": 2.9794,
      "theoretical_loss": 3.8807806811679595,
      "tokens_seen": 542587904
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004220060180541625,
      "loss": 2.7943,
      "theoretical_loss": 3.8807309066162885,
      "tokens_seen": 542653440
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004219959879638917,
      "loss": 3.0174,
      "theoretical_loss": 3.8806811397584218,
      "tokens_seen": 542718976
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042198595787362085,
      "loss": 3.0142,
      "theoretical_loss": 3.8806313805922414,
      "tokens_seen": 542784512
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042197592778335004,
      "loss": 2.6437,
      "theoretical_loss": 3.88058162911563,
      "tokens_seen": 542850048
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042196589769307927,
      "loss": 3.1293,
      "theoretical_loss": 3.8805318853264708,
      "tokens_seen": 542915584
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004219558676028084,
      "loss": 3.0661,
      "theoretical_loss": 3.8804821492226482,
      "tokens_seen": 542981120
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042194583751253763,
      "loss": 2.8919,
      "theoretical_loss": 3.880432420802048,
      "tokens_seen": 543046656
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042193580742226676,
      "loss": 2.8511,
      "theoretical_loss": 3.880382700062556,
      "tokens_seen": 543112192
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.000421925777331996,
      "loss": 2.9383,
      "theoretical_loss": 3.880332987002059,
      "tokens_seen": 543177728
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042191574724172523,
      "loss": 2.8971,
      "theoretical_loss": 3.8802832816184445,
      "tokens_seen": 543243264
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042190571715145436,
      "loss": 2.8064,
      "theoretical_loss": 3.880233583909602,
      "tokens_seen": 543308800
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004218956870611836,
      "loss": 2.9946,
      "theoretical_loss": 3.8801838938734203,
      "tokens_seen": 543374336
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004218856569709127,
      "loss": 2.9893,
      "theoretical_loss": 3.8801342115077895,
      "tokens_seen": 543439872
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042187562688064196,
      "loss": 3.1243,
      "theoretical_loss": 3.8800845368106014,
      "tokens_seen": 543505408
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042186559679037114,
      "loss": 3.0212,
      "theoretical_loss": 3.880034869779747,
      "tokens_seen": 543570944
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004218555667001003,
      "loss": 3.2732,
      "theoretical_loss": 3.8799852104131194,
      "tokens_seen": 543636480
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004218455366098295,
      "loss": 2.9189,
      "theoretical_loss": 3.879935558708613,
      "tokens_seen": 543702016
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042183550651955873,
      "loss": 3.024,
      "theoretical_loss": 3.879885914664121,
      "tokens_seen": 543767552
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042182547642928786,
      "loss": 3.0686,
      "theoretical_loss": 3.8798362782775393,
      "tokens_seen": 543833088
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004218154463390171,
      "loss": 2.9981,
      "theoretical_loss": 3.879786649546764,
      "tokens_seen": 543898624
    },
    {
      "epoch": 1.07,
      "objective/train/docs_used": 620805,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7509706020355225,
      "objective/train/theoretical_loss": 3.8797494330215407,
      "objective/train/tokens_used": 564407776,
      "theoretical_loss": 3.8797494330215407,
      "tokens_seen": 543947776
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004218054162487462,
      "loss": 2.807,
      "theoretical_loss": 3.879737028469692,
      "tokens_seen": 543964160
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042179538615847546,
      "loss": 2.9816,
      "theoretical_loss": 3.8796874150442213,
      "tokens_seen": 544029696
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00042178535606820464,
      "loss": 2.8923,
      "theoretical_loss": 3.8796378092682495,
      "tokens_seen": 544095232
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0004217753259779338,
      "loss": 2.6996,
      "theoretical_loss": 3.8795882111396773,
      "tokens_seen": 544160768
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.000421765295887663,
      "loss": 2.9286,
      "theoretical_loss": 3.8795386206564038,
      "tokens_seen": 544226304
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004217552657973922,
      "loss": 2.9942,
      "theoretical_loss": 3.8794890378163314,
      "tokens_seen": 544291840
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00042174523570712136,
      "loss": 3.0129,
      "theoretical_loss": 3.879439462617361,
      "tokens_seen": 544357376
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004217352056168506,
      "loss": 2.9842,
      "theoretical_loss": 3.8793898950573955,
      "tokens_seen": 544422912
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004217251755265797,
      "loss": 3.0301,
      "theoretical_loss": 3.8793403351343385,
      "tokens_seen": 544488448
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00042171514543630896,
      "loss": 2.7943,
      "theoretical_loss": 3.8792907828460947,
      "tokens_seen": 544553984
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004217051153460381,
      "loss": 2.812,
      "theoretical_loss": 3.8792412381905685,
      "tokens_seen": 544619520
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004216950852557673,
      "loss": 2.8875,
      "theoretical_loss": 3.879191701165667,
      "tokens_seen": 544685056
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004216850551654965,
      "loss": 2.8014,
      "theoretical_loss": 3.8791421717692964,
      "tokens_seen": 544750592
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004216750250752257,
      "loss": 2.9776,
      "theoretical_loss": 3.8790926499993645,
      "tokens_seen": 544816128
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00042166499498495487,
      "loss": 3.0308,
      "theoretical_loss": 3.87904313585378,
      "tokens_seen": 544881664
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004216549648946841,
      "loss": 2.9902,
      "theoretical_loss": 3.878993629330452,
      "tokens_seen": 544947200
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00042164493480441323,
      "loss": 2.8341,
      "theoretical_loss": 3.878944130427291,
      "tokens_seen": 545012736
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00042163490471414246,
      "loss": 2.7959,
      "theoretical_loss": 3.878894639142208,
      "tokens_seen": 545078272
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004216248746238716,
      "loss": 3.1664,
      "theoretical_loss": 3.8788451554731145,
      "tokens_seen": 545143808
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004216148445336008,
      "loss": 2.8896,
      "theoretical_loss": 3.878795679417923,
      "tokens_seen": 545209344
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00042160481444333,
      "loss": 3.0888,
      "theoretical_loss": 3.8787462109745476,
      "tokens_seen": 545274880
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004215947843530592,
      "loss": 3.0511,
      "theoretical_loss": 3.878696750140902,
      "tokens_seen": 545340416
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00042158475426278837,
      "loss": 2.7537,
      "theoretical_loss": 3.8786472969149015,
      "tokens_seen": 545405952
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00042157472417251755,
      "loss": 2.9606,
      "theoretical_loss": 3.8785978512944626,
      "tokens_seen": 545471488
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00042156469408224673,
      "loss": 2.8147,
      "theoretical_loss": 3.878548413277501,
      "tokens_seen": 545537024
    },
    {
      "epoch": 1.08,
      "objective/train/docs_used": 621470,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.958339214324951,
      "objective/train/theoretical_loss": 3.8785113397533095,
      "objective/train/tokens_used": 566046176,
      "theoretical_loss": 3.8785113397533095,
      "tokens_seen": 545586176
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00042155466399197597,
      "loss": 2.9194,
      "theoretical_loss": 3.878498982861935,
      "tokens_seen": 545602560
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004215446339017051,
      "loss": 2.8887,
      "theoretical_loss": 3.8784495600456825,
      "tokens_seen": 545668096
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00042153460381143433,
      "loss": 3.0382,
      "theoretical_loss": 3.8784001448266636,
      "tokens_seen": 545733632
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004215245737211635,
      "loss": 2.8951,
      "theoretical_loss": 3.8783507372027977,
      "tokens_seen": 545799168
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004215145436308927,
      "loss": 2.7978,
      "theoretical_loss": 3.8783013371720054,
      "tokens_seen": 545864704
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004215045135406219,
      "loss": 3.1933,
      "theoretical_loss": 3.8782519447322086,
      "tokens_seen": 545930240
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00042149448345035105,
      "loss": 2.7269,
      "theoretical_loss": 3.8782025598813297,
      "tokens_seen": 545995776
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00042148445336008024,
      "loss": 3.0344,
      "theoretical_loss": 3.878153182617292,
      "tokens_seen": 546061312
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00042147442326980947,
      "loss": 2.9469,
      "theoretical_loss": 3.8781038129380203,
      "tokens_seen": 546126848
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004214643931795386,
      "loss": 2.9681,
      "theoretical_loss": 3.8780544508414394,
      "tokens_seen": 546192384
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00042145436308926783,
      "loss": 2.7063,
      "theoretical_loss": 3.8780050963254746,
      "tokens_seen": 546257920
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00042144433299899696,
      "loss": 2.8749,
      "theoretical_loss": 3.877955749388052,
      "tokens_seen": 546323456
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004214343029087262,
      "loss": 2.9719,
      "theoretical_loss": 3.8779064100270997,
      "tokens_seen": 546388992
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004214242728184554,
      "loss": 2.941,
      "theoretical_loss": 3.877857078240546,
      "tokens_seen": 546454528
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00042141424272818456,
      "loss": 3.0534,
      "theoretical_loss": 3.8778077540263194,
      "tokens_seen": 546520064
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00042140421263791374,
      "loss": 2.7779,
      "theoretical_loss": 3.877758437382351,
      "tokens_seen": 546585600
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004213941825476429,
      "loss": 2.7718,
      "theoretical_loss": 3.877709128306569,
      "tokens_seen": 546651136
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004213841524573721,
      "loss": 2.7369,
      "theoretical_loss": 3.8776598267969073,
      "tokens_seen": 546716672
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00042137412236710134,
      "loss": 2.6951,
      "theoretical_loss": 3.877610532851297,
      "tokens_seen": 546782208
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00042136409227683046,
      "loss": 3.0101,
      "theoretical_loss": 3.8775612464676716,
      "tokens_seen": 546847744
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004213540621865597,
      "loss": 2.9433,
      "theoretical_loss": 3.8775119676439647,
      "tokens_seen": 546913280
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004213440320962889,
      "loss": 2.6857,
      "theoretical_loss": 3.877462696378111,
      "tokens_seen": 546978816
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00042133400200601806,
      "loss": 2.8557,
      "theoretical_loss": 3.877413432668046,
      "tokens_seen": 547044352
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00042132397191574724,
      "loss": 3.0713,
      "theoretical_loss": 3.8773641765117066,
      "tokens_seen": 547109888
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004213139418254764,
      "loss": 3.0662,
      "theoretical_loss": 3.8773149279070296,
      "tokens_seen": 547175424
    },
    {
      "epoch": 1.08,
      "objective/train/docs_used": 622904,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.055859327316284,
      "objective/train/theoretical_loss": 3.8772779964080595,
      "objective/train/tokens_used": 567684576,
      "theoretical_loss": 3.8772779964080595,
      "tokens_seen": 547224576
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004213039117352056,
      "loss": 2.6809,
      "theoretical_loss": 3.8772656868519526,
      "tokens_seen": 547240960
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00042129388164493484,
      "loss": 3.0167,
      "theoretical_loss": 3.877216453344415,
      "tokens_seen": 547306496
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00042128385155466397,
      "loss": 2.8911,
      "theoretical_loss": 3.8771672273823556,
      "tokens_seen": 547372032
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004212738214643932,
      "loss": 3.198,
      "theoretical_loss": 3.8771180089637154,
      "tokens_seen": 547437568
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00042126379137412233,
      "loss": 2.8962,
      "theoretical_loss": 3.8770687980864356,
      "tokens_seen": 547503104
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00042125376128385156,
      "loss": 3.0193,
      "theoretical_loss": 3.8770195947484583,
      "tokens_seen": 547568640
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00042124373119358075,
      "loss": 2.8844,
      "theoretical_loss": 3.8769703989477255,
      "tokens_seen": 547634176
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004212337011033099,
      "loss": 2.9321,
      "theoretical_loss": 3.8769212106821813,
      "tokens_seen": 547699712
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004212236710130391,
      "loss": 2.9915,
      "theoretical_loss": 3.8768720299497703,
      "tokens_seen": 547765248
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004212136409227683,
      "loss": 2.6956,
      "theoretical_loss": 3.8768228567484377,
      "tokens_seen": 547830784
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00042120361083249747,
      "loss": 2.4523,
      "theoretical_loss": 3.8767736910761297,
      "tokens_seen": 547896320
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004211935807422267,
      "loss": 3.2263,
      "theoretical_loss": 3.8767245329307927,
      "tokens_seen": 547961856
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00042118355065195583,
      "loss": 2.8253,
      "theoretical_loss": 3.8766753823103746,
      "tokens_seen": 548027392
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00042117352056168507,
      "loss": 2.6039,
      "theoretical_loss": 3.8766262392128237,
      "tokens_seen": 548092928
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004211634904714143,
      "loss": 2.9644,
      "theoretical_loss": 3.876577103636089,
      "tokens_seen": 548158464
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00042115346038114343,
      "loss": 2.9272,
      "theoretical_loss": 3.8765279755781217,
      "tokens_seen": 548224000
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00042114343029087266,
      "loss": 2.8769,
      "theoretical_loss": 3.876478855036872,
      "tokens_seen": 548289536
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004211334002006018,
      "loss": 3.018,
      "theoretical_loss": 3.876429742010291,
      "tokens_seen": 548355072
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00042112337011033103,
      "loss": 3.134,
      "theoretical_loss": 3.8763806364963314,
      "tokens_seen": 548420608
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004211133400200602,
      "loss": 2.9645,
      "theoretical_loss": 3.876331538492947,
      "tokens_seen": 548486144
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004211033099297894,
      "loss": 2.9324,
      "theoretical_loss": 3.8762824479980917,
      "tokens_seen": 548551680
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00042109327983951857,
      "loss": 2.7315,
      "theoretical_loss": 3.8762333650097203,
      "tokens_seen": 548617216
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00042108324974924775,
      "loss": 2.7846,
      "theoretical_loss": 3.8761842895257885,
      "tokens_seen": 548682752
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00042107321965897693,
      "loss": 2.798,
      "theoretical_loss": 3.8761352215442524,
      "tokens_seen": 548748288
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00042106318956870617,
      "loss": 2.8967,
      "theoretical_loss": 3.8760861610630695,
      "tokens_seen": 548813824
    },
    {
      "epoch": 1.08,
      "objective/train/docs_used": 623446,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.478111743927002,
      "objective/train/theoretical_loss": 3.876049370623061,
      "objective/train/tokens_used": 569322976,
      "theoretical_loss": 3.876049370623061,
      "tokens_seen": 548862976
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004210531594784353,
      "loss": 3.0132,
      "theoretical_loss": 3.8760371080801983,
      "tokens_seen": 548879360
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00042104312938816453,
      "loss": 2.8714,
      "theoretical_loss": 3.8759880625935974,
      "tokens_seen": 548944896
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004210330992978937,
      "loss": 3.1069,
      "theoretical_loss": 3.8759390246012266,
      "tokens_seen": 549010432
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004210230692076229,
      "loss": 2.8727,
      "theoretical_loss": 3.8758899941010463,
      "tokens_seen": 549075968
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004210130391173521,
      "loss": 2.7593,
      "theoretical_loss": 3.8758409710910176,
      "tokens_seen": 549141504
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00042100300902708125,
      "loss": 2.9946,
      "theoretical_loss": 3.8757919555691025,
      "tokens_seen": 549207040
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00042099297893681044,
      "loss": 2.9204,
      "theoretical_loss": 3.875742947533264,
      "tokens_seen": 549272576
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00042098294884653967,
      "loss": 2.8915,
      "theoretical_loss": 3.8756939469814666,
      "tokens_seen": 549338112
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004209729187562688,
      "loss": 2.9225,
      "theoretical_loss": 3.8756449539116735,
      "tokens_seen": 549403648
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00042096288866599803,
      "loss": 3.0075,
      "theoretical_loss": 3.8755959683218504,
      "tokens_seen": 549469184
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00042095285857572716,
      "loss": 2.6968,
      "theoretical_loss": 3.8755469902099633,
      "tokens_seen": 549534720
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004209428284854564,
      "loss": 2.8839,
      "theoretical_loss": 3.875498019573979,
      "tokens_seen": 549600256
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004209327983951856,
      "loss": 3.1297,
      "theoretical_loss": 3.875449056411866,
      "tokens_seen": 549665792
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00042092276830491476,
      "loss": 2.8652,
      "theoretical_loss": 3.875400100721592,
      "tokens_seen": 549731328
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00042091273821464394,
      "loss": 3.0584,
      "theoretical_loss": 3.8753511525011257,
      "tokens_seen": 549796864
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004209027081243731,
      "loss": 2.7467,
      "theoretical_loss": 3.875302211748439,
      "tokens_seen": 549862400
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004208926780341023,
      "loss": 3.218,
      "theoretical_loss": 3.8752532784615004,
      "tokens_seen": 549927936
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00042088264794383154,
      "loss": 2.906,
      "theoretical_loss": 3.8752043526382827,
      "tokens_seen": 549993472
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00042087261785356066,
      "loss": 2.7583,
      "theoretical_loss": 3.875155434276759,
      "tokens_seen": 550059008
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004208625877632899,
      "loss": 2.6835,
      "theoretical_loss": 3.8751065233749005,
      "tokens_seen": 550124544
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004208525576730191,
      "loss": 2.6696,
      "theoretical_loss": 3.8750576199306837,
      "tokens_seen": 550190080
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00042084252758274826,
      "loss": 3.2374,
      "theoretical_loss": 3.8750087239420807,
      "tokens_seen": 550255616
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00042083249749247744,
      "loss": 2.8243,
      "theoretical_loss": 3.8749598354070693,
      "tokens_seen": 550321152
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004208224674022066,
      "loss": 2.954,
      "theoretical_loss": 3.8749109543236253,
      "tokens_seen": 550386688
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004208124373119358,
      "loss": 2.8086,
      "theoretical_loss": 3.8748620806897254,
      "tokens_seen": 550452224
    },
    {
      "epoch": 1.08,
      "objective/train/docs_used": 624875,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.574239730834961,
      "objective/train/theoretical_loss": 3.8748254303518475,
      "objective/train/tokens_used": 570961376,
      "theoretical_loss": 3.8748254303518475,
      "tokens_seen": 550501376
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00042080240722166504,
      "loss": 3.106,
      "theoretical_loss": 3.874813214503348,
      "tokens_seen": 550517760
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00042079237713139417,
      "loss": 2.974,
      "theoretical_loss": 3.8747643557624714,
      "tokens_seen": 550583296
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004207823470411234,
      "loss": 2.9065,
      "theoretical_loss": 3.874715504465075,
      "tokens_seen": 550648832
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00042077231695085253,
      "loss": 2.9074,
      "theoretical_loss": 3.87466666060914,
      "tokens_seen": 550714368
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00042076228686058176,
      "loss": 2.8743,
      "theoretical_loss": 3.8746178241926468,
      "tokens_seen": 550779904
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00042075225677031095,
      "loss": 3.0312,
      "theoretical_loss": 3.8745689952135773,
      "tokens_seen": 550845440
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004207422266800401,
      "loss": 2.8656,
      "theoretical_loss": 3.8745201736699144,
      "tokens_seen": 550910976
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004207321965897693,
      "loss": 3.0798,
      "theoretical_loss": 3.8744713595596414,
      "tokens_seen": 550976512
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004207221664994985,
      "loss": 3.0155,
      "theoretical_loss": 3.874422552880743,
      "tokens_seen": 551042048
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00042071213640922767,
      "loss": 2.8655,
      "theoretical_loss": 3.8743737536312035,
      "tokens_seen": 551107584
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004207021063189569,
      "loss": 2.895,
      "theoretical_loss": 3.87432496180901,
      "tokens_seen": 551173120
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00042069207622868603,
      "loss": 2.878,
      "theoretical_loss": 3.8742761774121472,
      "tokens_seen": 551238656
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00042068204613841527,
      "loss": 2.9071,
      "theoretical_loss": 3.874227400438604,
      "tokens_seen": 551304192
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00042067201604814445,
      "loss": 2.8582,
      "theoretical_loss": 3.8741786308863677,
      "tokens_seen": 551369728
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00042066198595787363,
      "loss": 2.8426,
      "theoretical_loss": 3.8741298687534282,
      "tokens_seen": 551435264
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004206519558676028,
      "loss": 2.9456,
      "theoretical_loss": 3.874081114037775,
      "tokens_seen": 551500800
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.000420641925777332,
      "loss": 2.6779,
      "theoretical_loss": 3.874032366737398,
      "tokens_seen": 551566336
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004206318956870612,
      "loss": 2.8382,
      "theoretical_loss": 3.873983626850289,
      "tokens_seen": 551631872
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004206218655967904,
      "loss": 2.7917,
      "theoretical_loss": 3.8739348943744396,
      "tokens_seen": 551697408
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00042061183550651954,
      "loss": 2.7561,
      "theoretical_loss": 3.8738861693078435,
      "tokens_seen": 551762944
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00042060180541624877,
      "loss": 3.085,
      "theoretical_loss": 3.8738374516484937,
      "tokens_seen": 551828480
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004205917753259779,
      "loss": 2.8885,
      "theoretical_loss": 3.8737887413943852,
      "tokens_seen": 551894016
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00042058174523570713,
      "loss": 2.8237,
      "theoretical_loss": 3.8737400385435126,
      "tokens_seen": 551959552
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004205717151454363,
      "loss": 3.0146,
      "theoretical_loss": 3.8736913430938724,
      "tokens_seen": 552025088
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004205616850551655,
      "loss": 2.8244,
      "theoretical_loss": 3.8736426550434615,
      "tokens_seen": 552090624
    },
    {
      "epoch": 1.08,
      "objective/train/docs_used": 625640,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8349080085754395,
      "objective/train/theoretical_loss": 3.8736061438601923,
      "objective/train/tokens_used": 572599776,
      "theoretical_loss": 3.8736061438601923,
      "tokens_seen": 552139776
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004205516549648947,
      "loss": 2.9482,
      "theoretical_loss": 3.8735939743902765,
      "tokens_seen": 552156160
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004205416248746239,
      "loss": 2.7414,
      "theoretical_loss": 3.873545301132317,
      "tokens_seen": 552221696
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00042053159478435304,
      "loss": 2.9545,
      "theoretical_loss": 3.873496635267581,
      "tokens_seen": 552287232
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004205215646940823,
      "loss": 2.6246,
      "theoretical_loss": 3.8734479767940693,
      "tokens_seen": 552352768
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004205115346038114,
      "loss": 2.9388,
      "theoretical_loss": 3.8733993257097827,
      "tokens_seen": 552418304
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00042050150451354064,
      "loss": 3.0344,
      "theoretical_loss": 3.8733506820127213,
      "tokens_seen": 552483840
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004204914744232698,
      "loss": 2.99,
      "theoretical_loss": 3.8733020457008887,
      "tokens_seen": 552549376
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.000420481444332999,
      "loss": 3.006,
      "theoretical_loss": 3.8732534167722874,
      "tokens_seen": 552614912
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004204714142427282,
      "loss": 2.7293,
      "theoretical_loss": 3.873204795224921,
      "tokens_seen": 552680448
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00042046138415245736,
      "loss": 2.6337,
      "theoretical_loss": 3.873156181056795,
      "tokens_seen": 552745984
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00042045135406218654,
      "loss": 3.1811,
      "theoretical_loss": 3.8731075742659136,
      "tokens_seen": 552811520
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004204413239719158,
      "loss": 3.0877,
      "theoretical_loss": 3.8730589748502835,
      "tokens_seen": 552877056
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004204312938816449,
      "loss": 2.8474,
      "theoretical_loss": 3.8730103828079114,
      "tokens_seen": 552942592
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00042042126379137414,
      "loss": 3.0734,
      "theoretical_loss": 3.8729617981368047,
      "tokens_seen": 553008128
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004204112337011033,
      "loss": 3.238,
      "theoretical_loss": 3.872913220834973,
      "tokens_seen": 553073664
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004204012036108325,
      "loss": 2.9764,
      "theoretical_loss": 3.872864650900424,
      "tokens_seen": 553139200
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00042039117352056174,
      "loss": 2.9151,
      "theoretical_loss": 3.8728160883311684,
      "tokens_seen": 553204736
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00042038114343029086,
      "loss": 3.0027,
      "theoretical_loss": 3.8727675331252174,
      "tokens_seen": 553270272
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004203711133400201,
      "loss": 2.6286,
      "theoretical_loss": 3.872718985280582,
      "tokens_seen": 553335808
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004203610832497493,
      "loss": 3.0266,
      "theoretical_loss": 3.8726704447952747,
      "tokens_seen": 553401344
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00042035105315947846,
      "loss": 3.1303,
      "theoretical_loss": 3.8726219116673084,
      "tokens_seen": 553466880
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00042034102306920764,
      "loss": 2.82,
      "theoretical_loss": 3.8725733858946967,
      "tokens_seen": 553532416
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004203309929789368,
      "loss": 2.9129,
      "theoretical_loss": 3.8725248674754553,
      "tokens_seen": 553597952
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.000420320962888666,
      "loss": 2.728,
      "theoretical_loss": 3.8724763564075984,
      "tokens_seen": 553663488
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00042031093279839524,
      "loss": 2.947,
      "theoretical_loss": 3.8724278526891425,
      "tokens_seen": 553729024
    },
    {
      "epoch": 1.08,
      "objective/train/docs_used": 626970,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.11159610748291,
      "objective/train/theoretical_loss": 3.872391479722152,
      "objective/train/tokens_used": 574238176,
      "theoretical_loss": 3.872391479722152,
      "tokens_seen": 553778176
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00042030090270812437,
      "loss": 2.9388,
      "theoretical_loss": 3.872379356318105,
      "tokens_seen": 553794560
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004202908726178536,
      "loss": 2.8643,
      "theoretical_loss": 3.8723308672925025,
      "tokens_seen": 553860096
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00042028084252758273,
      "loss": 2.8689,
      "theoretical_loss": 3.8722823856103554,
      "tokens_seen": 553925632
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00042027081243731196,
      "loss": 3.1274,
      "theoretical_loss": 3.8722339112696815,
      "tokens_seen": 553991168
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00042026078234704115,
      "loss": 2.9725,
      "theoretical_loss": 3.8721854442685,
      "tokens_seen": 554056704
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004202507522567703,
      "loss": 2.8734,
      "theoretical_loss": 3.8721369846048344,
      "tokens_seen": 554122240
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004202407221664995,
      "loss": 2.7302,
      "theoretical_loss": 3.8720885322767042,
      "tokens_seen": 554187776
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004202306920762287,
      "loss": 2.8491,
      "theoretical_loss": 3.872040087282132,
      "tokens_seen": 554253312
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00042022066198595787,
      "loss": 2.9801,
      "theoretical_loss": 3.8719916496191407,
      "tokens_seen": 554318848
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004202106318956871,
      "loss": 2.8109,
      "theoretical_loss": 3.871943219285755,
      "tokens_seen": 554384384
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00042020060180541623,
      "loss": 3.0374,
      "theoretical_loss": 3.8718947962799986,
      "tokens_seen": 554449920
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00042019057171514547,
      "loss": 2.6717,
      "theoretical_loss": 3.871846380599897,
      "tokens_seen": 554515456
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00042018054162487465,
      "loss": 2.6651,
      "theoretical_loss": 3.8717979722434777,
      "tokens_seen": 554580992
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00042017051153460383,
      "loss": 2.9567,
      "theoretical_loss": 3.871749571208766,
      "tokens_seen": 554646528
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.000420160481444333,
      "loss": 2.9258,
      "theoretical_loss": 3.87170117749379,
      "tokens_seen": 554712064
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004201504513540622,
      "loss": 2.8143,
      "theoretical_loss": 3.8716527910965786,
      "tokens_seen": 554777600
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004201404212637914,
      "loss": 3.085,
      "theoretical_loss": 3.8716044120151603,
      "tokens_seen": 554843136
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004201303911735206,
      "loss": 2.8709,
      "theoretical_loss": 3.8715560402475657,
      "tokens_seen": 554908672
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00042012036108324974,
      "loss": 3.0459,
      "theoretical_loss": 3.8715076757918254,
      "tokens_seen": 554974208
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00042011033099297897,
      "loss": 3.0664,
      "theoretical_loss": 3.8714593186459703,
      "tokens_seen": 555039744
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004201003009027081,
      "loss": 2.8809,
      "theoretical_loss": 3.871410968808034,
      "tokens_seen": 555105280
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00042009027081243733,
      "loss": 2.9343,
      "theoretical_loss": 3.8713626262760474,
      "tokens_seen": 555170816
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004200802407221665,
      "loss": 2.6519,
      "theoretical_loss": 3.871314291048046,
      "tokens_seen": 555236352
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004200702106318957,
      "loss": 2.856,
      "theoretical_loss": 3.871265963122064,
      "tokens_seen": 555301888
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004200601805416249,
      "loss": 2.5503,
      "theoretical_loss": 3.8712176424961373,
      "tokens_seen": 555367424
    },
    {
      "epoch": 1.08,
      "objective/train/docs_used": 627698,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.411449432373047,
      "objective/train/theoretical_loss": 3.871181406816171,
      "objective/train/tokens_used": 575876576,
      "theoretical_loss": 3.871181406816171,
      "tokens_seen": 555416576
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004200501504513541,
      "loss": 3.0096,
      "theoretical_loss": 3.8711693291683003,
      "tokens_seen": 555432960
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00042004012036108324,
      "loss": 3.0007,
      "theoretical_loss": 3.871121023136591,
      "tokens_seen": 555498496
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004200300902708125,
      "loss": 2.7156,
      "theoretical_loss": 3.871072724399047,
      "tokens_seen": 555564032
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004200200601805416,
      "loss": 2.8436,
      "theoretical_loss": 3.871024432953706,
      "tokens_seen": 555629568
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00042001003009027084,
      "loss": 2.6881,
      "theoretical_loss": 3.870976148798608,
      "tokens_seen": 555695104
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00042,
      "loss": 3.0327,
      "theoretical_loss": 3.8709278719317917,
      "tokens_seen": 555760640
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004199899699097292,
      "loss": 2.7813,
      "theoretical_loss": 3.8708796023512986,
      "tokens_seen": 555826176
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004199799398194584,
      "loss": 2.7957,
      "theoretical_loss": 3.8708313400551697,
      "tokens_seen": 555891712
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041996990972918756,
      "loss": 2.955,
      "theoretical_loss": 3.8707830850414475,
      "tokens_seen": 555957248
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041995987963891674,
      "loss": 2.7701,
      "theoretical_loss": 3.8707348373081745,
      "tokens_seen": 556022784
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.000419949849548646,
      "loss": 2.7127,
      "theoretical_loss": 3.870686596853395,
      "tokens_seen": 556088320
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004199398194583751,
      "loss": 2.9101,
      "theoretical_loss": 3.870638363675152,
      "tokens_seen": 556153856
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041992978936810434,
      "loss": 3.0043,
      "theoretical_loss": 3.870590137771492,
      "tokens_seen": 556219392
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041991975927783347,
      "loss": 2.8558,
      "theoretical_loss": 3.870541919140461,
      "tokens_seen": 556284928
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004199097291875627,
      "loss": 2.9357,
      "theoretical_loss": 3.8704937077801045,
      "tokens_seen": 556350464
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004198996990972919,
      "loss": 3.0304,
      "theoretical_loss": 3.8704455036884706,
      "tokens_seen": 556416000
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041988966900702106,
      "loss": 2.5689,
      "theoretical_loss": 3.8703973068636075,
      "tokens_seen": 556481536
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041987963891675025,
      "loss": 2.8771,
      "theoretical_loss": 3.870349117303564,
      "tokens_seen": 556547072
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004198696088264795,
      "loss": 2.7484,
      "theoretical_loss": 3.87030093500639,
      "tokens_seen": 556612608
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004198595787362086,
      "loss": 2.9133,
      "theoretical_loss": 3.8702527599701355,
      "tokens_seen": 556678144
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041984954864593784,
      "loss": 2.7249,
      "theoretical_loss": 3.8702045921928523,
      "tokens_seen": 556743680
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041983951855566697,
      "loss": 2.9768,
      "theoretical_loss": 3.8701564316725916,
      "tokens_seen": 556809216
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004198294884653962,
      "loss": 2.7145,
      "theoretical_loss": 3.8701082784074075,
      "tokens_seen": 556874752
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004198194583751254,
      "loss": 3.0215,
      "theoretical_loss": 3.870060132395351,
      "tokens_seen": 556940288
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041980942828485457,
      "loss": 2.9044,
      "theoretical_loss": 3.8700119936344786,
      "tokens_seen": 557005824
    },
    {
      "epoch": 1.08,
      "objective/train/docs_used": 628817,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7914021015167236,
      "objective/train/theoretical_loss": 3.8699758943212434,
      "objective/train/tokens_used": 577514976,
      "theoretical_loss": 3.8699758943212434,
      "tokens_seen": 557054976
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041979939819458375,
      "loss": 2.962,
      "theoretical_loss": 3.8699638621228445,
      "tokens_seen": 557071360
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041978936810431293,
      "loss": 2.7353,
      "theoretical_loss": 3.8699157378585043,
      "tokens_seen": 557136896
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004197793380140421,
      "loss": 2.9422,
      "theoretical_loss": 3.869867620839514,
      "tokens_seen": 557202432
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041976930792377135,
      "loss": 3.1109,
      "theoretical_loss": 3.8698195110639317,
      "tokens_seen": 557267968
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041975927783350047,
      "loss": 2.8179,
      "theoretical_loss": 3.8697714085298145,
      "tokens_seen": 557333504
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004197492477432297,
      "loss": 2.9445,
      "theoretical_loss": 3.8697233132352222,
      "tokens_seen": 557399040
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041973921765295883,
      "loss": 2.9836,
      "theoretical_loss": 3.869675225178213,
      "tokens_seen": 557464576
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041972918756268807,
      "loss": 2.9379,
      "theoretical_loss": 3.8696271443568477,
      "tokens_seen": 557530112
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041971915747241725,
      "loss": 2.803,
      "theoretical_loss": 3.869579070769187,
      "tokens_seen": 557595648
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041970912738214643,
      "loss": 3.052,
      "theoretical_loss": 3.8695310044132927,
      "tokens_seen": 557661184
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004196990972918756,
      "loss": 2.9134,
      "theoretical_loss": 3.8694829452872277,
      "tokens_seen": 557726720
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041968906720160485,
      "loss": 2.7973,
      "theoretical_loss": 3.8694348933890543,
      "tokens_seen": 557792256
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.000419679037111334,
      "loss": 3.0926,
      "theoretical_loss": 3.8693868487168372,
      "tokens_seen": 557857792
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004196690070210632,
      "loss": 3.0218,
      "theoretical_loss": 3.8693388112686407,
      "tokens_seen": 557923328
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004196589769307924,
      "loss": 2.981,
      "theoretical_loss": 3.8692907810425305,
      "tokens_seen": 557988864
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004196489468405216,
      "loss": 3.0689,
      "theoretical_loss": 3.869242758036573,
      "tokens_seen": 558054400
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004196389167502508,
      "loss": 2.6322,
      "theoretical_loss": 3.8691947422488333,
      "tokens_seen": 558119936
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041962888665997994,
      "loss": 2.9541,
      "theoretical_loss": 3.8691467336773817,
      "tokens_seen": 558185472
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041961885656970917,
      "loss": 2.9968,
      "theoretical_loss": 3.8690987323202846,
      "tokens_seen": 558251008
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004196088264794383,
      "loss": 3.0707,
      "theoretical_loss": 3.8690507381756123,
      "tokens_seen": 558316544
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041959879638916753,
      "loss": 2.7787,
      "theoretical_loss": 3.8690027512414344,
      "tokens_seen": 558382080
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004195887662988967,
      "loss": 2.7311,
      "theoretical_loss": 3.868954771515821,
      "tokens_seen": 558447616
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004195787362086259,
      "loss": 2.9757,
      "theoretical_loss": 3.8689067989968438,
      "tokens_seen": 558513152
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004195687061183551,
      "loss": 3.019,
      "theoretical_loss": 3.868858833682575,
      "tokens_seen": 558578688
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004195586760280843,
      "loss": 2.808,
      "theoretical_loss": 3.868810875571088,
      "tokens_seen": 558644224
    },
    {
      "epoch": 1.08,
      "objective/train/docs_used": 629246,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.687079429626465,
      "objective/train/theoretical_loss": 3.8687749117131385,
      "objective/train/tokens_used": 579153376,
      "theoretical_loss": 3.8687749117131385,
      "tokens_seen": 558693376
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041954864593781344,
      "loss": 2.8716,
      "theoretical_loss": 3.868762924660455,
      "tokens_seen": 558709760
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004195386158475427,
      "loss": 2.9467,
      "theoretical_loss": 3.8687149809487518,
      "tokens_seen": 558775296
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004195285857572718,
      "loss": 2.692,
      "theoretical_loss": 3.8686670444340527,
      "tokens_seen": 558840832
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041951855566700104,
      "loss": 2.7968,
      "theoretical_loss": 3.8686191151144333,
      "tokens_seen": 558906368
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004195085255767302,
      "loss": 2.6929,
      "theoretical_loss": 3.868571192987971,
      "tokens_seen": 558971904
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004194984954864594,
      "loss": 3.0387,
      "theoretical_loss": 3.8685232780527423,
      "tokens_seen": 559037440
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004194884653961886,
      "loss": 2.7125,
      "theoretical_loss": 3.868475370306826,
      "tokens_seen": 559102976
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041947843530591776,
      "loss": 2.8212,
      "theoretical_loss": 3.8684274697483003,
      "tokens_seen": 559168512
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041946840521564694,
      "loss": 2.8875,
      "theoretical_loss": 3.8683795763752444,
      "tokens_seen": 559234048
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004194583751253762,
      "loss": 3.2464,
      "theoretical_loss": 3.8683316901857396,
      "tokens_seen": 559299584
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004194483450351053,
      "loss": 2.8747,
      "theoretical_loss": 3.868283811177866,
      "tokens_seen": 559365120
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041943831494483454,
      "loss": 2.6532,
      "theoretical_loss": 3.868235939349706,
      "tokens_seen": 559430656
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041942828485456367,
      "loss": 2.9353,
      "theoretical_loss": 3.868188074699341,
      "tokens_seen": 559496192
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004194182547642929,
      "loss": 2.8988,
      "theoretical_loss": 3.8681402172248553,
      "tokens_seen": 559561728
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004194082246740221,
      "loss": 2.664,
      "theoretical_loss": 3.8680923669243326,
      "tokens_seen": 559627264
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041939819458375126,
      "loss": 2.8466,
      "theoretical_loss": 3.8680445237958567,
      "tokens_seen": 559692800
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041938816449348045,
      "loss": 3.101,
      "theoretical_loss": 3.8679966878375143,
      "tokens_seen": 559758336
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004193781344032097,
      "loss": 2.7125,
      "theoretical_loss": 3.867948859047391,
      "tokens_seen": 559823872
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004193681043129388,
      "loss": 2.7661,
      "theoretical_loss": 3.8679010374235734,
      "tokens_seen": 559889408
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041935807422266804,
      "loss": 2.7567,
      "theoretical_loss": 3.8678532229641496,
      "tokens_seen": 559954944
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041934804413239717,
      "loss": 2.8698,
      "theoretical_loss": 3.867805415667208,
      "tokens_seen": 560020480
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004193380140421264,
      "loss": 2.8603,
      "theoretical_loss": 3.867757615530837,
      "tokens_seen": 560086016
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004193279839518556,
      "loss": 2.8624,
      "theoretical_loss": 3.8677098225531266,
      "tokens_seen": 560151552
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041931795386158477,
      "loss": 3.0192,
      "theoretical_loss": 3.867662036732168,
      "tokens_seen": 560217088
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041930792377131395,
      "loss": 2.552,
      "theoretical_loss": 3.8676142580660517,
      "tokens_seen": 560282624
    },
    {
      "epoch": 1.08,
      "objective/train/docs_used": 630481,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8848719596862793,
      "objective/train/theoretical_loss": 3.867578428760682,
      "objective/train/tokens_used": 580791776,
      "theoretical_loss": 3.867578428760682,
      "tokens_seen": 560331776
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041929789368104313,
      "loss": 2.9827,
      "theoretical_loss": 3.8675664865528705,
      "tokens_seen": 560348160
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004192878635907723,
      "loss": 2.9074,
      "theoretical_loss": 3.8675187221907166,
      "tokens_seen": 560413696
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041927783350050155,
      "loss": 2.7092,
      "theoretical_loss": 3.867470964977683,
      "tokens_seen": 560479232
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041926780341023067,
      "loss": 2.9471,
      "theoretical_loss": 3.867423214911865,
      "tokens_seen": 560544768
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004192577733199599,
      "loss": 2.9313,
      "theoretical_loss": 3.867375471991357,
      "tokens_seen": 560610304
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041924774322968904,
      "loss": 2.7035,
      "theoretical_loss": 3.8673277362142544,
      "tokens_seen": 560675840
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041923771313941827,
      "loss": 2.9026,
      "theoretical_loss": 3.867280007578654,
      "tokens_seen": 560741376
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041922768304914745,
      "loss": 2.7299,
      "theoretical_loss": 3.867232286082653,
      "tokens_seen": 560806912
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041921765295887663,
      "loss": 3.0076,
      "theoretical_loss": 3.867184571724349,
      "tokens_seen": 560872448
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004192076228686058,
      "loss": 2.8349,
      "theoretical_loss": 3.867136864501841,
      "tokens_seen": 560937984
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041919759277833505,
      "loss": 3.0804,
      "theoretical_loss": 3.867089164413228,
      "tokens_seen": 561003520
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004191875626880642,
      "loss": 2.8924,
      "theoretical_loss": 3.8670414714566093,
      "tokens_seen": 561069056
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004191775325977934,
      "loss": 2.9753,
      "theoretical_loss": 3.866993785630087,
      "tokens_seen": 561134592
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041916750250752254,
      "loss": 2.901,
      "theoretical_loss": 3.8669461069317617,
      "tokens_seen": 561200128
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004191574724172518,
      "loss": 2.9177,
      "theoretical_loss": 3.866898435359736,
      "tokens_seen": 561265664
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041914744232698095,
      "loss": 2.7428,
      "theoretical_loss": 3.866850770912113,
      "tokens_seen": 561331200
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041913741223671014,
      "loss": 2.6984,
      "theoretical_loss": 3.8668031135869962,
      "tokens_seen": 561396736
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004191273821464393,
      "loss": 3.0854,
      "theoretical_loss": 3.86675546338249,
      "tokens_seen": 561462272
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004191173520561685,
      "loss": 2.815,
      "theoretical_loss": 3.866707820296699,
      "tokens_seen": 561527808
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004191073219658977,
      "loss": 2.8582,
      "theoretical_loss": 3.8666601843277304,
      "tokens_seen": 561593344
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004190972918756269,
      "loss": 2.9834,
      "theoretical_loss": 3.866612555473689,
      "tokens_seen": 561658880
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041908726178535604,
      "loss": 2.9236,
      "theoretical_loss": 3.8665649337326835,
      "tokens_seen": 561724416
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004190772316950853,
      "loss": 2.8714,
      "theoretical_loss": 3.8665173191028215,
      "tokens_seen": 561789952
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004190672016048144,
      "loss": 2.7921,
      "theoretical_loss": 3.8664697115822118,
      "tokens_seen": 561855488
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041905717151454364,
      "loss": 2.8708,
      "theoretical_loss": 3.866422111168964,
      "tokens_seen": 561921024
    },
    {
      "epoch": 1.08,
      "objective/train/docs_used": 630917,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.474799871444702,
      "objective/train/theoretical_loss": 3.866386415522097,
      "objective/train/tokens_used": 582430176,
      "theoretical_loss": 3.866386415522097,
      "tokens_seen": 561970176
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004190471414242728,
      "loss": 2.7912,
      "theoretical_loss": 3.8663745178611872,
      "tokens_seen": 561986560
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.000419037111334002,
      "loss": 2.9873,
      "theoretical_loss": 3.866326931656994,
      "tokens_seen": 562052096
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004190270812437312,
      "loss": 2.8812,
      "theoretical_loss": 3.8662793525544954,
      "tokens_seen": 562117632
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004190170511534604,
      "loss": 2.6993,
      "theoretical_loss": 3.8662317805518036,
      "tokens_seen": 562183168
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041900702106318954,
      "loss": 3.0518,
      "theoretical_loss": 3.8661842156470314,
      "tokens_seen": 562248704
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004189969909729188,
      "loss": 2.944,
      "theoretical_loss": 3.866136657838293,
      "tokens_seen": 562314240
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004189869608826479,
      "loss": 2.7422,
      "theoretical_loss": 3.866089107123703,
      "tokens_seen": 562379776
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041897693079237714,
      "loss": 2.8368,
      "theoretical_loss": 3.8660415635013767,
      "tokens_seen": 562445312
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004189669007021063,
      "loss": 2.9691,
      "theoretical_loss": 3.86599402696943,
      "tokens_seen": 562510848
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004189568706118355,
      "loss": 2.8235,
      "theoretical_loss": 3.8659464975259787,
      "tokens_seen": 562576384
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004189468405215647,
      "loss": 2.9171,
      "theoretical_loss": 3.865898975169142,
      "tokens_seen": 562641920
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041893681043129387,
      "loss": 3.0027,
      "theoretical_loss": 3.865851459897037,
      "tokens_seen": 562707456
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041892678034102305,
      "loss": 2.9641,
      "theoretical_loss": 3.865803951707782,
      "tokens_seen": 562772992
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004189167502507523,
      "loss": 2.9334,
      "theoretical_loss": 3.8657564505994975,
      "tokens_seen": 562838528
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041890672016048146,
      "loss": 2.9198,
      "theoretical_loss": 3.8657089565703036,
      "tokens_seen": 562904064
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041889669007021065,
      "loss": 3.1526,
      "theoretical_loss": 3.865661469618321,
      "tokens_seen": 562969600
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004188866599799399,
      "loss": 3.0167,
      "theoretical_loss": 3.8656139897416715,
      "tokens_seen": 563035136
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.000418876629889669,
      "loss": 2.8326,
      "theoretical_loss": 3.865566516938478,
      "tokens_seen": 563100672
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041886659979939824,
      "loss": 2.9995,
      "theoretical_loss": 3.8655190512068627,
      "tokens_seen": 563166208
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041885656970912737,
      "loss": 2.9358,
      "theoretical_loss": 3.8654715925449503,
      "tokens_seen": 563231744
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004188465396188566,
      "loss": 2.7895,
      "theoretical_loss": 3.8654241409508656,
      "tokens_seen": 563297280
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004188365095285858,
      "loss": 2.9822,
      "theoretical_loss": 3.865376696422733,
      "tokens_seen": 563362816
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041882647943831497,
      "loss": 3.0189,
      "theoretical_loss": 3.865329258958679,
      "tokens_seen": 563428352
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041881644934804415,
      "loss": 2.7797,
      "theoretical_loss": 3.8652818285568307,
      "tokens_seen": 563493888
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041880641925777333,
      "loss": 2.861,
      "theoretical_loss": 3.8652344052153147,
      "tokens_seen": 563559424
    },
    {
      "epoch": 1.08,
      "objective/train/docs_used": 632338,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9370810985565186,
      "objective/train/theoretical_loss": 3.865198842341395,
      "objective/train/tokens_used": 584068576,
      "theoretical_loss": 3.865198842341395,
      "tokens_seen": 563608576
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004187963891675025,
      "loss": 2.935,
      "theoretical_loss": 3.86518698893226,
      "tokens_seen": 563624960
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041878635907723175,
      "loss": 2.9378,
      "theoretical_loss": 3.8651395797057946,
      "tokens_seen": 563690496
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004187763289869609,
      "loss": 2.7841,
      "theoretical_loss": 3.8650921775340494,
      "tokens_seen": 563756032
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004187662988966901,
      "loss": 2.921,
      "theoretical_loss": 3.8650447824151533,
      "tokens_seen": 563821568
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041875626880641924,
      "loss": 2.959,
      "theoretical_loss": 3.864997394347238,
      "tokens_seen": 563887104
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041874623871614847,
      "loss": 2.6183,
      "theoretical_loss": 3.864950013328435,
      "tokens_seen": 563952640
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041873620862587765,
      "loss": 2.9683,
      "theoretical_loss": 3.864902639356877,
      "tokens_seen": 564018176
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041872617853560683,
      "loss": 3.0088,
      "theoretical_loss": 3.864855272430697,
      "tokens_seen": 564083712
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.000418716148445336,
      "loss": 3.0662,
      "theoretical_loss": 3.864807912548029,
      "tokens_seen": 564149248
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041870611835506525,
      "loss": 2.9283,
      "theoretical_loss": 3.8647605597070074,
      "tokens_seen": 564214784
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004186960882647944,
      "loss": 2.8571,
      "theoretical_loss": 3.864713213905768,
      "tokens_seen": 564280320
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004186860581745236,
      "loss": 2.82,
      "theoretical_loss": 3.864665875142446,
      "tokens_seen": 564345856
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041867602808425274,
      "loss": 3.119,
      "theoretical_loss": 3.8646185434151787,
      "tokens_seen": 564411392
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.000418665997993982,
      "loss": 3.0035,
      "theoretical_loss": 3.864571218722103,
      "tokens_seen": 564476928
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041865596790371115,
      "loss": 2.8305,
      "theoretical_loss": 3.8645239010613572,
      "tokens_seen": 564542464
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041864593781344034,
      "loss": 2.8717,
      "theoretical_loss": 3.8644765904310803,
      "tokens_seen": 564608000
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004186359077231695,
      "loss": 2.7008,
      "theoretical_loss": 3.864429286829412,
      "tokens_seen": 564673536
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004186258776328987,
      "loss": 2.8845,
      "theoretical_loss": 3.8643819902544925,
      "tokens_seen": 564739072
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004186158475426279,
      "loss": 2.9494,
      "theoretical_loss": 3.864334700704462,
      "tokens_seen": 564804608
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004186058174523571,
      "loss": 2.9526,
      "theoretical_loss": 3.8642874181774634,
      "tokens_seen": 564870144
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041859578736208624,
      "loss": 2.8049,
      "theoretical_loss": 3.8642401426716377,
      "tokens_seen": 564935680
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004185857572718155,
      "loss": 2.703,
      "theoretical_loss": 3.8641928741851292,
      "tokens_seen": 565001216
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004185757271815446,
      "loss": 3.0939,
      "theoretical_loss": 3.864145612716081,
      "tokens_seen": 565066752
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041856569709127384,
      "loss": 2.748,
      "theoretical_loss": 3.8640983582626385,
      "tokens_seen": 565132288
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.000418555667001003,
      "loss": 3.0962,
      "theoretical_loss": 3.8640511108229454,
      "tokens_seen": 565197824
    },
    {
      "epoch": 1.08,
      "objective/train/docs_used": 633117,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.0730414390563965,
      "objective/train/theoretical_loss": 3.864015679844834,
      "objective/train/tokens_used": 585706976,
      "theoretical_loss": 3.864015679844834,
      "tokens_seen": 565246976
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004185456369107322,
      "loss": 3.0086,
      "theoretical_loss": 3.8640038703951483,
      "tokens_seen": 565263360
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004185356068204614,
      "loss": 2.9326,
      "theoretical_loss": 3.8639566369773943,
      "tokens_seen": 565328896
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004185255767301906,
      "loss": 2.6394,
      "theoretical_loss": 3.86390941056783,
      "tokens_seen": 565394432
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041851554663991974,
      "loss": 2.9686,
      "theoretical_loss": 3.8638621911646043,
      "tokens_seen": 565459968
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.000418505516549649,
      "loss": 2.9378,
      "theoretical_loss": 3.863814978765865,
      "tokens_seen": 565525504
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004184954864593781,
      "loss": 2.7113,
      "theoretical_loss": 3.8637677733697617,
      "tokens_seen": 565591040
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041848545636910734,
      "loss": 2.9256,
      "theoretical_loss": 3.8637205749744448,
      "tokens_seen": 565656576
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004184754262788365,
      "loss": 2.8471,
      "theoretical_loss": 3.8636733835780652,
      "tokens_seen": 565722112
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004184653961885657,
      "loss": 2.896,
      "theoretical_loss": 3.863626199178774,
      "tokens_seen": 565787648
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004184553660982949,
      "loss": 2.755,
      "theoretical_loss": 3.863579021774724,
      "tokens_seen": 565853184
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041844533600802407,
      "loss": 2.9088,
      "theoretical_loss": 3.8635318513640677,
      "tokens_seen": 565918720
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041843530591775325,
      "loss": 2.8796,
      "theoretical_loss": 3.863484687944959,
      "tokens_seen": 565984256
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004184252758274825,
      "loss": 2.6829,
      "theoretical_loss": 3.863437531515552,
      "tokens_seen": 566049792
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004184152457372116,
      "loss": 3.0297,
      "theoretical_loss": 3.8633903820740016,
      "tokens_seen": 566115328
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041840521564694085,
      "loss": 2.7752,
      "theoretical_loss": 3.863343239618464,
      "tokens_seen": 566180864
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041839518555666997,
      "loss": 3.0822,
      "theoretical_loss": 3.863296104147096,
      "tokens_seen": 566246400
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004183851554663992,
      "loss": 2.9419,
      "theoretical_loss": 3.8632489756580535,
      "tokens_seen": 566311936
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004183751253761284,
      "loss": 3.0283,
      "theoretical_loss": 3.8632018541494952,
      "tokens_seen": 566377472
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041836509528585757,
      "loss": 2.9068,
      "theoretical_loss": 3.8631547396195796,
      "tokens_seen": 566443008
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041835506519558675,
      "loss": 3.0821,
      "theoretical_loss": 3.863107632066466,
      "tokens_seen": 566508544
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.000418345035105316,
      "loss": 2.8347,
      "theoretical_loss": 3.8630605314883137,
      "tokens_seen": 566574080
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004183350050150451,
      "loss": 2.8839,
      "theoretical_loss": 3.863013437883284,
      "tokens_seen": 566639616
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041832497492477435,
      "loss": 2.713,
      "theoretical_loss": 3.862966351249538,
      "tokens_seen": 566705152
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004183149448345035,
      "loss": 2.9915,
      "theoretical_loss": 3.8629192715852376,
      "tokens_seen": 566770688
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004183049147442327,
      "loss": 2.868,
      "theoretical_loss": 3.862872198888546,
      "tokens_seen": 566836224
    },
    {
      "epoch": 1.08,
      "objective/train/docs_used": 634194,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.527719020843506,
      "objective/train/theoretical_loss": 3.8628368989374158,
      "objective/train/tokens_used": 587345376,
      "theoretical_loss": 3.8628368989374158,
      "tokens_seen": 566885376
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004182948846539619,
      "loss": 2.473,
      "theoretical_loss": 3.862825133157626,
      "tokens_seen": 566901760
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004182848545636911,
      "loss": 2.9303,
      "theoretical_loss": 3.8627780743906426,
      "tokens_seen": 566967296
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041827482447342025,
      "loss": 2.9108,
      "theoretical_loss": 3.862731022585759,
      "tokens_seen": 567032832
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041826479438314944,
      "loss": 2.8899,
      "theoretical_loss": 3.862683977741143,
      "tokens_seen": 567098368
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004182547642928786,
      "loss": 2.7323,
      "theoretical_loss": 3.8626369398549585,
      "tokens_seen": 567163904
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041824473420260785,
      "loss": 2.9033,
      "theoretical_loss": 3.862589908925374,
      "tokens_seen": 567229440
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.000418234704112337,
      "loss": 3.1807,
      "theoretical_loss": 3.8625428849505563,
      "tokens_seen": 567294976
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004182246740220662,
      "loss": 2.795,
      "theoretical_loss": 3.862495867928674,
      "tokens_seen": 567360512
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004182146439317954,
      "loss": 3.0679,
      "theoretical_loss": 3.8624488578578964,
      "tokens_seen": 567426048
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004182046138415246,
      "loss": 2.8983,
      "theoretical_loss": 3.8624018547363925,
      "tokens_seen": 567491584
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041819458375125376,
      "loss": 2.9737,
      "theoretical_loss": 3.862354858562333,
      "tokens_seen": 567557120
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041818455366098294,
      "loss": 2.8067,
      "theoretical_loss": 3.862307869333889,
      "tokens_seen": 567622656
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004181745235707121,
      "loss": 2.8586,
      "theoretical_loss": 3.8622608870492323,
      "tokens_seen": 567688192
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041816449348044136,
      "loss": 2.8927,
      "theoretical_loss": 3.862213911706535,
      "tokens_seen": 567753728
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041815446339017054,
      "loss": 2.7804,
      "theoretical_loss": 3.862166943303971,
      "tokens_seen": 567819264
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004181444332998997,
      "loss": 2.6391,
      "theoretical_loss": 3.862119981839713,
      "tokens_seen": 567884800
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004181344032096289,
      "loss": 2.6862,
      "theoretical_loss": 3.8620730273119364,
      "tokens_seen": 567950336
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004181243731193581,
      "loss": 2.8688,
      "theoretical_loss": 3.862026079718816,
      "tokens_seen": 568015872
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004181143430290873,
      "loss": 3.0498,
      "theoretical_loss": 3.8619791390585285,
      "tokens_seen": 568081408
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041810431293881644,
      "loss": 2.8153,
      "theoretical_loss": 3.8619322053292495,
      "tokens_seen": 568146944
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004180942828485457,
      "loss": 2.723,
      "theoretical_loss": 3.8618852785291566,
      "tokens_seen": 568212480
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004180842527582748,
      "loss": 2.8977,
      "theoretical_loss": 3.8618383586564278,
      "tokens_seen": 568278016
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041807422266800404,
      "loss": 2.8766,
      "theoretical_loss": 3.861791445709242,
      "tokens_seen": 568343552
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004180641925777332,
      "loss": 2.8196,
      "theoretical_loss": 3.8617445396857786,
      "tokens_seen": 568409088
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004180541624874624,
      "loss": 2.5105,
      "theoretical_loss": 3.861697640584217,
      "tokens_seen": 568474624
    },
    {
      "epoch": 1.08,
      "objective/train/docs_used": 634895,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.1014013290405273,
      "objective/train/theoretical_loss": 3.8616624707994496,
      "objective/train/tokens_used": 588983776,
      "theoretical_loss": 3.8616624707994496,
      "tokens_seen": 568523776
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004180441323971916,
      "loss": 2.6694,
      "theoretical_loss": 3.861650748402738,
      "tokens_seen": 568540160
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004180341023069208,
      "loss": 2.9858,
      "theoretical_loss": 3.861603863139524,
      "tokens_seen": 568605696
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041802407221664994,
      "loss": 2.4696,
      "theoretical_loss": 3.861556984792756,
      "tokens_seen": 568671232
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004180140421263792,
      "loss": 2.7752,
      "theoretical_loss": 3.861510113360618,
      "tokens_seen": 568736768
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004180040120361083,
      "loss": 2.6513,
      "theoretical_loss": 3.861463248841292,
      "tokens_seen": 568802304
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041799398194583754,
      "loss": 2.6531,
      "theoretical_loss": 3.861416391232963,
      "tokens_seen": 568867840
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004179839518555667,
      "loss": 2.8723,
      "theoretical_loss": 3.8613695405338158,
      "tokens_seen": 568933376
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004179739217652959,
      "loss": 2.6643,
      "theoretical_loss": 3.861322696742036,
      "tokens_seen": 568998912
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004179638916750251,
      "loss": 2.8317,
      "theoretical_loss": 3.8612758598558097,
      "tokens_seen": 569064448
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041795386158475427,
      "loss": 2.9949,
      "theoretical_loss": 3.8612290298733236,
      "tokens_seen": 569129984
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041794383149448345,
      "loss": 2.7817,
      "theoretical_loss": 3.8611822067927655,
      "tokens_seen": 569195520
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004179338014042127,
      "loss": 2.8349,
      "theoretical_loss": 3.861135390612324,
      "tokens_seen": 569261056
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004179237713139418,
      "loss": 2.6289,
      "theoretical_loss": 3.8610885813301876,
      "tokens_seen": 569326592
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041791374122367105,
      "loss": 2.9583,
      "theoretical_loss": 3.8610417789445464,
      "tokens_seen": 569392128
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041790371113340017,
      "loss": 2.8696,
      "theoretical_loss": 3.86099498345359,
      "tokens_seen": 569457664
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004178936810431294,
      "loss": 2.7377,
      "theoretical_loss": 3.86094819485551,
      "tokens_seen": 569523200
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004178836509528586,
      "loss": 2.5355,
      "theoretical_loss": 3.8609014131484978,
      "tokens_seen": 569588736
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041787362086258777,
      "loss": 2.7731,
      "theoretical_loss": 3.860854638330746,
      "tokens_seen": 569654272
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041786359077231695,
      "loss": 2.8211,
      "theoretical_loss": 3.860807870400447,
      "tokens_seen": 569719808
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004178535606820462,
      "loss": 2.9783,
      "theoretical_loss": 3.8607611093557956,
      "tokens_seen": 569785344
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004178435305917753,
      "loss": 2.6345,
      "theoretical_loss": 3.860714355194986,
      "tokens_seen": 569850880
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041783350050150455,
      "loss": 2.885,
      "theoretical_loss": 3.8606676079162128,
      "tokens_seen": 569916416
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004178234704112337,
      "loss": 3.0496,
      "theoretical_loss": 3.860620867517672,
      "tokens_seen": 569981952
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004178134403209629,
      "loss": 2.9936,
      "theoretical_loss": 3.86057413399756,
      "tokens_seen": 570047488
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004178034102306921,
      "loss": 2.5621,
      "theoretical_loss": 3.8605274073540743,
      "tokens_seen": 570113024
    },
    {
      "epoch": 1.08,
      "objective/train/docs_used": 635428,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.758199691772461,
      "objective/train/theoretical_loss": 3.860492366883161,
      "objective/train/tokens_used": 590622176,
      "theoretical_loss": 3.860492366883161,
      "tokens_seen": 570162176
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004177933801404213,
      "loss": 2.7609,
      "theoretical_loss": 3.8604806875854116,
      "tokens_seen": 570178560
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041778335005015045,
      "loss": 2.6497,
      "theoretical_loss": 3.8604339746897725,
      "tokens_seen": 570244096
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041777331995987964,
      "loss": 2.9305,
      "theoretical_loss": 3.860387268665354,
      "tokens_seen": 570309632
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004177632898696088,
      "loss": 2.5576,
      "theoretical_loss": 3.860340569510357,
      "tokens_seen": 570375168
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041775325977933805,
      "loss": 2.7111,
      "theoretical_loss": 3.8602938772229827,
      "tokens_seen": 570440704
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004177432296890672,
      "loss": 2.9168,
      "theoretical_loss": 3.8602471918014305,
      "tokens_seen": 570506240
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004177331995987964,
      "loss": 2.7029,
      "theoretical_loss": 3.8602005132439037,
      "tokens_seen": 570571776
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004177231695085256,
      "loss": 2.7222,
      "theoretical_loss": 3.8601538415486045,
      "tokens_seen": 570637312
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004177131394182548,
      "loss": 2.7994,
      "theoretical_loss": 3.8601071767137363,
      "tokens_seen": 570702848
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041770310932798396,
      "loss": 2.7614,
      "theoretical_loss": 3.8600605187375026,
      "tokens_seen": 570768384
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041769307923771314,
      "loss": 2.6008,
      "theoretical_loss": 3.860013867618109,
      "tokens_seen": 570833920
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004176830491474423,
      "loss": 2.9842,
      "theoretical_loss": 3.8599672233537596,
      "tokens_seen": 570899456
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041767301905717156,
      "loss": 2.601,
      "theoretical_loss": 3.8599205859426604,
      "tokens_seen": 570964992
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004176629889669007,
      "loss": 2.7662,
      "theoretical_loss": 3.8598739553830193,
      "tokens_seen": 571030528
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004176529588766299,
      "loss": 2.9515,
      "theoretical_loss": 3.859827331673042,
      "tokens_seen": 571096064
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041764292878635904,
      "loss": 2.8044,
      "theoretical_loss": 3.8597807148109378,
      "tokens_seen": 571161600
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004176328986960883,
      "loss": 2.6835,
      "theoretical_loss": 3.8597341047949145,
      "tokens_seen": 571227136
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041762286860581746,
      "loss": 2.8642,
      "theoretical_loss": 3.8596875016231817,
      "tokens_seen": 571292672
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041761283851554664,
      "loss": 2.9252,
      "theoretical_loss": 3.8596409052939498,
      "tokens_seen": 571358208
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004176028084252758,
      "loss": 2.2926,
      "theoretical_loss": 3.8595943158054284,
      "tokens_seen": 571423744
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.000417592778335005,
      "loss": 2.3754,
      "theoretical_loss": 3.85954773315583,
      "tokens_seen": 571489280
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004175827482447342,
      "loss": 3.0418,
      "theoretical_loss": 3.8595011573433657,
      "tokens_seen": 571554816
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004175727181544634,
      "loss": 2.8298,
      "theoretical_loss": 3.8594545883662494,
      "tokens_seen": 571620352
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041756268806419255,
      "loss": 2.6487,
      "theoretical_loss": 3.8594080262226935,
      "tokens_seen": 571685888
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004175526579739218,
      "loss": 2.6435,
      "theoretical_loss": 3.859361470910912,
      "tokens_seen": 571751424
    },
    {
      "epoch": 1.08,
      "objective/train/docs_used": 638980,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.1796209812164307,
      "objective/train/theoretical_loss": 3.859326558909354,
      "objective/train/tokens_used": 592260576,
      "theoretical_loss": 3.859326558909354,
      "tokens_seen": 571800576
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041754262788365096,
      "loss": 2.9137,
      "theoretical_loss": 3.8593149224291197,
      "tokens_seen": 571816960
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041753259779338015,
      "loss": 2.54,
      "theoretical_loss": 3.8592683807755326,
      "tokens_seen": 571882496
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004175225677031093,
      "loss": 2.74,
      "theoretical_loss": 3.8592218459483663,
      "tokens_seen": 571948032
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004175125376128385,
      "loss": 2.6262,
      "theoretical_loss": 3.8591753179458372,
      "tokens_seen": 572013568
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004175025075225677,
      "loss": 2.8701,
      "theoretical_loss": 3.859128796766163,
      "tokens_seen": 572079104
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004174924774322969,
      "loss": 2.6111,
      "theoretical_loss": 3.859082282407562,
      "tokens_seen": 572144640
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041748244734202605,
      "loss": 2.7199,
      "theoretical_loss": 3.8590357748682527,
      "tokens_seen": 572210176
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004174724172517553,
      "loss": 2.4907,
      "theoretical_loss": 3.858989274146454,
      "tokens_seen": 572275712
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004174623871614844,
      "loss": 2.8167,
      "theoretical_loss": 3.858942780240387,
      "tokens_seen": 572341248
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041745235707121365,
      "loss": 2.6303,
      "theoretical_loss": 3.858896293148272,
      "tokens_seen": 572406784
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041744232698094283,
      "loss": 3.005,
      "theoretical_loss": 3.85884981286833,
      "tokens_seen": 572472320
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.000417432296890672,
      "loss": 2.8463,
      "theoretical_loss": 3.858803339398783,
      "tokens_seen": 572537856
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004174222668004012,
      "loss": 3.0947,
      "theoretical_loss": 3.858756872737855,
      "tokens_seen": 572603392
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041741223671013037,
      "loss": 3.0426,
      "theoretical_loss": 3.8587104128837675,
      "tokens_seen": 572668928
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004174022066198596,
      "loss": 2.8641,
      "theoretical_loss": 3.8586639598347463,
      "tokens_seen": 572734464
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004173921765295888,
      "loss": 2.8205,
      "theoretical_loss": 3.8586175135890155,
      "tokens_seen": 572800000
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041738214643931797,
      "loss": 2.9435,
      "theoretical_loss": 3.8585710741448,
      "tokens_seen": 572865536
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041737211634904715,
      "loss": 2.9395,
      "theoretical_loss": 3.8585246415003267,
      "tokens_seen": 572931072
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004173620862587764,
      "loss": 2.7106,
      "theoretical_loss": 3.8584782156538218,
      "tokens_seen": 572996608
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004173520561685055,
      "loss": 2.8422,
      "theoretical_loss": 3.858431796603513,
      "tokens_seen": 573062144
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041734202607823475,
      "loss": 2.4501,
      "theoretical_loss": 3.8583853843476277,
      "tokens_seen": 573127680
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004173319959879639,
      "loss": 2.7858,
      "theoretical_loss": 3.8583389788843956,
      "tokens_seen": 573193216
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004173219658976931,
      "loss": 2.5905,
      "theoretical_loss": 3.858292580212045,
      "tokens_seen": 573258752
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004173119358074223,
      "loss": 2.659,
      "theoretical_loss": 3.8582461883288075,
      "tokens_seen": 573324288
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004173019057171515,
      "loss": 2.6405,
      "theoretical_loss": 3.858199803232913,
      "tokens_seen": 573389824
    },
    {
      "epoch": 1.08,
      "objective/train/docs_used": 641212,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7385621070861816,
      "objective/train/theoretical_loss": 3.8581650188641214,
      "objective/train/tokens_used": 593504736,
      "theoretical_loss": 3.8581650188641214,
      "tokens_seen": 573438976
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041729187562688065,
      "loss": 2.822,
      "theoretical_loss": 3.858153424922592,
      "tokens_seen": 573455360
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041728184553660984,
      "loss": 2.5345,
      "theoretical_loss": 3.858107053396078,
      "tokens_seen": 573520896
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.000417271815446339,
      "loss": 2.6229,
      "theoretical_loss": 3.858060688651603,
      "tokens_seen": 573586432
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041726178535606825,
      "loss": 2.5662,
      "theoretical_loss": 3.8580143306874,
      "tokens_seen": 573651968
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004172517552657974,
      "loss": 2.9459,
      "theoretical_loss": 3.857967979501704,
      "tokens_seen": 573717504
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004172417251755266,
      "loss": 2.9108,
      "theoretical_loss": 3.857921635092749,
      "tokens_seen": 573783040
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004172316950852558,
      "loss": 2.7796,
      "theoretical_loss": 3.8578752974587704,
      "tokens_seen": 573848576
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.000417221664994985,
      "loss": 2.9508,
      "theoretical_loss": 3.857828966598005,
      "tokens_seen": 573914112
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041721163490471416,
      "loss": 2.4994,
      "theoretical_loss": 3.857782642508688,
      "tokens_seen": 573979648
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041720160481444334,
      "loss": 2.8715,
      "theoretical_loss": 3.857736325189058,
      "tokens_seen": 574045184
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004171915747241725,
      "loss": 2.666,
      "theoretical_loss": 3.8576900146373525,
      "tokens_seen": 574110720
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041718154463390176,
      "loss": 2.6323,
      "theoretical_loss": 3.8576437108518102,
      "tokens_seen": 574176256
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004171715145436309,
      "loss": 2.7769,
      "theoretical_loss": 3.8575974138306703,
      "tokens_seen": 574241792
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004171614844533601,
      "loss": 2.8701,
      "theoretical_loss": 3.857551123572174,
      "tokens_seen": 574307328
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041715145436308924,
      "loss": 2.7098,
      "theoretical_loss": 3.8575048400745597,
      "tokens_seen": 574372864
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004171414242728185,
      "loss": 2.9271,
      "theoretical_loss": 3.8574585633360705,
      "tokens_seen": 574438400
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041713139418254766,
      "loss": 2.7082,
      "theoretical_loss": 3.8574122933549475,
      "tokens_seen": 574503936
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041712136409227684,
      "loss": 2.8392,
      "theoretical_loss": 3.857366030129434,
      "tokens_seen": 574569472
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.000417111334002006,
      "loss": 2.5627,
      "theoretical_loss": 3.857319773657772,
      "tokens_seen": 574635008
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004171013039117352,
      "loss": 2.5853,
      "theoretical_loss": 3.8572735239382068,
      "tokens_seen": 574700544
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004170912738214644,
      "loss": 2.8761,
      "theoretical_loss": 3.8572272809689823,
      "tokens_seen": 574766080
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004170812437311936,
      "loss": 2.9132,
      "theoretical_loss": 3.857181044748344,
      "tokens_seen": 574831616
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041707121364092275,
      "loss": 2.6828,
      "theoretical_loss": 3.857134815274538,
      "tokens_seen": 574897152
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.000417061183550652,
      "loss": 2.5846,
      "theoretical_loss": 3.85708859254581,
      "tokens_seen": 574962688
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041705115346038116,
      "loss": 2.9014,
      "theoretical_loss": 3.8570423765604076,
      "tokens_seen": 575028224
    },
    {
      "epoch": 1.08,
      "objective/train/docs_used": 641212,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5287535190582275,
      "objective/train/theoretical_loss": 3.8570077189956096,
      "objective/train/tokens_used": 593504736,
      "theoretical_loss": 3.8570077189956096,
      "tokens_seen": 575077376
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041704112337011035,
      "loss": 2.5591,
      "theoretical_loss": 3.856996167316579,
      "tokens_seen": 575093760
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004170310932798395,
      "loss": 2.623,
      "theoretical_loss": 3.8569499648125727,
      "tokens_seen": 575159296
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004170210631895687,
      "loss": 2.9332,
      "theoretical_loss": 3.8569037690466375,
      "tokens_seen": 575224832
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004170110330992979,
      "loss": 2.7544,
      "theoretical_loss": 3.8568575800170235,
      "tokens_seen": 575290368
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004170010030090271,
      "loss": 2.7303,
      "theoretical_loss": 3.856811397721981,
      "tokens_seen": 575355904
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041699097291875625,
      "loss": 2.5582,
      "theoretical_loss": 3.856765222159762,
      "tokens_seen": 575421440
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004169809428284855,
      "loss": 2.8197,
      "theoretical_loss": 3.856719053328616,
      "tokens_seen": 575486976
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004169709127382146,
      "loss": 2.7173,
      "theoretical_loss": 3.8566728912267982,
      "tokens_seen": 575552512
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041696088264794385,
      "loss": 2.8001,
      "theoretical_loss": 3.85662673585256,
      "tokens_seen": 575618048
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041695085255767303,
      "loss": 2.9218,
      "theoretical_loss": 3.8565805872041556,
      "tokens_seen": 575683584
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004169408224674022,
      "loss": 2.7323,
      "theoretical_loss": 3.8565344452798396,
      "tokens_seen": 575749120
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004169307923771314,
      "loss": 2.8589,
      "theoretical_loss": 3.856488310077866,
      "tokens_seen": 575814656
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004169207622868606,
      "loss": 2.8005,
      "theoretical_loss": 3.8564421815964924,
      "tokens_seen": 575880192
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041691073219658975,
      "loss": 2.7526,
      "theoretical_loss": 3.856396059833974,
      "tokens_seen": 575945728
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.000416900702106319,
      "loss": 2.6452,
      "theoretical_loss": 3.856349944788567,
      "tokens_seen": 576011264
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004168906720160481,
      "loss": 2.8848,
      "theoretical_loss": 3.8563038364585314,
      "tokens_seen": 576076800
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041688064192577735,
      "loss": 2.548,
      "theoretical_loss": 3.856257734842123,
      "tokens_seen": 576142336
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041687061183550653,
      "loss": 2.8584,
      "theoretical_loss": 3.856211639937602,
      "tokens_seen": 576207872
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004168605817452357,
      "loss": 3.0388,
      "theoretical_loss": 3.856165551743228,
      "tokens_seen": 576273408
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004168505516549649,
      "loss": 2.7249,
      "theoretical_loss": 3.8561194702572603,
      "tokens_seen": 576338944
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004168405215646941,
      "loss": 2.5879,
      "theoretical_loss": 3.856073395477962,
      "tokens_seen": 576404480
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041683049147442326,
      "loss": 3.0491,
      "theoretical_loss": 3.856027327403592,
      "tokens_seen": 576470016
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004168204613841525,
      "loss": 2.7849,
      "theoretical_loss": 3.8559812660324138,
      "tokens_seen": 576535552
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004168104312938816,
      "loss": 2.5804,
      "theoretical_loss": 3.8559352113626906,
      "tokens_seen": 576601088
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041680040120361085,
      "loss": 2.7666,
      "theoretical_loss": 3.855889163392685,
      "tokens_seen": 576666624
    },
    {
      "epoch": 1.08,
      "objective/train/docs_used": 641212,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.898465633392334,
      "objective/train/theoretical_loss": 3.8558546318108267,
      "objective/train/tokens_used": 593504736,
      "theoretical_loss": 3.8558546318108267,
      "tokens_seen": 576715776
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041679037111334,
      "loss": 2.7956,
      "theoretical_loss": 3.855843122120662,
      "tokens_seen": 576732160
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004167803410230692,
      "loss": 2.9043,
      "theoretical_loss": 3.8557970875448855,
      "tokens_seen": 576797696
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004167703109327984,
      "loss": 2.7581,
      "theoretical_loss": 3.8557510596636217,
      "tokens_seen": 576863232
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004167602808425276,
      "loss": 2.6658,
      "theoretical_loss": 3.8557050384751363,
      "tokens_seen": 576928768
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041675025075225676,
      "loss": 2.7947,
      "theoretical_loss": 3.855659023977696,
      "tokens_seen": 576994304
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.000416740220661986,
      "loss": 3.0682,
      "theoretical_loss": 3.8556130161695688,
      "tokens_seen": 577059840
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004167301905717151,
      "loss": 2.7337,
      "theoretical_loss": 3.8555670150490213,
      "tokens_seen": 577125376
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.00041672016048144436,
      "loss": 2.8836,
      "theoretical_loss": 3.855521020614324,
      "tokens_seen": 577190912
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0004167101303911735,
      "loss": 2.6244,
      "theoretical_loss": 3.8554750328637444,
      "tokens_seen": 577256448
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0004167001003009027,
      "loss": 2.6494,
      "theoretical_loss": 3.8554290517955536,
      "tokens_seen": 577321984
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0004166900702106319,
      "loss": 2.845,
      "theoretical_loss": 3.855383077408022,
      "tokens_seen": 577387520
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0004166800401203611,
      "loss": 2.8374,
      "theoretical_loss": 3.8553371096994207,
      "tokens_seen": 577453056
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.00041667001003009026,
      "loss": 2.6816,
      "theoretical_loss": 3.8552911486680217,
      "tokens_seen": 577518592
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.00041665997993981944,
      "loss": 2.7496,
      "theoretical_loss": 3.855245194312097,
      "tokens_seen": 577584128
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0004166499498495487,
      "loss": 2.6757,
      "theoretical_loss": 3.8551992466299208,
      "tokens_seen": 577649664
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.00041663991975927786,
      "loss": 2.7504,
      "theoretical_loss": 3.8551533056197664,
      "tokens_seen": 577715200
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.00041662988966900704,
      "loss": 2.8373,
      "theoretical_loss": 3.8551073712799075,
      "tokens_seen": 577780736
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0004166198595787362,
      "loss": 2.8251,
      "theoretical_loss": 3.85506144360862,
      "tokens_seen": 577846272
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0004166098294884654,
      "loss": 2.8219,
      "theoretical_loss": 3.8550155226041802,
      "tokens_seen": 577911808
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0004165997993981946,
      "loss": 2.8431,
      "theoretical_loss": 3.854969608264863,
      "tokens_seen": 577977344
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0004165897693079238,
      "loss": 2.9726,
      "theoretical_loss": 3.854923700588947,
      "tokens_seen": 578042880
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.00041657973921765295,
      "loss": 2.8628,
      "theoretical_loss": 3.8548777995747088,
      "tokens_seen": 578108416
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0004165697091273822,
      "loss": 3.1248,
      "theoretical_loss": 3.8548319052204265,
      "tokens_seen": 578173952
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.00041655967903711136,
      "loss": 2.6876,
      "theoretical_loss": 3.8547860175243795,
      "tokens_seen": 578239488
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.00041654964894684055,
      "loss": 2.6237,
      "theoretical_loss": 3.854740136484848,
      "tokens_seen": 578305024
    },
    {
      "epoch": 1.09,
      "objective/train/docs_used": 641212,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9665675163269043,
      "objective/train/theoretical_loss": 3.854705730072502,
      "objective/train/tokens_used": 593504736,
      "theoretical_loss": 3.854705730072502,
      "tokens_seen": 578354176
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0004165396188565697,
      "loss": 2.8729,
      "theoretical_loss": 3.854694262100111,
      "tokens_seen": 578370560
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0004165295887662989,
      "loss": 2.616,
      "theoretical_loss": 3.85464839436845,
      "tokens_seen": 578436096
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0004165195586760281,
      "loss": 2.9423,
      "theoretical_loss": 3.854602533288147,
      "tokens_seen": 578501632
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0004165095285857573,
      "loss": 2.4722,
      "theoretical_loss": 3.8545566788574828,
      "tokens_seen": 578567168
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.00041649949849548645,
      "loss": 2.7993,
      "theoretical_loss": 3.854510831074742,
      "tokens_seen": 578632704
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0004164894684052157,
      "loss": 2.8733,
      "theoretical_loss": 3.8544649899382053,
      "tokens_seen": 578698240
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0004164794383149448,
      "loss": 2.8087,
      "theoretical_loss": 3.85441915544616,
      "tokens_seen": 578763776
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.00041646940822467405,
      "loss": 2.9179,
      "theoretical_loss": 3.854373327596888,
      "tokens_seen": 578829312
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.00041645937813440323,
      "loss": 2.9232,
      "theoretical_loss": 3.854327506388677,
      "tokens_seen": 578894848
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0004164493480441324,
      "loss": 2.8144,
      "theoretical_loss": 3.854281691819811,
      "tokens_seen": 578960384
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0004164393179538616,
      "loss": 2.7681,
      "theoretical_loss": 3.8542358838885775,
      "tokens_seen": 579025920
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0004164292878635908,
      "loss": 2.7656,
      "theoretical_loss": 3.854190082593264,
      "tokens_seen": 579091456
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.00041641925777331995,
      "loss": 2.6538,
      "theoretical_loss": 3.854144287932158,
      "tokens_seen": 579156992
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0004164092276830492,
      "loss": 2.8346,
      "theoretical_loss": 3.8540984999035475,
      "tokens_seen": 579222528
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0004163991975927783,
      "loss": 2.5503,
      "theoretical_loss": 3.8540527185057223,
      "tokens_seen": 579288064
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.00041638916750250755,
      "loss": 2.8447,
      "theoretical_loss": 3.854006943736972,
      "tokens_seen": 579353600
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.00041637913741223673,
      "loss": 2.8478,
      "theoretical_loss": 3.853961175595587,
      "tokens_seen": 579419136
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0004163691073219659,
      "loss": 2.7447,
      "theoretical_loss": 3.8539154140798586,
      "tokens_seen": 579484672
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0004163590772316951,
      "loss": 2.7067,
      "theoretical_loss": 3.853869659188078,
      "tokens_seen": 579550208
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0004163490471414243,
      "loss": 2.8598,
      "theoretical_loss": 3.8538239109185377,
      "tokens_seen": 579615744
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.00041633901705115346,
      "loss": 2.5542,
      "theoretical_loss": 3.8537781692695305,
      "tokens_seen": 579681280
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0004163289869608827,
      "loss": 2.9101,
      "theoretical_loss": 3.8537324342393506,
      "tokens_seen": 579746816
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0004163189568706118,
      "loss": 2.7138,
      "theoretical_loss": 3.8536867058262914,
      "tokens_seen": 579812352
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.00041630892678034105,
      "loss": 2.9372,
      "theoretical_loss": 3.8536409840286483,
      "tokens_seen": 579877888
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0004162988966900702,
      "loss": 2.6821,
      "theoretical_loss": 3.8535952688447166,
      "tokens_seen": 579943424
    },
    {
      "epoch": 1.09,
      "objective/train/docs_used": 641212,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7990593910217285,
      "objective/train/theoretical_loss": 3.8535609867959906,
      "objective/train/tokens_used": 593504736,
      "theoretical_loss": 3.8535609867959906,
      "tokens_seen": 579992576
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0004162888665997994,
      "loss": 2.7535,
      "theoretical_loss": 3.853549560272792,
      "tokens_seen": 580008960
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0004162788365095286,
      "loss": 2.7862,
      "theoretical_loss": 3.8535038583111723,
      "tokens_seen": 580074496
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0004162688064192578,
      "loss": 2.8232,
      "theoretical_loss": 3.8534581629581535,
      "tokens_seen": 580140032
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.00041625877632898696,
      "loss": 2.8708,
      "theoretical_loss": 3.8534124742120346,
      "tokens_seen": 580205568
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0004162487462387162,
      "loss": 2.7073,
      "theoretical_loss": 3.853366792071114,
      "tokens_seen": 580271104
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0004162387161484453,
      "loss": 2.4495,
      "theoretical_loss": 3.8533211165336905,
      "tokens_seen": 580336640
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.00041622868605817456,
      "loss": 2.9621,
      "theoretical_loss": 3.8532754475980644,
      "tokens_seen": 580402176
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0004162186559679037,
      "loss": 2.8054,
      "theoretical_loss": 3.8532297852625366,
      "tokens_seen": 580467712
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0004162086258776329,
      "loss": 2.8751,
      "theoretical_loss": 3.8531841295254075,
      "tokens_seen": 580533248
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0004161985957873621,
      "loss": 2.7955,
      "theoretical_loss": 3.8531384803849793,
      "tokens_seen": 580598784
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0004161885656970913,
      "loss": 2.9094,
      "theoretical_loss": 3.8530928378395544,
      "tokens_seen": 580664320
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.00041617853560682046,
      "loss": 2.7033,
      "theoretical_loss": 3.8530472018874358,
      "tokens_seen": 580729856
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.00041616850551654964,
      "loss": 2.7345,
      "theoretical_loss": 3.8530015725269267,
      "tokens_seen": 580795392
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0004161584754262788,
      "loss": 2.8526,
      "theoretical_loss": 3.852955949756332,
      "tokens_seen": 580860928
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.00041614844533600806,
      "loss": 2.8981,
      "theoretical_loss": 3.8529103335739565,
      "tokens_seen": 580926464
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0004161384152457372,
      "loss": 3.1064,
      "theoretical_loss": 3.8528647239781053,
      "tokens_seen": 580992000
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0004161283851554664,
      "loss": 2.7506,
      "theoretical_loss": 3.852819120967085,
      "tokens_seen": 581057536
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.00041611835506519555,
      "loss": 2.8473,
      "theoretical_loss": 3.8527735245392023,
      "tokens_seen": 581123072
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0004161083249749248,
      "loss": 2.7136,
      "theoretical_loss": 3.852727934692765,
      "tokens_seen": 581188608
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.00041609829488465397,
      "loss": 2.5776,
      "theoretical_loss": 3.852682351426081,
      "tokens_seen": 581254144
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.00041608826479438315,
      "loss": 2.5465,
      "theoretical_loss": 3.8526367747374577,
      "tokens_seen": 581319680
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.00041607823470411233,
      "loss": 2.8932,
      "theoretical_loss": 3.852591204625206,
      "tokens_seen": 581385216
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.00041606820461384156,
      "loss": 2.8596,
      "theoretical_loss": 3.8525456410876355,
      "tokens_seen": 581450752
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0004160581745235707,
      "loss": 2.5327,
      "theoretical_loss": 3.8525000841230566,
      "tokens_seen": 581516288
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0004160481444332999,
      "loss": 2.9714,
      "theoretical_loss": 3.85245453372978,
      "tokens_seen": 581581824
    },
    {
      "epoch": 1.09,
      "objective/train/docs_used": 641212,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.1994011402130127,
      "objective/train/theoretical_loss": 3.8524203752462247,
      "objective/train/tokens_used": 593504736,
      "theoretical_loss": 3.8524203752462247,
      "tokens_seen": 581630976
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.00041603811434302905,
      "loss": 2.7316,
      "theoretical_loss": 3.8524089899061185,
      "tokens_seen": 581647360
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0004160280842527583,
      "loss": 2.8617,
      "theoretical_loss": 3.8523634526503834,
      "tokens_seen": 581712896
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.00041601805416248747,
      "loss": 2.7499,
      "theoretical_loss": 3.8523179219608883,
      "tokens_seen": 581778432
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.00041600802407221665,
      "loss": 2.5885,
      "theoretical_loss": 3.8522723978359474,
      "tokens_seen": 581843968
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.00041599799398194583,
      "loss": 2.7726,
      "theoretical_loss": 3.852226880273874,
      "tokens_seen": 581909504
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.000415987963891675,
      "loss": 2.7785,
      "theoretical_loss": 3.8521813692729836,
      "tokens_seen": 581975040
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0004159779338014042,
      "loss": 2.6607,
      "theoretical_loss": 3.852135864831591,
      "tokens_seen": 582040576
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.00041596790371113343,
      "loss": 2.622,
      "theoretical_loss": 3.8520903669480138,
      "tokens_seen": 582106112
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.00041595787362086256,
      "loss": 2.7967,
      "theoretical_loss": 3.8520448756205674,
      "tokens_seen": 582171648
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0004159478435305918,
      "loss": 2.9222,
      "theoretical_loss": 3.85199939084757,
      "tokens_seen": 582237184
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0004159378134403209,
      "loss": 2.6956,
      "theoretical_loss": 3.8519539126273394,
      "tokens_seen": 582302720
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.00041592778335005015,
      "loss": 2.7274,
      "theoretical_loss": 3.8519084409581943,
      "tokens_seen": 582368256
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.00041591775325977934,
      "loss": 2.6459,
      "theoretical_loss": 3.8518629758384537,
      "tokens_seen": 582433792
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0004159077231695085,
      "loss": 2.7572,
      "theoretical_loss": 3.8518175172664377,
      "tokens_seen": 582499328
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.00041589769307923775,
      "loss": 2.7572,
      "theoretical_loss": 3.851772065240467,
      "tokens_seen": 582564864
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.00041588766298896693,
      "loss": 2.8565,
      "theoretical_loss": 3.851726619758862,
      "tokens_seen": 582630400
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0004158776328986961,
      "loss": 2.7034,
      "theoretical_loss": 3.851681180819945,
      "tokens_seen": 582695936
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0004158676028084253,
      "loss": 2.7775,
      "theoretical_loss": 3.851635748422039,
      "tokens_seen": 582761472
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0004158575727181545,
      "loss": 2.9109,
      "theoretical_loss": 3.8515903225634656,
      "tokens_seen": 582827008
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.00041584754262788366,
      "loss": 2.7499,
      "theoretical_loss": 3.851544903242549,
      "tokens_seen": 582892544
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0004158375125376129,
      "loss": 2.854,
      "theoretical_loss": 3.8514994904576136,
      "tokens_seen": 582958080
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.000415827482447342,
      "loss": 3.1417,
      "theoretical_loss": 3.851454084206985,
      "tokens_seen": 583023616
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.00041581745235707126,
      "loss": 2.7627,
      "theoretical_loss": 3.8514086844889865,
      "tokens_seen": 583089152
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0004158074222668004,
      "loss": 2.6031,
      "theoretical_loss": 3.851363291301946,
      "tokens_seen": 583154688
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0004157973921765296,
      "loss": 2.9589,
      "theoretical_loss": 3.85131790464419,
      "tokens_seen": 583220224
    },
    {
      "epoch": 1.09,
      "objective/train/docs_used": 641212,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.419975757598877,
      "objective/train/theoretical_loss": 3.8512838689347095,
      "objective/train/tokens_used": 593504736,
      "theoretical_loss": 3.8512838689347095,
      "tokens_seen": 583269376
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0004157873620862588,
      "loss": 2.8707,
      "theoretical_loss": 3.8512725245140453,
      "tokens_seen": 583285760
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.000415777331995988,
      "loss": 2.8639,
      "theoretical_loss": 3.85122715090984,
      "tokens_seen": 583351296
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.00041576730190571716,
      "loss": 2.7196,
      "theoretical_loss": 3.8511817838299023,
      "tokens_seen": 583416832
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0004157572718154464,
      "loss": 2.828,
      "theoretical_loss": 3.8511364232725622,
      "tokens_seen": 583482368
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0004157472417251755,
      "loss": 2.8964,
      "theoretical_loss": 3.8510910692361486,
      "tokens_seen": 583547904
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.00041573721163490476,
      "loss": 2.6145,
      "theoretical_loss": 3.851045721718992,
      "tokens_seen": 583613440
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0004157271815446339,
      "loss": 2.6236,
      "theoretical_loss": 3.851000380719424,
      "tokens_seen": 583678976
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0004157171514543631,
      "loss": 2.8353,
      "theoretical_loss": 3.850955046235776,
      "tokens_seen": 583744512
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0004157071213640923,
      "loss": 2.9949,
      "theoretical_loss": 3.85090971826638,
      "tokens_seen": 583810048
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0004156970912738215,
      "loss": 2.61,
      "theoretical_loss": 3.850864396809569,
      "tokens_seen": 583875584
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.00041568706118355066,
      "loss": 2.7655,
      "theoretical_loss": 3.8508190818636763,
      "tokens_seen": 583941120
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.00041567703109327984,
      "loss": 2.8568,
      "theoretical_loss": 3.8507737734270355,
      "tokens_seen": 584006656
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.000415667001003009,
      "loss": 2.6818,
      "theoretical_loss": 3.850728471497982,
      "tokens_seen": 584072192
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.00041565697091273826,
      "loss": 2.8308,
      "theoretical_loss": 3.8506831760748517,
      "tokens_seen": 584137728
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0004156469408224674,
      "loss": 2.9406,
      "theoretical_loss": 3.8506378871559788,
      "tokens_seen": 584203264
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0004156369107321966,
      "loss": 2.736,
      "theoretical_loss": 3.850592604739701,
      "tokens_seen": 584268800
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.00041562688064192575,
      "loss": 2.8512,
      "theoretical_loss": 3.850547328824356,
      "tokens_seen": 584334336
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.000415616850551655,
      "loss": 2.5336,
      "theoretical_loss": 3.8505020594082797,
      "tokens_seen": 584399872
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.00041560682046138417,
      "loss": 2.9136,
      "theoretical_loss": 3.850456796489812,
      "tokens_seen": 584465408
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.00041559679037111335,
      "loss": 2.5841,
      "theoretical_loss": 3.8504115400672916,
      "tokens_seen": 584530944
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.00041558676028084253,
      "loss": 2.7805,
      "theoretical_loss": 3.850366290139057,
      "tokens_seen": 584596480
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.00041557673019057176,
      "loss": 2.685,
      "theoretical_loss": 3.85032104670345,
      "tokens_seen": 584662016
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0004155667001003009,
      "loss": 2.9191,
      "theoretical_loss": 3.85027580975881,
      "tokens_seen": 584727552
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0004155566700100301,
      "loss": 3.0622,
      "theoretical_loss": 3.8502305793034797,
      "tokens_seen": 584793088
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.00041554663991975925,
      "loss": 2.9408,
      "theoretical_loss": 3.8501853553357996,
      "tokens_seen": 584858624
    },
    {
      "epoch": 1.09,
      "objective/train/docs_used": 641212,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.363860607147217,
      "objective/train/theoretical_loss": 3.850151441616564,
      "objective/train/tokens_used": 593504736,
      "theoretical_loss": 3.850151441616564,
      "tokens_seen": 584907776
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0004155366098294885,
      "loss": 2.6347,
      "theoretical_loss": 3.850140137854114,
      "tokens_seen": 584924160
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.00041552657973921767,
      "loss": 2.8598,
      "theoretical_loss": 3.850094926856765,
      "tokens_seen": 584989696
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.00041551654964894685,
      "loss": 2.8034,
      "theoretical_loss": 3.8500497223420966,
      "tokens_seen": 585055232
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.00041550651955867603,
      "loss": 2.6854,
      "theoretical_loss": 3.8500045243084537,
      "tokens_seen": 585120768
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0004154964894684052,
      "loss": 2.7721,
      "theoretical_loss": 3.849959332754181,
      "tokens_seen": 585186304
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0004154864593781344,
      "loss": 3.0961,
      "theoretical_loss": 3.849914147677624,
      "tokens_seen": 585251840
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.00041547642928786363,
      "loss": 2.9701,
      "theoretical_loss": 3.849868969077129,
      "tokens_seen": 585317376
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.00041546639919759276,
      "loss": 2.7004,
      "theoretical_loss": 3.8498237969510436,
      "tokens_seen": 585382912
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.000415456369107322,
      "loss": 2.7834,
      "theoretical_loss": 3.849778631297715,
      "tokens_seen": 585448448
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0004154463390170511,
      "loss": 2.7596,
      "theoretical_loss": 3.8497334721154903,
      "tokens_seen": 585513984
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.00041543630892678035,
      "loss": 2.7112,
      "theoretical_loss": 3.84968831940272,
      "tokens_seen": 585579520
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.00041542627883650954,
      "loss": 2.9759,
      "theoretical_loss": 3.8496431731577516,
      "tokens_seen": 585645056
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0004154162487462387,
      "loss": 2.7964,
      "theoretical_loss": 3.849598033378936,
      "tokens_seen": 585710592
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0004154062186559679,
      "loss": 2.8015,
      "theoretical_loss": 3.849552900064624,
      "tokens_seen": 585776128
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.00041539618856569713,
      "loss": 3.0477,
      "theoretical_loss": 3.8495077732131655,
      "tokens_seen": 585841664
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.00041538615847542626,
      "loss": 2.8499,
      "theoretical_loss": 3.8494626528229134,
      "tokens_seen": 585907200
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0004153761283851555,
      "loss": 2.6039,
      "theoretical_loss": 3.8494175388922196,
      "tokens_seen": 585972736
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0004153660982948846,
      "loss": 2.5178,
      "theoretical_loss": 3.849372431419437,
      "tokens_seen": 586038272
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.00041535606820461386,
      "loss": 2.827,
      "theoretical_loss": 3.8493273304029194,
      "tokens_seen": 586103808
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.00041534603811434304,
      "loss": 2.8356,
      "theoretical_loss": 3.8492822358410206,
      "tokens_seen": 586169344
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0004153360080240722,
      "loss": 2.7122,
      "theoretical_loss": 3.8492371477320955,
      "tokens_seen": 586234880
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0004153259779338014,
      "loss": 2.8285,
      "theoretical_loss": 3.8491920660744996,
      "tokens_seen": 586300416
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0004153159478435306,
      "loss": 2.9381,
      "theoretical_loss": 3.849146990866589,
      "tokens_seen": 586365952
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.00041530591775325976,
      "loss": 2.871,
      "theoretical_loss": 3.8491019221067195,
      "tokens_seen": 586431488
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.000415295887662989,
      "loss": 2.9606,
      "theoretical_loss": 3.849056859793249,
      "tokens_seen": 586497024
    },
    {
      "epoch": 1.09,
      "objective/train/docs_used": 641212,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.706367015838623,
      "objective/train/theoretical_loss": 3.8490230672876073,
      "objective/train/tokens_used": 593504736,
      "theoretical_loss": 3.8490230672876073,
      "tokens_seen": 586546176
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0004152858575727181,
      "loss": 2.7225,
      "theoretical_loss": 3.8490118039245353,
      "tokens_seen": 586562560
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.00041527582748244736,
      "loss": 2.711,
      "theoretical_loss": 3.848966754498936,
      "tokens_seen": 586628096
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0004152657973921765,
      "loss": 2.9355,
      "theoretical_loss": 3.848921711514811,
      "tokens_seen": 586693632
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0004152557673019057,
      "loss": 2.9359,
      "theoretical_loss": 3.848876674970519,
      "tokens_seen": 586759168
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0004152457372116349,
      "loss": 2.8788,
      "theoretical_loss": 3.848831644864421,
      "tokens_seen": 586824704
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0004152357071213641,
      "loss": 2.892,
      "theoretical_loss": 3.8487866211948774,
      "tokens_seen": 586890240
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.00041522567703109327,
      "loss": 2.8509,
      "theoretical_loss": 3.8487416039602493,
      "tokens_seen": 586955776
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0004152156469408225,
      "loss": 2.6335,
      "theoretical_loss": 3.8486965931588992,
      "tokens_seen": 587021312
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.00041520561685055163,
      "loss": 2.764,
      "theoretical_loss": 3.8486515887891892,
      "tokens_seen": 587086848
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.00041519558676028086,
      "loss": 2.7661,
      "theoretical_loss": 3.8486065908494824,
      "tokens_seen": 587152384
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.00041518555667001,
      "loss": 2.8219,
      "theoretical_loss": 3.8485615993381432,
      "tokens_seen": 587217920
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0004151755265797392,
      "loss": 2.6805,
      "theoretical_loss": 3.8485166142535356,
      "tokens_seen": 587283456
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0004151654964894684,
      "loss": 2.7319,
      "theoretical_loss": 3.8484716355940245,
      "tokens_seen": 587348992
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0004151554663991976,
      "loss": 2.8708,
      "theoretical_loss": 3.8484266633579756,
      "tokens_seen": 587414528
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0004151454363089268,
      "loss": 2.8686,
      "theoretical_loss": 3.848381697543755,
      "tokens_seen": 587480064
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.00041513540621865595,
      "loss": 2.865,
      "theoretical_loss": 3.8483367381497295,
      "tokens_seen": 587545600
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0004151253761283852,
      "loss": 2.8516,
      "theoretical_loss": 3.8482917851742666,
      "tokens_seen": 587611136
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.00041511534603811437,
      "loss": 2.8969,
      "theoretical_loss": 3.848246838615734,
      "tokens_seen": 587676672
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.00041510531594784355,
      "loss": 2.7833,
      "theoretical_loss": 3.8482018984724995,
      "tokens_seen": 587742208
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.00041509528585757273,
      "loss": 2.8398,
      "theoretical_loss": 3.8481569647429343,
      "tokens_seen": 587807744
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.00041508525576730196,
      "loss": 2.6893,
      "theoretical_loss": 3.8481120374254063,
      "tokens_seen": 587873280
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0004150752256770311,
      "loss": 2.9593,
      "theoretical_loss": 3.848067116518287,
      "tokens_seen": 587938816
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0004150651955867603,
      "loss": 2.9808,
      "theoretical_loss": 3.848022202019946,
      "tokens_seen": 588004352
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.00041505516549648945,
      "loss": 2.7058,
      "theoretical_loss": 3.847977293928756,
      "tokens_seen": 588069888
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0004150451354062187,
      "loss": 2.6618,
      "theoretical_loss": 3.8479323922430893,
      "tokens_seen": 588135424
    },
    {
      "epoch": 1.09,
      "objective/train/docs_used": 641212,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.562694787979126,
      "objective/train/theoretical_loss": 3.847898720181484,
      "objective/train/tokens_used": 593504736,
      "theoretical_loss": 3.847898720181484,
      "tokens_seen": 588184576
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.00041503510531594787,
      "loss": 2.3911,
      "theoretical_loss": 3.8478874969613175,
      "tokens_seen": 588200960
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.00041502507522567705,
      "loss": 2.5121,
      "theoretical_loss": 3.847842608081815,
      "tokens_seen": 588266496
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.00041501504513540623,
      "loss": 2.7847,
      "theoretical_loss": 3.8477977256029554,
      "tokens_seen": 588332032
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0004150050150451354,
      "loss": 2.5692,
      "theoretical_loss": 3.847752849523112,
      "tokens_seen": 588397568
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0004149949849548646,
      "loss": 2.977,
      "theoretical_loss": 3.847707979840662,
      "tokens_seen": 588463104
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.00041498495486459383,
      "loss": 2.8128,
      "theoretical_loss": 3.8476631165539796,
      "tokens_seen": 588528640
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.00041497492477432296,
      "loss": 2.7995,
      "theoretical_loss": 3.8476182596614414,
      "tokens_seen": 588594176
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0004149648946840522,
      "loss": 2.771,
      "theoretical_loss": 3.8475734091614253,
      "tokens_seen": 588659712
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0004149548645937813,
      "loss": 2.8067,
      "theoretical_loss": 3.847528565052307,
      "tokens_seen": 588725248
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.00041494483450351055,
      "loss": 2.6756,
      "theoretical_loss": 3.847483727332466,
      "tokens_seen": 588790784
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.00041493480441323974,
      "loss": 2.6244,
      "theoretical_loss": 3.8474388960002797,
      "tokens_seen": 588856320
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0004149247743229689,
      "loss": 2.6471,
      "theoretical_loss": 3.8473940710541283,
      "tokens_seen": 588921856
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0004149147442326981,
      "loss": 2.4718,
      "theoretical_loss": 3.847349252492392,
      "tokens_seen": 588987392
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.00041490471414242733,
      "loss": 2.9592,
      "theoretical_loss": 3.84730444031345,
      "tokens_seen": 589052928
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.00041489468405215646,
      "loss": 2.5508,
      "theoretical_loss": 3.847259634515684,
      "tokens_seen": 589118464
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0004148846539618857,
      "loss": 3.1104,
      "theoretical_loss": 3.847214835097476,
      "tokens_seen": 589184000
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0004148746238716148,
      "loss": 2.821,
      "theoretical_loss": 3.8471700420572077,
      "tokens_seen": 589249536
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.00041486459378134406,
      "loss": 2.8675,
      "theoretical_loss": 3.8471252553932618,
      "tokens_seen": 589315072
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.00041485456369107324,
      "loss": 2.7168,
      "theoretical_loss": 3.847080475104022,
      "tokens_seen": 589380608
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0004148445336008024,
      "loss": 2.5669,
      "theoretical_loss": 3.847035701187872,
      "tokens_seen": 589446144
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0004148345035105316,
      "loss": 2.7023,
      "theoretical_loss": 3.8469909336431964,
      "tokens_seen": 589511680
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0004148244734202608,
      "loss": 2.8024,
      "theoretical_loss": 3.846946172468381,
      "tokens_seen": 589577216
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.00041481444332998996,
      "loss": 2.839,
      "theoretical_loss": 3.846901417661811,
      "tokens_seen": 589642752
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0004148044132397192,
      "loss": 2.5471,
      "theoretical_loss": 3.846856669221872,
      "tokens_seen": 589708288
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0004147943831494483,
      "loss": 2.6021,
      "theoretical_loss": 3.846811927146952,
      "tokens_seen": 589773824
    },
    {
      "epoch": 1.09,
      "objective/train/docs_used": 641212,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.017899990081787,
      "objective/train/theoretical_loss": 3.846778374766836,
      "objective/train/tokens_used": 593504736,
      "theoretical_loss": 3.846778374766836,
      "tokens_seen": 589822976
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.00041478435305917756,
      "loss": 2.9278,
      "theoretical_loss": 3.846767191435438,
      "tokens_seen": 589839360
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0004147743229689067,
      "loss": 2.6644,
      "theoretical_loss": 3.8467224620857183,
      "tokens_seen": 589904896
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0004147642928786359,
      "loss": 2.7847,
      "theoretical_loss": 3.846677739096182,
      "tokens_seen": 589970432
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0004147542627883651,
      "loss": 2.4443,
      "theoretical_loss": 3.846633022465218,
      "tokens_seen": 590035968
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0004147442326980943,
      "loss": 2.6099,
      "theoretical_loss": 3.846588312191215,
      "tokens_seen": 590101504
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.00041473420260782347,
      "loss": 2.8192,
      "theoretical_loss": 3.8465436082725653,
      "tokens_seen": 590167040
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0004147241725175527,
      "loss": 2.6361,
      "theoretical_loss": 3.846498910707659,
      "tokens_seen": 590232576
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.00041471414242728183,
      "loss": 2.7931,
      "theoretical_loss": 3.8464542194948876,
      "tokens_seen": 590298112
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.00041470411233701106,
      "loss": 2.8275,
      "theoretical_loss": 3.8464095346326435,
      "tokens_seen": 590363648
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0004146940822467402,
      "loss": 2.5265,
      "theoretical_loss": 3.8463648561193198,
      "tokens_seen": 590429184
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0004146840521564694,
      "loss": 2.4411,
      "theoretical_loss": 3.84632018395331,
      "tokens_seen": 590494720
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0004146740220661986,
      "loss": 2.6881,
      "theoretical_loss": 3.846275518133007,
      "tokens_seen": 590560256
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0004146639919759278,
      "loss": 2.9586,
      "theoretical_loss": 3.846230858656806,
      "tokens_seen": 590625792
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.00041465396188565697,
      "loss": 2.8721,
      "theoretical_loss": 3.846186205523102,
      "tokens_seen": 590691328
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.00041464393179538615,
      "loss": 2.5559,
      "theoretical_loss": 3.846141558730291,
      "tokens_seen": 590756864
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.00041463390170511533,
      "loss": 2.877,
      "theoretical_loss": 3.8460969182767695,
      "tokens_seen": 590822400
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.00041462387161484457,
      "loss": 2.8501,
      "theoretical_loss": 3.8460522841609333,
      "tokens_seen": 590887936
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0004146138415245737,
      "loss": 2.7009,
      "theoretical_loss": 3.8460076563811807,
      "tokens_seen": 590953472
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.00041460381143430293,
      "loss": 2.6128,
      "theoretical_loss": 3.8459630349359104,
      "tokens_seen": 591019008
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0004145937813440321,
      "loss": 2.4929,
      "theoretical_loss": 3.8459184198235192,
      "tokens_seen": 591084544
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0004145837512537613,
      "loss": 2.8232,
      "theoretical_loss": 3.8458738110424076,
      "tokens_seen": 591150080
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0004145737211634905,
      "loss": 2.6952,
      "theoretical_loss": 3.845829208590975,
      "tokens_seen": 591215616
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.00041456369107321965,
      "loss": 2.8109,
      "theoretical_loss": 3.8457846124676225,
      "tokens_seen": 591281152
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.00041455366098294884,
      "loss": 3.0057,
      "theoretical_loss": 3.8457400226707494,
      "tokens_seen": 591346688
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.00041454363089267807,
      "loss": 2.8575,
      "theoretical_loss": 3.8456954391987592,
      "tokens_seen": 591412224
    },
    {
      "epoch": 1.09,
      "objective/train/docs_used": 641212,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5051016807556152,
      "objective/train/theoretical_loss": 3.845662005744509,
      "objective/train/tokens_used": 593504736,
      "theoretical_loss": 3.845662005744509,
      "tokens_seen": 591461376
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0004145336008024072,
      "loss": 2.763,
      "theoretical_loss": 3.8456508620500527,
      "tokens_seen": 591477760
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.00041452357071213643,
      "loss": 2.7731,
      "theoretical_loss": 3.8456062912230324,
      "tokens_seen": 591543296
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.00041451354062186556,
      "loss": 2.8782,
      "theoretical_loss": 3.8455617267161024,
      "tokens_seen": 591608832
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0004145035105315948,
      "loss": 2.9828,
      "theoretical_loss": 3.8455171685276666,
      "tokens_seen": 591674368
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.000414493480441324,
      "loss": 2.6916,
      "theoretical_loss": 3.845472616656129,
      "tokens_seen": 591739904
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.00041448345035105316,
      "loss": 2.3959,
      "theoretical_loss": 3.845428071099895,
      "tokens_seen": 591805440
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.00041447342026078234,
      "loss": 2.7328,
      "theoretical_loss": 3.8453835318573697,
      "tokens_seen": 591870976
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0004144633901705115,
      "loss": 2.8305,
      "theoretical_loss": 3.8453389989269593,
      "tokens_seen": 591936512
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0004144533600802407,
      "loss": 2.7196,
      "theoretical_loss": 3.8452944723070708,
      "tokens_seen": 592002048
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.00041444332998996994,
      "loss": 2.6286,
      "theoretical_loss": 3.845249951996111,
      "tokens_seen": 592067584
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.00041443329989969906,
      "loss": 2.4818,
      "theoretical_loss": 3.8452054379924894,
      "tokens_seen": 592133120
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0004144232698094283,
      "loss": 2.6504,
      "theoretical_loss": 3.845160930294613,
      "tokens_seen": 592198656
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.00041441323971915753,
      "loss": 2.715,
      "theoretical_loss": 3.845116428900891,
      "tokens_seen": 592264192
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.00041440320962888666,
      "loss": 2.7967,
      "theoretical_loss": 3.8450719338097326,
      "tokens_seen": 592329728
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0004143931795386159,
      "loss": 3.0511,
      "theoretical_loss": 3.8450274450195496,
      "tokens_seen": 592395264
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.000414383149448345,
      "loss": 2.8589,
      "theoretical_loss": 3.8449829625287517,
      "tokens_seen": 592460800
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.00041437311935807426,
      "loss": 2.8515,
      "theoretical_loss": 3.8449384863357503,
      "tokens_seen": 592526336
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.00041436308926780344,
      "loss": 3.0156,
      "theoretical_loss": 3.8448940164389573,
      "tokens_seen": 592591872
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0004143530591775326,
      "loss": 2.5779,
      "theoretical_loss": 3.844849552836786,
      "tokens_seen": 592657408
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0004143430290872618,
      "loss": 2.5526,
      "theoretical_loss": 3.844805095527648,
      "tokens_seen": 592722944
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.000414332998996991,
      "loss": 2.5768,
      "theoretical_loss": 3.844760644509959,
      "tokens_seen": 592788480
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.00041432296890672016,
      "loss": 2.8752,
      "theoretical_loss": 3.8447161997821313,
      "tokens_seen": 592854016
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0004143129388164494,
      "loss": 2.502,
      "theoretical_loss": 3.844671761342581,
      "tokens_seen": 592919552
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0004143029087261785,
      "loss": 2.7409,
      "theoretical_loss": 3.8446273291897226,
      "tokens_seen": 592985088
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.00041429287863590776,
      "loss": 2.8724,
      "theoretical_loss": 3.844582903321973,
      "tokens_seen": 593050624
    },
    {
      "epoch": 1.09,
      "objective/train/docs_used": 641212,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6726303100585938,
      "objective/train/theoretical_loss": 3.8445495880448104,
      "objective/train/tokens_used": 593504736,
      "theoretical_loss": 3.8445495880448104,
      "tokens_seen": 593099776
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0004142828485456369,
      "loss": 2.8083,
      "theoretical_loss": 3.844538483737748,
      "tokens_seen": 593116160
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0004142728184553661,
      "loss": 3.0371,
      "theoretical_loss": 3.8444940704354655,
      "tokens_seen": 593181696
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0004142627883650953,
      "loss": 2.7278,
      "theoretical_loss": 3.844449663413542,
      "tokens_seen": 593247232
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0004142527582748245,
      "loss": 2.7245,
      "theoretical_loss": 3.844405262670397,
      "tokens_seen": 593312768
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.00041424272818455367,
      "loss": 2.728,
      "theoretical_loss": 3.8443608682044488,
      "tokens_seen": 593378304
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0004142326980942829,
      "loss": 2.7423,
      "theoretical_loss": 3.8443164800141165,
      "tokens_seen": 593443840
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.00041422266800401203,
      "loss": 2.8193,
      "theoretical_loss": 3.8442720980978207,
      "tokens_seen": 593509376
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.00041421263791374126,
      "loss": 3.4325,
      "theoretical_loss": 3.844217322819618,
      "tokens_seen": 593590272
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.0004142026078234704,
      "loss": 2.9414,
      "theoretical_loss": 3.844172954916165,
      "tokens_seen": 593655808
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.0004141925777331996,
      "loss": 2.9461,
      "theoretical_loss": 3.844128593281643,
      "tokens_seen": 593721344
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.0004141825476429288,
      "loss": 2.8163,
      "theoretical_loss": 3.8440842379144717,
      "tokens_seen": 593786880
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.000414172517552658,
      "loss": 3.198,
      "theoretical_loss": 3.844039888813076,
      "tokens_seen": 593852416
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.00041416248746238717,
      "loss": 3.0478,
      "theoretical_loss": 3.8439955459758792,
      "tokens_seen": 593917952
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.00041415245737211635,
      "loss": 2.9957,
      "theoretical_loss": 3.8439512094013057,
      "tokens_seen": 593983488
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.00041414242728184553,
      "loss": 3.0462,
      "theoretical_loss": 3.843906879087779,
      "tokens_seen": 594049024
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.00041413239719157477,
      "loss": 2.8059,
      "theoretical_loss": 3.8438625550337253,
      "tokens_seen": 594114560
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.0004141223671013039,
      "loss": 2.8339,
      "theoretical_loss": 3.84381823723757,
      "tokens_seen": 594180096
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.00041411233701103313,
      "loss": 3.0042,
      "theoretical_loss": 3.8437739256977403,
      "tokens_seen": 594245632
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.0004141023069207623,
      "loss": 2.868,
      "theoretical_loss": 3.8437296204126623,
      "tokens_seen": 594311168
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.0004140922768304915,
      "loss": 3.0221,
      "theoretical_loss": 3.843685321380763,
      "tokens_seen": 594376704
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.0004140822467402207,
      "loss": 2.856,
      "theoretical_loss": 3.8436410286004725,
      "tokens_seen": 594442240
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.00041407221664994985,
      "loss": 2.9139,
      "theoretical_loss": 3.8435967420702175,
      "tokens_seen": 594507776
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.00041406218655967904,
      "loss": 2.8862,
      "theoretical_loss": 3.8435524617884287,
      "tokens_seen": 594573312
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.00041405215646940827,
      "loss": 2.977,
      "theoretical_loss": 3.8435081877535344,
      "tokens_seen": 594638848
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.0004140421263791374,
      "loss": 2.78,
      "theoretical_loss": 3.843463919963966,
      "tokens_seen": 594704384
    },
    {
      "epoch": 2.0,
      "objective/train/docs_used": 708692,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.850222587585449,
      "objective/train/theoretical_loss": 3.843441788410689,
      "objective/train/tokens_used": 615197152,
      "theoretical_loss": 3.843441788410689,
      "tokens_seen": 594737152
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.00041403209628886663,
      "loss": 2.9877,
      "theoretical_loss": 3.8434196584181546,
      "tokens_seen": 594769920
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.00041402206619859576,
      "loss": 2.9706,
      "theoretical_loss": 3.8433754031145315,
      "tokens_seen": 594835456
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.000414012036108325,
      "loss": 2.9933,
      "theoretical_loss": 3.8433311540515285,
      "tokens_seen": 594900992
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.0004140020060180542,
      "loss": 3.0618,
      "theoretical_loss": 3.8432869112275787,
      "tokens_seen": 594966528
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.00041399197592778336,
      "loss": 3.0248,
      "theoretical_loss": 3.8432426746411146,
      "tokens_seen": 595032064
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.00041398194583751254,
      "loss": 2.9282,
      "theoretical_loss": 3.8431984442905707,
      "tokens_seen": 595097600
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.0004139719157472417,
      "loss": 2.9484,
      "theoretical_loss": 3.8431542201743807,
      "tokens_seen": 595163136
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.0004139618856569709,
      "loss": 2.8406,
      "theoretical_loss": 3.8431100022909797,
      "tokens_seen": 595228672
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.00041395185556670014,
      "loss": 2.7116,
      "theoretical_loss": 3.8430657906388035,
      "tokens_seen": 595294208
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.00041394182547642926,
      "loss": 2.8288,
      "theoretical_loss": 3.843021585216288,
      "tokens_seen": 595359744
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.0004139317953861585,
      "loss": 2.886,
      "theoretical_loss": 3.84297738602187,
      "tokens_seen": 595425280
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.0004139217652958877,
      "loss": 2.9879,
      "theoretical_loss": 3.8429331930539865,
      "tokens_seen": 595490816
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.00041391173520561686,
      "loss": 2.8221,
      "theoretical_loss": 3.8428890063110748,
      "tokens_seen": 595556352
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.00041390170511534604,
      "loss": 2.8475,
      "theoretical_loss": 3.8428448257915733,
      "tokens_seen": 595621888
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.0004138916750250752,
      "loss": 3.1417,
      "theoretical_loss": 3.8428006514939215,
      "tokens_seen": 595687424
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.0004138816449348044,
      "loss": 2.7316,
      "theoretical_loss": 3.8427564834165584,
      "tokens_seen": 595752960
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.00041387161484453364,
      "loss": 2.8183,
      "theoretical_loss": 3.842712321557924,
      "tokens_seen": 595818496
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.00041386158475426277,
      "loss": 2.9654,
      "theoretical_loss": 3.842668165916459,
      "tokens_seen": 595884032
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.000413851554663992,
      "loss": 2.7973,
      "theoretical_loss": 3.8426240164906043,
      "tokens_seen": 595949568
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.00041384152457372113,
      "loss": 3.0307,
      "theoretical_loss": 3.842579873278801,
      "tokens_seen": 596015104
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.00041383149448345036,
      "loss": 2.7348,
      "theoretical_loss": 3.842535736279493,
      "tokens_seen": 596080640
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.00041382146439317954,
      "loss": 2.9633,
      "theoretical_loss": 3.842491605491122,
      "tokens_seen": 596146176
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.0004138114343029087,
      "loss": 2.9599,
      "theoretical_loss": 3.842447480912131,
      "tokens_seen": 596211712
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.0004138014042126379,
      "loss": 2.7724,
      "theoretical_loss": 3.8424033625409644,
      "tokens_seen": 596277248
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.0004137913741223671,
      "loss": 3.0336,
      "theoretical_loss": 3.8423592503760666,
      "tokens_seen": 596342784
    },
    {
      "epoch": 2.0,
      "objective/train/docs_used": 713945,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.0268349647521973,
      "objective/train/theoretical_loss": 3.8423371966204827,
      "objective/train/tokens_used": 616835552,
      "theoretical_loss": 3.8423371966204827,
      "tokens_seen": 596375552
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.00041378134403209627,
      "loss": 3.0419,
      "theoretical_loss": 3.842315144415883,
      "tokens_seen": 596408320
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.0004137713139418255,
      "loss": 2.9421,
      "theoretical_loss": 3.842271044658859,
      "tokens_seen": 596473856
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.00041376128385155463,
      "loss": 2.99,
      "theoretical_loss": 3.84222695110344,
      "tokens_seen": 596539392
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.00041375125376128387,
      "loss": 2.9369,
      "theoretical_loss": 3.842182863748074,
      "tokens_seen": 596604928
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.00041374122367101305,
      "loss": 3.0614,
      "theoretical_loss": 3.8421387825912072,
      "tokens_seen": 596670464
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.00041373119358074223,
      "loss": 3.1159,
      "theoretical_loss": 3.842094707631288,
      "tokens_seen": 596736000
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.0004137211634904714,
      "loss": 2.7884,
      "theoretical_loss": 3.8420506388667652,
      "tokens_seen": 596801536
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.0004137111334002006,
      "loss": 2.977,
      "theoretical_loss": 3.842006576296087,
      "tokens_seen": 596867072
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.00041370110330992977,
      "loss": 2.9942,
      "theoretical_loss": 3.841962519917703,
      "tokens_seen": 596932608
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.000413691073219659,
      "loss": 2.86,
      "theoretical_loss": 3.8419184697300635,
      "tokens_seen": 596998144
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.00041368104312938813,
      "loss": 2.8966,
      "theoretical_loss": 3.841874425731619,
      "tokens_seen": 597063680
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.00041367101303911737,
      "loss": 2.7813,
      "theoretical_loss": 3.841830387920821,
      "tokens_seen": 597129216
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.00041366098294884655,
      "loss": 2.6166,
      "theoretical_loss": 3.8417863562961205,
      "tokens_seen": 597194752
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.00041365095285857573,
      "loss": 2.9736,
      "theoretical_loss": 3.841742330855971,
      "tokens_seen": 597260288
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.00041364092276830497,
      "loss": 3.0773,
      "theoretical_loss": 3.841698311598824,
      "tokens_seen": 597325824
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.0004136308926780341,
      "loss": 3.1391,
      "theoretical_loss": 3.8416542985231343,
      "tokens_seen": 597391360
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.00041362086258776333,
      "loss": 2.9423,
      "theoretical_loss": 3.841610291627355,
      "tokens_seen": 597456896
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.0004136108324974925,
      "loss": 2.874,
      "theoretical_loss": 3.841566290909941,
      "tokens_seen": 597522432
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.0004136008024072217,
      "loss": 2.9495,
      "theoretical_loss": 3.8415222963693467,
      "tokens_seen": 597587968
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.0004135907723169509,
      "loss": 3.039,
      "theoretical_loss": 3.8414783080040285,
      "tokens_seen": 597653504
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.00041358074222668005,
      "loss": 2.8841,
      "theoretical_loss": 3.8414343258124424,
      "tokens_seen": 597719040
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.00041357071213640924,
      "loss": 2.803,
      "theoretical_loss": 3.841390349793045,
      "tokens_seen": 597784576
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.00041356068204613847,
      "loss": 2.9264,
      "theoretical_loss": 3.8413463799442935,
      "tokens_seen": 597850112
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.0004135506519558676,
      "loss": 3.0903,
      "theoretical_loss": 3.8413024162646465,
      "tokens_seen": 597915648
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.00041354062186559683,
      "loss": 2.7137,
      "theoretical_loss": 3.8412584587525616,
      "tokens_seen": 597981184
    },
    {
      "epoch": 2.0,
      "objective/train/docs_used": 718818,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.732727289199829,
      "objective/train/theoretical_loss": 3.841236482308873,
      "objective/train/tokens_used": 618473952,
      "theoretical_loss": 3.841236482308873,
      "tokens_seen": 598013952
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.00041353059177532596,
      "loss": 3.0269,
      "theoretical_loss": 3.841214507406498,
      "tokens_seen": 598046720
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.0004135205616850552,
      "loss": 2.9763,
      "theoretical_loss": 3.8411705622249155,
      "tokens_seen": 598112256
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.0004135105315947844,
      "loss": 3.026,
      "theoretical_loss": 3.841126623206274,
      "tokens_seen": 598177792
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.00041350050150451356,
      "loss": 3.0752,
      "theoretical_loss": 3.841082690349034,
      "tokens_seen": 598243328
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.00041349047141424274,
      "loss": 2.9446,
      "theoretical_loss": 3.8410387636516568,
      "tokens_seen": 598308864
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.0004134804413239719,
      "loss": 2.8162,
      "theoretical_loss": 3.840994843112604,
      "tokens_seen": 598374400
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.0004134704112337011,
      "loss": 2.8233,
      "theoretical_loss": 3.8409509287303383,
      "tokens_seen": 598439936
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.00041346038114343034,
      "loss": 2.9554,
      "theoretical_loss": 3.840907020503322,
      "tokens_seen": 598505472
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.00041345035105315946,
      "loss": 2.7912,
      "theoretical_loss": 3.840863118430019,
      "tokens_seen": 598571008
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.0004134403209628887,
      "loss": 2.9135,
      "theoretical_loss": 3.840819222508893,
      "tokens_seen": 598636544
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.0004134302908726179,
      "loss": 3.072,
      "theoretical_loss": 3.840775332738408,
      "tokens_seen": 598702080
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.00041342026078234706,
      "loss": 2.8755,
      "theoretical_loss": 3.84073144911703,
      "tokens_seen": 598767616
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.00041341023069207624,
      "loss": 2.984,
      "theoretical_loss": 3.8406875716432243,
      "tokens_seen": 598833152
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.0004134002006018054,
      "loss": 3.0003,
      "theoretical_loss": 3.8406437003154568,
      "tokens_seen": 598898688
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.0004133901705115346,
      "loss": 2.9602,
      "theoretical_loss": 3.840599835132194,
      "tokens_seen": 598964224
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.00041338014042126384,
      "loss": 2.85,
      "theoretical_loss": 3.840555976091904,
      "tokens_seen": 599029760
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.00041337011033099297,
      "loss": 2.9622,
      "theoretical_loss": 3.840512123193054,
      "tokens_seen": 599095296
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.0004133600802407222,
      "loss": 2.9202,
      "theoretical_loss": 3.8404682764341125,
      "tokens_seen": 599160832
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.00041335005015045133,
      "loss": 2.9157,
      "theoretical_loss": 3.840424435813548,
      "tokens_seen": 599226368
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.00041334002006018056,
      "loss": 2.9191,
      "theoretical_loss": 3.84038060132983,
      "tokens_seen": 599291904
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.00041332998996990975,
      "loss": 2.898,
      "theoretical_loss": 3.8403367729814297,
      "tokens_seen": 599357440
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.0004133199598796389,
      "loss": 2.9767,
      "theoretical_loss": 3.840292950766816,
      "tokens_seen": 599422976
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.0004133099297893681,
      "loss": 2.8749,
      "theoretical_loss": 3.840249134684461,
      "tokens_seen": 599488512
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.0004132998996990973,
      "loss": 2.9314,
      "theoretical_loss": 3.8402053247328363,
      "tokens_seen": 599554048
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.00041328986960882647,
      "loss": 2.8579,
      "theoretical_loss": 3.8401615209104136,
      "tokens_seen": 599619584
    },
    {
      "epoch": 2.0,
      "objective/train/docs_used": 723725,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7808403968811035,
      "objective/train/theoretical_loss": 3.8401396212971757,
      "objective/train/tokens_used": 620112352,
      "theoretical_loss": 3.8401396212971757,
      "tokens_seen": 599652352
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.0004132798395185557,
      "loss": 2.7878,
      "theoretical_loss": 3.8401177232156654,
      "tokens_seen": 599685120
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.00041326980942828483,
      "loss": 2.812,
      "theoretical_loss": 3.8400739316470665,
      "tokens_seen": 599750656
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.00041325977933801407,
      "loss": 3.0098,
      "theoretical_loss": 3.8400301462030892,
      "tokens_seen": 599816192
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.00041324974924774325,
      "loss": 2.9269,
      "theoretical_loss": 3.8399863668822087,
      "tokens_seen": 599881728
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.00041323971915747243,
      "loss": 2.8626,
      "theoretical_loss": 3.8399425936828995,
      "tokens_seen": 599947264
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.0004132296890672016,
      "loss": 2.9165,
      "theoretical_loss": 3.839898826603637,
      "tokens_seen": 600012800
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.0004132196589769308,
      "loss": 2.9539,
      "theoretical_loss": 3.839855065642898,
      "tokens_seen": 600078336
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.00041320962888665997,
      "loss": 2.9778,
      "theoretical_loss": 3.839811310799158,
      "tokens_seen": 600143872
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.0004131995987963892,
      "loss": 2.8436,
      "theoretical_loss": 3.839767562070895,
      "tokens_seen": 600209408
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.00041318956870611833,
      "loss": 2.8844,
      "theoretical_loss": 3.839723819456586,
      "tokens_seen": 600274944
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.00041317953861584757,
      "loss": 3.0573,
      "theoretical_loss": 3.8396800829547093,
      "tokens_seen": 600340480
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.0004131695085255767,
      "loss": 2.857,
      "theoretical_loss": 3.839636352563744,
      "tokens_seen": 600406016
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.00041315947843530593,
      "loss": 2.8885,
      "theoretical_loss": 3.83959262828217,
      "tokens_seen": 600471552
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.0004131494483450351,
      "loss": 2.8889,
      "theoretical_loss": 3.8395489101084657,
      "tokens_seen": 600537088
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.0004131394182547643,
      "loss": 2.8557,
      "theoretical_loss": 3.8395051980411123,
      "tokens_seen": 600602624
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.0004131293881644935,
      "loss": 2.9485,
      "theoretical_loss": 3.8394614920785903,
      "tokens_seen": 600668160
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.0004131193580742227,
      "loss": 2.8408,
      "theoretical_loss": 3.8394177922193817,
      "tokens_seen": 600733696
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.00041310932798395184,
      "loss": 2.8704,
      "theoretical_loss": 3.8393740984619686,
      "tokens_seen": 600799232
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.0004130992978936811,
      "loss": 2.8666,
      "theoretical_loss": 3.8393304108048323,
      "tokens_seen": 600864768
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.0004130892678034102,
      "loss": 2.9289,
      "theoretical_loss": 3.839286729246458,
      "tokens_seen": 600930304
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.00041307923771313944,
      "loss": 2.8433,
      "theoretical_loss": 3.839243053785327,
      "tokens_seen": 600995840
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.0004130692076228686,
      "loss": 2.8201,
      "theoretical_loss": 3.8391993844199255,
      "tokens_seen": 601061376
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.0004130591775325978,
      "loss": 2.9049,
      "theoretical_loss": 3.839155721148737,
      "tokens_seen": 601126912
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.000413049147442327,
      "loss": 2.8023,
      "theoretical_loss": 3.8391120639702474,
      "tokens_seen": 601192448
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.00041303911735205616,
      "loss": 2.9873,
      "theoretical_loss": 3.839068412882942,
      "tokens_seen": 601257984
    },
    {
      "epoch": 2.0,
      "objective/train/docs_used": 728779,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.1727378368377686,
      "objective/train/theoretical_loss": 3.8390465896230106,
      "objective/train/tokens_used": 621750752,
      "theoretical_loss": 3.8390465896230106,
      "tokens_seen": 601290752
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.00041302908726178534,
      "loss": 2.9188,
      "theoretical_loss": 3.8390247678853076,
      "tokens_seen": 601323520
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.0004130190571715146,
      "loss": 2.8634,
      "theoretical_loss": 3.8389811289758304,
      "tokens_seen": 601389056
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.0004130090270812437,
      "loss": 2.6309,
      "theoretical_loss": 3.838937496152999,
      "tokens_seen": 601454592
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.00041299899699097294,
      "loss": 2.9133,
      "theoretical_loss": 3.8388938694153003,
      "tokens_seen": 601520128
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.00041298896690070207,
      "loss": 2.9708,
      "theoretical_loss": 3.838850248761223,
      "tokens_seen": 601585664
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.0004129789368104313,
      "loss": 2.7942,
      "theoretical_loss": 3.8388066341892566,
      "tokens_seen": 601651200
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.0004129689067201605,
      "loss": 2.8799,
      "theoretical_loss": 3.838763025697891,
      "tokens_seen": 601716736
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.00041295887662988966,
      "loss": 2.881,
      "theoretical_loss": 3.8387194232856148,
      "tokens_seen": 601782272
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.00041294884653961884,
      "loss": 2.9561,
      "theoretical_loss": 3.8386758269509205,
      "tokens_seen": 601847808
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.0004129388164493481,
      "loss": 2.7078,
      "theoretical_loss": 3.8386322366922982,
      "tokens_seen": 601913344
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.0004129287863590772,
      "loss": 2.9614,
      "theoretical_loss": 3.83858865250824,
      "tokens_seen": 601978880
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.00041291875626880644,
      "loss": 2.8647,
      "theoretical_loss": 3.8385450743972376,
      "tokens_seen": 602044416
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.0004129087261785356,
      "loss": 2.8494,
      "theoretical_loss": 3.8385015023577846,
      "tokens_seen": 602109952
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.0004128986960882648,
      "loss": 2.8509,
      "theoretical_loss": 3.8384579363883744,
      "tokens_seen": 602175488
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.00041288866599799404,
      "loss": 2.8159,
      "theoretical_loss": 3.8384143764875,
      "tokens_seen": 602241024
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.00041287863590772317,
      "loss": 3.0505,
      "theoretical_loss": 3.8383708226536575,
      "tokens_seen": 602306560
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.0004128686058174524,
      "loss": 2.7001,
      "theoretical_loss": 3.83832727488534,
      "tokens_seen": 602372096
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.00041285857572718153,
      "loss": 2.7911,
      "theoretical_loss": 3.838283733181044,
      "tokens_seen": 602437632
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.00041284854563691076,
      "loss": 3.0129,
      "theoretical_loss": 3.8382401975392653,
      "tokens_seen": 602503168
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.00041283851554663995,
      "loss": 2.7819,
      "theoretical_loss": 3.8381966679585005,
      "tokens_seen": 602568704
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.0004128284854563691,
      "loss": 2.8773,
      "theoretical_loss": 3.838153144437247,
      "tokens_seen": 602634240
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.0004128184553660983,
      "loss": 2.819,
      "theoretical_loss": 3.838109626974002,
      "tokens_seen": 602699776
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.0004128084252758275,
      "loss": 3.0012,
      "theoretical_loss": 3.8380661155672646,
      "tokens_seen": 602765312
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.00041279839518555667,
      "loss": 2.8227,
      "theoretical_loss": 3.838022610215532,
      "tokens_seen": 602830848
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.0004127883650952859,
      "loss": 2.6656,
      "theoretical_loss": 3.837979110917305,
      "tokens_seen": 602896384
    },
    {
      "epoch": 2.0,
      "objective/train/docs_used": 731087,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8379440307617188,
      "objective/train/theoretical_loss": 3.837957363537787,
      "objective/train/tokens_used": 623389152,
      "theoretical_loss": 3.837957363537787,
      "tokens_seen": 602929152
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.00041277833500501503,
      "loss": 2.9037,
      "theoretical_loss": 3.837935617671082,
      "tokens_seen": 602961920
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.00041276830491474427,
      "loss": 2.9334,
      "theoretical_loss": 3.837892130475365,
      "tokens_seen": 603027456
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.00041275827482447345,
      "loss": 2.8944,
      "theoretical_loss": 3.837848649328653,
      "tokens_seen": 603092992
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.00041274824473420263,
      "loss": 2.8149,
      "theoretical_loss": 3.8378051742294494,
      "tokens_seen": 603158528
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.0004127382146439318,
      "loss": 2.8712,
      "theoretical_loss": 3.8377617051762547,
      "tokens_seen": 603224064
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.000412728184553661,
      "loss": 2.8573,
      "theoretical_loss": 3.8377182421675715,
      "tokens_seen": 603289600
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.00041271815446339017,
      "loss": 2.8153,
      "theoretical_loss": 3.8376747852019033,
      "tokens_seen": 603355136
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.0004127081243731194,
      "loss": 2.7117,
      "theoretical_loss": 3.837631334277753,
      "tokens_seen": 603420672
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.00041269809428284854,
      "loss": 2.8129,
      "theoretical_loss": 3.8375878893936255,
      "tokens_seen": 603486208
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.00041268806419257777,
      "loss": 2.9922,
      "theoretical_loss": 3.837544450548025,
      "tokens_seen": 603551744
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.0004126780341023069,
      "loss": 2.7937,
      "theoretical_loss": 3.8375010177394566,
      "tokens_seen": 603617280
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.00041266800401203613,
      "loss": 2.6681,
      "theoretical_loss": 3.8374575909664257,
      "tokens_seen": 603682816
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.0004126579739217653,
      "loss": 2.8324,
      "theoretical_loss": 3.837414170227439,
      "tokens_seen": 603748352
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.0004126479438314945,
      "loss": 3.0201,
      "theoretical_loss": 3.8373707555210035,
      "tokens_seen": 603813888
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.0004126379137412237,
      "loss": 2.7794,
      "theoretical_loss": 3.8373273468456253,
      "tokens_seen": 603879424
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.0004126278836509529,
      "loss": 2.9055,
      "theoretical_loss": 3.8372839441998137,
      "tokens_seen": 603944960
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.00041261785356068204,
      "loss": 2.978,
      "theoretical_loss": 3.8372405475820757,
      "tokens_seen": 604010496
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.0004126078234704113,
      "loss": 2.9492,
      "theoretical_loss": 3.8371971569909205,
      "tokens_seen": 604076032
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.0004125977933801404,
      "loss": 2.8896,
      "theoretical_loss": 3.837153772424858,
      "tokens_seen": 604141568
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.00041258776328986964,
      "loss": 2.9618,
      "theoretical_loss": 3.837110393882398,
      "tokens_seen": 604207104
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.0004125777331995988,
      "loss": 2.9102,
      "theoretical_loss": 3.83706702136205,
      "tokens_seen": 604272640
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.000412567703109328,
      "loss": 3.0573,
      "theoretical_loss": 3.8370236548623264,
      "tokens_seen": 604338176
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.0004125576730190572,
      "loss": 2.7037,
      "theoretical_loss": 3.8369802943817377,
      "tokens_seen": 604403712
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.00041254764292878636,
      "loss": 2.8876,
      "theoretical_loss": 3.8369369399187963,
      "tokens_seen": 604469248
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.00041253761283851554,
      "loss": 2.9452,
      "theoretical_loss": 3.836893591472015,
      "tokens_seen": 604534784
    },
    {
      "epoch": 2.0,
      "objective/train/docs_used": 736119,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.0697572231292725,
      "objective/train/theoretical_loss": 3.83687191950422,
      "objective/train/tokens_used": 625027552,
      "theoretical_loss": 3.83687191950422,
      "tokens_seen": 604567552
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.0004125275827482448,
      "loss": 3.0517,
      "theoretical_loss": 3.836850249039906,
      "tokens_seen": 604600320
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.0004125175526579739,
      "loss": 2.8523,
      "theoretical_loss": 3.8368069126209843,
      "tokens_seen": 604665856
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.00041250752256770314,
      "loss": 2.9029,
      "theoretical_loss": 3.836763582213763,
      "tokens_seen": 604731392
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.00041249749247743227,
      "loss": 2.6393,
      "theoretical_loss": 3.8367202578167574,
      "tokens_seen": 604796928
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.0004124874623871615,
      "loss": 2.7526,
      "theoretical_loss": 3.8366769394284823,
      "tokens_seen": 604862464
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.0004124774322968907,
      "loss": 2.9766,
      "theoretical_loss": 3.836633627047453,
      "tokens_seen": 604928000
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.00041246740220661986,
      "loss": 2.8323,
      "theoretical_loss": 3.8365903206721867,
      "tokens_seen": 604993536
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.00041245737211634904,
      "loss": 2.823,
      "theoretical_loss": 3.8365470203012,
      "tokens_seen": 605059072
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.0004124473420260783,
      "loss": 2.9371,
      "theoretical_loss": 3.83650372593301,
      "tokens_seen": 605124608
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.0004124373119358074,
      "loss": 2.8824,
      "theoretical_loss": 3.836460437566134,
      "tokens_seen": 605190144
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.00041242728184553664,
      "loss": 2.9946,
      "theoretical_loss": 3.8364171551990918,
      "tokens_seen": 605255680
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.00041241725175526577,
      "loss": 3.0648,
      "theoretical_loss": 3.836373878830401,
      "tokens_seen": 605321216
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.000412407221664995,
      "loss": 2.8366,
      "theoretical_loss": 3.8363306084585815,
      "tokens_seen": 605386752
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.0004123971915747242,
      "loss": 2.9835,
      "theoretical_loss": 3.836287344082153,
      "tokens_seen": 605452288
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.00041238716148445337,
      "loss": 2.8837,
      "theoretical_loss": 3.8362440856996365,
      "tokens_seen": 605517824
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.00041237713139418255,
      "loss": 2.726,
      "theoretical_loss": 3.8362008333095527,
      "tokens_seen": 605583360
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.00041236710130391173,
      "loss": 2.8451,
      "theoretical_loss": 3.8361575869104225,
      "tokens_seen": 605648896
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.0004123570712136409,
      "loss": 2.7774,
      "theoretical_loss": 3.836114346500769,
      "tokens_seen": 605714432
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.00041234704112337015,
      "loss": 2.7372,
      "theoretical_loss": 3.8360711120791144,
      "tokens_seen": 605779968
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.00041233701103309927,
      "loss": 2.8797,
      "theoretical_loss": 3.8360278836439816,
      "tokens_seen": 605845504
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.0004123269809428285,
      "loss": 3.046,
      "theoretical_loss": 3.835984661193895,
      "tokens_seen": 605911040
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.00041231695085255763,
      "loss": 2.7871,
      "theoretical_loss": 3.8359414447273776,
      "tokens_seen": 605976576
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.00041230692076228687,
      "loss": 2.8635,
      "theoretical_loss": 3.835898234242954,
      "tokens_seen": 606042112
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.00041229689067201605,
      "loss": 2.9083,
      "theoretical_loss": 3.8358550297391507,
      "tokens_seen": 606107648
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.00041228686058174523,
      "loss": 2.666,
      "theoretical_loss": 3.8358118312144924,
      "tokens_seen": 606173184
    },
    {
      "epoch": 2.0,
      "objective/train/docs_used": 740966,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.874371290206909,
      "objective/train/theoretical_loss": 3.8357902341938823,
      "objective/train/tokens_used": 626665952,
      "theoretical_loss": 3.8357902341938823,
      "tokens_seen": 606205952
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.0004122768304914744,
      "loss": 2.7719,
      "theoretical_loss": 3.835768638667506,
      "tokens_seen": 606238720
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.00041226680040120365,
      "loss": 2.8218,
      "theoretical_loss": 3.8357254520967174,
      "tokens_seen": 606304256
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.0004122567703109328,
      "loss": 2.8802,
      "theoretical_loss": 3.835682271500655,
      "tokens_seen": 606369792
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.000412246740220662,
      "loss": 3.0319,
      "theoretical_loss": 3.8356390968778453,
      "tokens_seen": 606435328
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.00041223671013039114,
      "loss": 2.865,
      "theoretical_loss": 3.8355959282268177,
      "tokens_seen": 606500864
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.0004122266800401204,
      "loss": 2.9013,
      "theoretical_loss": 3.8355527655461,
      "tokens_seen": 606566400
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.00041221664994984955,
      "loss": 2.9206,
      "theoretical_loss": 3.835509608834223,
      "tokens_seen": 606631936
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.00041220661985957874,
      "loss": 2.9424,
      "theoretical_loss": 3.8354664580897153,
      "tokens_seen": 606697472
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.0004121965897693079,
      "loss": 2.8802,
      "theoretical_loss": 3.835423313311108,
      "tokens_seen": 606763008
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.0004121865596790371,
      "loss": 2.9102,
      "theoretical_loss": 3.8353801744969322,
      "tokens_seen": 606828544
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.0004121765295887663,
      "loss": 2.8268,
      "theoretical_loss": 3.835337041645719,
      "tokens_seen": 606894080
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.0004121664994984955,
      "loss": 2.9799,
      "theoretical_loss": 3.8352939147559995,
      "tokens_seen": 606959616
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.0004121564694082247,
      "loss": 2.7371,
      "theoretical_loss": 3.835250793826308,
      "tokens_seen": 607025152
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.0004121464393179539,
      "loss": 2.7789,
      "theoretical_loss": 3.835207678855176,
      "tokens_seen": 607090688
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.0004121364092276831,
      "loss": 2.9266,
      "theoretical_loss": 3.835164569841138,
      "tokens_seen": 607156224
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.00041212637913741224,
      "loss": 2.6294,
      "theoretical_loss": 3.8351214667827276,
      "tokens_seen": 607221760
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.0004121163490471415,
      "loss": 2.7727,
      "theoretical_loss": 3.835078369678479,
      "tokens_seen": 607287296
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.0004121063189568706,
      "loss": 2.9085,
      "theoretical_loss": 3.835035278526928,
      "tokens_seen": 607352832
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.00041209628886659984,
      "loss": 2.8811,
      "theoretical_loss": 3.8349921933266105,
      "tokens_seen": 607418368
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.000412086258776329,
      "loss": 2.8147,
      "theoretical_loss": 3.8349491140760614,
      "tokens_seen": 607483904
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.0004120762286860582,
      "loss": 2.9682,
      "theoretical_loss": 3.8349060407738182,
      "tokens_seen": 607549440
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.0004120661985957874,
      "loss": 2.9455,
      "theoretical_loss": 3.834862973418418,
      "tokens_seen": 607614976
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.00041205616850551656,
      "loss": 2.921,
      "theoretical_loss": 3.834819912008398,
      "tokens_seen": 607680512
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.00041204613841524574,
      "loss": 2.913,
      "theoretical_loss": 3.8347768565422973,
      "tokens_seen": 607746048
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.000412036108324975,
      "loss": 2.8299,
      "theoretical_loss": 3.8347338070186536,
      "tokens_seen": 607811584
    },
    {
      "epoch": 2.0,
      "objective/train/docs_used": 746024,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.745020627975464,
      "objective/train/theoretical_loss": 3.834712284484797,
      "objective/train/tokens_used": 628304352,
      "theoretical_loss": 3.834712284484797,
      "tokens_seen": 607844352
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.0004120260782347041,
      "loss": 2.9376,
      "theoretical_loss": 3.834690763436007,
      "tokens_seen": 607877120
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.00041201604814443334,
      "loss": 2.8025,
      "theoretical_loss": 3.8346477257928964,
      "tokens_seen": 607942656
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.00041200601805416247,
      "loss": 2.9184,
      "theoretical_loss": 3.8346046940878624,
      "tokens_seen": 608008192
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.0004119959879638917,
      "loss": 2.8953,
      "theoretical_loss": 3.834561668319447,
      "tokens_seen": 608073728
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.0004119859578736209,
      "loss": 2.7333,
      "theoretical_loss": 3.834518648486189,
      "tokens_seen": 608139264
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.00041197592778335006,
      "loss": 2.8579,
      "theoretical_loss": 3.8344756345866324,
      "tokens_seen": 608204800
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.00041196589769307924,
      "loss": 2.8816,
      "theoretical_loss": 3.834432626619318,
      "tokens_seen": 608270336
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.0004119558676028085,
      "loss": 2.8811,
      "theoretical_loss": 3.8343896245827898,
      "tokens_seen": 608335872
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.0004119458375125376,
      "loss": 2.997,
      "theoretical_loss": 3.8343466284755907,
      "tokens_seen": 608401408
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.00041193580742226684,
      "loss": 2.8126,
      "theoretical_loss": 3.8343036382962645,
      "tokens_seen": 608466944
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.00041192577733199597,
      "loss": 2.9976,
      "theoretical_loss": 3.834260654043356,
      "tokens_seen": 608532480
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.0004119157472417252,
      "loss": 2.8142,
      "theoretical_loss": 3.8342176757154087,
      "tokens_seen": 608598016
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.0004119057171514544,
      "loss": 2.7509,
      "theoretical_loss": 3.8341747033109703,
      "tokens_seen": 608663552
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.00041189568706118357,
      "loss": 2.6745,
      "theoretical_loss": 3.8341317368285845,
      "tokens_seen": 608729088
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.00041188565697091275,
      "loss": 2.8161,
      "theoretical_loss": 3.8340887762667992,
      "tokens_seen": 608794624
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.00041187562688064193,
      "loss": 2.9308,
      "theoretical_loss": 3.8340458216241613,
      "tokens_seen": 608860160
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.0004118655967903711,
      "loss": 2.8721,
      "theoretical_loss": 3.834002872899217,
      "tokens_seen": 608925696
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.00041185556670010035,
      "loss": 2.7833,
      "theoretical_loss": 3.833959930090516,
      "tokens_seen": 608991232
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.00041184553660982947,
      "loss": 2.8491,
      "theoretical_loss": 3.8339169931966053,
      "tokens_seen": 609056768
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.0004118355065195587,
      "loss": 2.8048,
      "theoretical_loss": 3.8338740622160343,
      "tokens_seen": 609122304
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.00041182547642928783,
      "loss": 2.8229,
      "theoretical_loss": 3.833831137147353,
      "tokens_seen": 609187840
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.00041181544633901707,
      "loss": 3.0615,
      "theoretical_loss": 3.8337882179891114,
      "tokens_seen": 609253376
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.00041180541624874625,
      "loss": 2.6738,
      "theoretical_loss": 3.83374530473986,
      "tokens_seen": 609318912
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.00041179538615847543,
      "loss": 2.706,
      "theoretical_loss": 3.833702397398149,
      "tokens_seen": 609384448
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.0004117853560682046,
      "loss": 2.7287,
      "theoretical_loss": 3.8336594959625314,
      "tokens_seen": 609449984
    },
    {
      "epoch": 2.0,
      "objective/train/docs_used": 751044,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.0871493816375732,
      "objective/train/theoretical_loss": 3.833638047459054,
      "objective/train/tokens_used": 629942752,
      "theoretical_loss": 3.833638047459054,
      "tokens_seen": 609482752
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.00041177532597793385,
      "loss": 3.0586,
      "theoretical_loss": 3.833616600431558,
      "tokens_seen": 609515520
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.000411765295887663,
      "loss": 2.8549,
      "theoretical_loss": 3.8335737108037815,
      "tokens_seen": 609581056
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.0004117552657973922,
      "loss": 2.872,
      "theoretical_loss": 3.833530827077756,
      "tokens_seen": 609646592
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.00041174523570712134,
      "loss": 2.8627,
      "theoretical_loss": 3.833487949252034,
      "tokens_seen": 609712128
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.0004117352056168506,
      "loss": 2.8217,
      "theoretical_loss": 3.8334450773251705,
      "tokens_seen": 609777664
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.00041172517552657975,
      "loss": 2.99,
      "theoretical_loss": 3.833402211295719,
      "tokens_seen": 609843200
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.00041171514543630894,
      "loss": 2.7994,
      "theoretical_loss": 3.833359351162236,
      "tokens_seen": 609908736
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.0004117051153460381,
      "loss": 2.9186,
      "theoretical_loss": 3.8333164969232767,
      "tokens_seen": 609974272
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004116950852557673,
      "loss": 2.9245,
      "theoretical_loss": 3.8332736485773964,
      "tokens_seen": 610039808
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004116850551654965,
      "loss": 2.7503,
      "theoretical_loss": 3.8332308061231526,
      "tokens_seen": 610105344
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004116750250752257,
      "loss": 2.7762,
      "theoretical_loss": 3.833187969559102,
      "tokens_seen": 610170880
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00041166499498495484,
      "loss": 3.0806,
      "theoretical_loss": 3.833145138883803,
      "tokens_seen": 610236416
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004116549648946841,
      "loss": 2.5722,
      "theoretical_loss": 3.8331023140958127,
      "tokens_seen": 610301952
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004116449348044132,
      "loss": 2.9739,
      "theoretical_loss": 3.833059495193691,
      "tokens_seen": 610367488
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00041163490471414244,
      "loss": 2.908,
      "theoretical_loss": 3.833016682175996,
      "tokens_seen": 610433024
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004116248746238716,
      "loss": 2.9373,
      "theoretical_loss": 3.8329738750412883,
      "tokens_seen": 610498560
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004116148445336008,
      "loss": 2.7929,
      "theoretical_loss": 3.832931073788127,
      "tokens_seen": 610564096
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00041160481444333,
      "loss": 2.9314,
      "theoretical_loss": 3.8328882784150746,
      "tokens_seen": 610629632
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004115947843530592,
      "loss": 2.9051,
      "theoretical_loss": 3.8328454889206904,
      "tokens_seen": 610695168
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00041158475426278834,
      "loss": 2.8858,
      "theoretical_loss": 3.8328027053035374,
      "tokens_seen": 610760704
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004115747241725176,
      "loss": 3.0172,
      "theoretical_loss": 3.8327599275621775,
      "tokens_seen": 610826240
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004115646940822467,
      "loss": 2.8714,
      "theoretical_loss": 3.832717155695173,
      "tokens_seen": 610891776
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00041155466399197594,
      "loss": 2.7776,
      "theoretical_loss": 3.832674389701088,
      "tokens_seen": 610957312
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004115446339017051,
      "loss": 2.892,
      "theoretical_loss": 3.8326316295784855,
      "tokens_seen": 611022848
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004115346038114343,
      "loss": 2.9538,
      "theoretical_loss": 3.8325888753259303,
      "tokens_seen": 611088384
    },
    {
      "epoch": 2.01,
      "objective/train/docs_used": 756149,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.102915048599243,
      "objective/train/theoretical_loss": 3.8325675004004713,
      "objective/train/tokens_used": 631581152,
      "theoretical_loss": 3.8325675004004713,
      "tokens_seen": 611121152
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004115245737211635,
      "loss": 2.9014,
      "theoretical_loss": 3.8325461269419865,
      "tokens_seen": 611153920
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00041151454363089267,
      "loss": 2.8073,
      "theoretical_loss": 3.8325033844252197,
      "tokens_seen": 611219456
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00041150451354062185,
      "loss": 2.8775,
      "theoretical_loss": 3.8324606477741963,
      "tokens_seen": 611284992
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004114944834503511,
      "loss": 2.8357,
      "theoretical_loss": 3.832417916987482,
      "tokens_seen": 611350528
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004114844533600802,
      "loss": 2.7828,
      "theoretical_loss": 3.832375192063643,
      "tokens_seen": 611416064
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00041147442326980944,
      "loss": 2.8968,
      "theoretical_loss": 3.832332473001247,
      "tokens_seen": 611481600
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00041146439317953857,
      "loss": 2.957,
      "theoretical_loss": 3.832289759798863,
      "tokens_seen": 611547136
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004114543630892678,
      "loss": 2.8542,
      "theoretical_loss": 3.832247052455058,
      "tokens_seen": 611612672
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.000411444332998997,
      "loss": 2.9705,
      "theoretical_loss": 3.832204350968401,
      "tokens_seen": 611678208
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00041143430290872617,
      "loss": 2.7964,
      "theoretical_loss": 3.832161655337462,
      "tokens_seen": 611743744
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00041142427281845535,
      "loss": 2.6554,
      "theoretical_loss": 3.832118965560809,
      "tokens_seen": 611809280
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004114142427281846,
      "loss": 2.8301,
      "theoretical_loss": 3.832076281637014,
      "tokens_seen": 611874816
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00041140421263791377,
      "loss": 2.6453,
      "theoretical_loss": 3.8320336035646476,
      "tokens_seen": 611940352
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00041139418254764295,
      "loss": 2.846,
      "theoretical_loss": 3.831990931342281,
      "tokens_seen": 612005888
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00041138415245737213,
      "loss": 2.8807,
      "theoretical_loss": 3.831948264968485,
      "tokens_seen": 612071424
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004113741223671013,
      "loss": 2.842,
      "theoretical_loss": 3.8319056044418334,
      "tokens_seen": 612136960
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00041136409227683055,
      "loss": 2.9488,
      "theoretical_loss": 3.8318629497608985,
      "tokens_seen": 612202496
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00041135406218655967,
      "loss": 2.8403,
      "theoretical_loss": 3.831820300924253,
      "tokens_seen": 612268032
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004113440320962889,
      "loss": 2.8262,
      "theoretical_loss": 3.8317776579304716,
      "tokens_seen": 612333568
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00041133400200601803,
      "loss": 2.9601,
      "theoretical_loss": 3.8317350207781278,
      "tokens_seen": 612399104
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00041132397191574727,
      "loss": 2.8308,
      "theoretical_loss": 3.831692389465797,
      "tokens_seen": 612464640
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00041131394182547645,
      "loss": 2.7944,
      "theoretical_loss": 3.831649763992054,
      "tokens_seen": 612530176
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00041130391173520563,
      "loss": 2.942,
      "theoretical_loss": 3.831607144355476,
      "tokens_seen": 612595712
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004112938816449348,
      "loss": 2.8861,
      "theoretical_loss": 3.8315645305546377,
      "tokens_seen": 612661248
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00041128385155466405,
      "loss": 2.8315,
      "theoretical_loss": 3.8315219225881165,
      "tokens_seen": 612726784
    },
    {
      "epoch": 2.01,
      "objective/train/docs_used": 761270,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.108196973800659,
      "objective/train/theoretical_loss": 3.83150062079228,
      "objective/train/tokens_used": 633219552,
      "theoretical_loss": 3.83150062079228,
      "tokens_seen": 612759552
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004112738214643932,
      "loss": 2.84,
      "theoretical_loss": 3.8314793204544895,
      "tokens_seen": 612792320
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004112637913741224,
      "loss": 2.7863,
      "theoretical_loss": 3.831436724152335,
      "tokens_seen": 612857856
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00041125376128385154,
      "loss": 2.9543,
      "theoretical_loss": 3.831394133680231,
      "tokens_seen": 612923392
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004112437311935808,
      "loss": 2.9146,
      "theoretical_loss": 3.8313515490367562,
      "tokens_seen": 612988928
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00041123370110330995,
      "loss": 2.8663,
      "theoretical_loss": 3.8313089702204906,
      "tokens_seen": 613054464
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00041122367101303914,
      "loss": 2.9834,
      "theoretical_loss": 3.831266397230013,
      "tokens_seen": 613120000
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004112136409227683,
      "loss": 2.9004,
      "theoretical_loss": 3.8312238300639048,
      "tokens_seen": 613185536
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004112036108324975,
      "loss": 2.9223,
      "theoretical_loss": 3.8311812687207456,
      "tokens_seen": 613251072
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004111935807422267,
      "loss": 2.9421,
      "theoretical_loss": 3.8311387131991173,
      "tokens_seen": 613316608
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004111835506519559,
      "loss": 2.9294,
      "theoretical_loss": 3.831096163497602,
      "tokens_seen": 613382144
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00041117352056168504,
      "loss": 2.7985,
      "theoretical_loss": 3.8310536196147815,
      "tokens_seen": 613447680
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004111634904714143,
      "loss": 2.8329,
      "theoretical_loss": 3.831011081549239,
      "tokens_seen": 613513216
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004111534603811434,
      "loss": 3.1471,
      "theoretical_loss": 3.8309685492995573,
      "tokens_seen": 613578752
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00041114343029087264,
      "loss": 2.9241,
      "theoretical_loss": 3.830926022864321,
      "tokens_seen": 613644288
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004111334002006018,
      "loss": 2.7,
      "theoretical_loss": 3.8308835022421137,
      "tokens_seen": 613709824
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.000411123370110331,
      "loss": 2.9448,
      "theoretical_loss": 3.83084098743152,
      "tokens_seen": 613775360
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004111133400200602,
      "loss": 2.6292,
      "theoretical_loss": 3.8307984784311255,
      "tokens_seen": 613840896
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004111033099297894,
      "loss": 2.8419,
      "theoretical_loss": 3.8307559752395166,
      "tokens_seen": 613906432
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00041109327983951854,
      "loss": 3.0578,
      "theoretical_loss": 3.8307134778552783,
      "tokens_seen": 613971968
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004110832497492478,
      "loss": 2.6516,
      "theoretical_loss": 3.8306709862769983,
      "tokens_seen": 614037504
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004110732196589769,
      "loss": 2.8264,
      "theoretical_loss": 3.830628500503263,
      "tokens_seen": 614103040
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00041106318956870614,
      "loss": 2.635,
      "theoretical_loss": 3.830586020532661,
      "tokens_seen": 614168576
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004110531594784353,
      "loss": 2.9087,
      "theoretical_loss": 3.830543546363781,
      "tokens_seen": 614234112
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004110431293881645,
      "loss": 2.8193,
      "theoretical_loss": 3.8305010779952102,
      "tokens_seen": 614299648
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004110330992978937,
      "loss": 2.8163,
      "theoretical_loss": 3.8304586154255387,
      "tokens_seen": 614365184
    },
    {
      "epoch": 2.01,
      "objective/train/docs_used": 766282,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.869065046310425,
      "objective/train/theoretical_loss": 3.8304373863148493,
      "objective/train/tokens_used": 634857952,
      "theoretical_loss": 3.8304373863148493,
      "tokens_seen": 614397952
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00041102306920762287,
      "loss": 2.9004,
      "theoretical_loss": 3.8304161586533563,
      "tokens_seen": 614430720
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00041101303911735205,
      "loss": 2.8674,
      "theoretical_loss": 3.830373707677253,
      "tokens_seen": 614496256
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004110030090270813,
      "loss": 2.8783,
      "theoretical_loss": 3.83033126249582,
      "tokens_seen": 614561792
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004109929789368104,
      "loss": 3.0358,
      "theoretical_loss": 3.8302888231076473,
      "tokens_seen": 614627328
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00041098294884653965,
      "loss": 2.8569,
      "theoretical_loss": 3.830246389511328,
      "tokens_seen": 614692864
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00041097291875626877,
      "loss": 2.7641,
      "theoretical_loss": 3.8302039617054526,
      "tokens_seen": 614758400
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.000410962888665998,
      "loss": 2.9641,
      "theoretical_loss": 3.830161539688616,
      "tokens_seen": 614823936
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004109528585757272,
      "loss": 2.8592,
      "theoretical_loss": 3.8301191234594096,
      "tokens_seen": 614889472
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00041094282848545637,
      "loss": 2.8512,
      "theoretical_loss": 3.830076713016428,
      "tokens_seen": 614955008
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00041093279839518555,
      "loss": 2.6787,
      "theoretical_loss": 3.8300343083582646,
      "tokens_seen": 615020544
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004109227683049148,
      "loss": 2.7377,
      "theoretical_loss": 3.829991909483515,
      "tokens_seen": 615086080
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004109127382146439,
      "loss": 3.0285,
      "theoretical_loss": 3.8299495163907733,
      "tokens_seen": 615151616
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00041090270812437315,
      "loss": 2.7276,
      "theoretical_loss": 3.829907129078636,
      "tokens_seen": 615217152
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004108926780341023,
      "loss": 2.821,
      "theoretical_loss": 3.829864747545699,
      "tokens_seen": 615282688
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004108826479438315,
      "loss": 2.8471,
      "theoretical_loss": 3.8298223717905584,
      "tokens_seen": 615348224
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004108726178535607,
      "loss": 2.6164,
      "theoretical_loss": 3.8297800018118116,
      "tokens_seen": 615413760
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00041086258776328987,
      "loss": 2.7788,
      "theoretical_loss": 3.8297376376080567,
      "tokens_seen": 615479296
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00041085255767301905,
      "loss": 2.6422,
      "theoretical_loss": 3.829695279177891,
      "tokens_seen": 615544832
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00041084252758274823,
      "loss": 2.924,
      "theoretical_loss": 3.8296529265199135,
      "tokens_seen": 615610368
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004108324974924774,
      "loss": 2.9178,
      "theoretical_loss": 3.8296105796327233,
      "tokens_seen": 615675904
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00041082246740220665,
      "loss": 2.8156,
      "theoretical_loss": 3.82956823851492,
      "tokens_seen": 615741440
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004108124373119358,
      "loss": 2.8831,
      "theoretical_loss": 3.829525903165103,
      "tokens_seen": 615806976
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.000410802407221665,
      "loss": 2.7111,
      "theoretical_loss": 3.8294835735818733,
      "tokens_seen": 615872512
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004107923771313942,
      "loss": 2.9392,
      "theoretical_loss": 3.8294412497638324,
      "tokens_seen": 615938048
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004107823470411234,
      "loss": 2.6815,
      "theoretical_loss": 3.829398931709581,
      "tokens_seen": 616003584
    },
    {
      "epoch": 2.01,
      "objective/train/docs_used": 771277,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5148863792419434,
      "objective/train/theoretical_loss": 3.829377774843439,
      "objective/train/tokens_used": 636496352,
      "theoretical_loss": 3.829377774843439,
      "tokens_seen": 616036352
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00041077231695085256,
      "loss": 2.7967,
      "theoretical_loss": 3.8293566194177213,
      "tokens_seen": 616069120
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00041076228686058174,
      "loss": 2.6174,
      "theoretical_loss": 3.829314312886856,
      "tokens_seen": 616134656
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004107522567703109,
      "loss": 2.7027,
      "theoretical_loss": 3.829272012115588,
      "tokens_seen": 616200192
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00041074222668004015,
      "loss": 2.5431,
      "theoretical_loss": 3.8292297171025202,
      "tokens_seen": 616265728
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004107321965897693,
      "loss": 2.7535,
      "theoretical_loss": 3.8291874278462577,
      "tokens_seen": 616331264
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004107221664994985,
      "loss": 2.8045,
      "theoretical_loss": 3.8291451443454037,
      "tokens_seen": 616396800
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00041071213640922764,
      "loss": 2.6862,
      "theoretical_loss": 3.8291028665985642,
      "tokens_seen": 616462336
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004107021063189569,
      "loss": 2.6272,
      "theoretical_loss": 3.8290605946043437,
      "tokens_seen": 616527872
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00041069207622868606,
      "loss": 2.736,
      "theoretical_loss": 3.829018328361349,
      "tokens_seen": 616593408
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00041068204613841524,
      "loss": 2.9518,
      "theoretical_loss": 3.8289760678681852,
      "tokens_seen": 616658944
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004106720160481444,
      "loss": 2.8972,
      "theoretical_loss": 3.8289338131234603,
      "tokens_seen": 616724480
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004106619859578736,
      "loss": 2.8132,
      "theoretical_loss": 3.8288915641257812,
      "tokens_seen": 616790016
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00041065195586760284,
      "loss": 2.9286,
      "theoretical_loss": 3.828849320873756,
      "tokens_seen": 616855552
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.000410641925777332,
      "loss": 2.6631,
      "theoretical_loss": 3.828807083365992,
      "tokens_seen": 616921088
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004106318956870612,
      "loss": 2.8104,
      "theoretical_loss": 3.8287648516010995,
      "tokens_seen": 616986624
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004106218655967904,
      "loss": 2.8351,
      "theoretical_loss": 3.8287226255776865,
      "tokens_seen": 617052160
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004106118355065196,
      "loss": 2.757,
      "theoretical_loss": 3.8286804052943637,
      "tokens_seen": 617117696
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00041060180541624874,
      "loss": 2.9234,
      "theoretical_loss": 3.828638190749741,
      "tokens_seen": 617183232
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.000410591775325978,
      "loss": 2.6066,
      "theoretical_loss": 3.8285959819424296,
      "tokens_seen": 617248768
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004105817452357071,
      "loss": 2.9455,
      "theoretical_loss": 3.828553778871039,
      "tokens_seen": 617314304
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00041057171514543634,
      "loss": 2.924,
      "theoretical_loss": 3.8285115815341833,
      "tokens_seen": 617379840
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004105616850551655,
      "loss": 2.7455,
      "theoretical_loss": 3.828469389930473,
      "tokens_seen": 617445376
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004105516549648947,
      "loss": 2.7622,
      "theoretical_loss": 3.8284272040585208,
      "tokens_seen": 617510912
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004105416248746239,
      "loss": 2.8468,
      "theoretical_loss": 3.8283850239169412,
      "tokens_seen": 617576448
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00041053159478435307,
      "loss": 2.74,
      "theoretical_loss": 3.8283428495043466,
      "tokens_seen": 617641984
    },
    {
      "epoch": 2.01,
      "objective/train/docs_used": 772734,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.1529695987701416,
      "objective/train/theoretical_loss": 3.8283217644459855,
      "objective/train/tokens_used": 638134752,
      "theoretical_loss": 3.8283217644459855,
      "tokens_seen": 617674752
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00041052156469408225,
      "loss": 2.6459,
      "theoretical_loss": 3.8283006808193516,
      "tokens_seen": 617707520
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004105115346038115,
      "loss": 3.0692,
      "theoretical_loss": 3.8282585178605704,
      "tokens_seen": 617773056
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004105015045135406,
      "loss": 2.944,
      "theoretical_loss": 3.828216360626619,
      "tokens_seen": 617838592
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00041049147442326985,
      "loss": 2.9468,
      "theoretical_loss": 3.8281742091161117,
      "tokens_seen": 617904128
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00041048144433299897,
      "loss": 2.9138,
      "theoretical_loss": 3.828132063327666,
      "tokens_seen": 617969664
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004104714142427282,
      "loss": 2.7093,
      "theoretical_loss": 3.8280899232598973,
      "tokens_seen": 618035200
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004104613841524574,
      "loss": 2.9611,
      "theoretical_loss": 3.8280477889114226,
      "tokens_seen": 618100736
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00041045135406218657,
      "loss": 2.8029,
      "theoretical_loss": 3.82800566028086,
      "tokens_seen": 618166272
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00041044132397191575,
      "loss": 2.9444,
      "theoretical_loss": 3.827963537366827,
      "tokens_seen": 618231808
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.000410431293881645,
      "loss": 2.8183,
      "theoretical_loss": 3.827921420167942,
      "tokens_seen": 618297344
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004104212637913741,
      "loss": 2.9094,
      "theoretical_loss": 3.827879308682825,
      "tokens_seen": 618362880
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00041041123370110335,
      "loss": 2.7128,
      "theoretical_loss": 3.8278372029100938,
      "tokens_seen": 618428416
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004104012036108325,
      "loss": 2.8372,
      "theoretical_loss": 3.8277951028483694,
      "tokens_seen": 618493952
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004103911735205617,
      "loss": 3.0235,
      "theoretical_loss": 3.8277530084962716,
      "tokens_seen": 618559488
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004103811434302909,
      "loss": 2.7758,
      "theoretical_loss": 3.827710919852422,
      "tokens_seen": 618625024
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00041037111334002007,
      "loss": 2.7514,
      "theoretical_loss": 3.827668836915441,
      "tokens_seen": 618690560
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00041036108324974925,
      "loss": 2.6761,
      "theoretical_loss": 3.8276267596839513,
      "tokens_seen": 618756096
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00041035105315947844,
      "loss": 2.8231,
      "theoretical_loss": 3.8275846881565743,
      "tokens_seen": 618821632
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004103410230692076,
      "loss": 2.8471,
      "theoretical_loss": 3.827542622331933,
      "tokens_seen": 618887168
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00041033099297893685,
      "loss": 2.7828,
      "theoretical_loss": 3.827500562208651,
      "tokens_seen": 618952704
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.000410320962888666,
      "loss": 3.0159,
      "theoretical_loss": 3.8274585077853525,
      "tokens_seen": 619018240
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004103109327983952,
      "loss": 2.8519,
      "theoretical_loss": 3.82741645906066,
      "tokens_seen": 619083776
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004103009027081244,
      "loss": 2.7824,
      "theoretical_loss": 3.8273744160331997,
      "tokens_seen": 619149312
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004102908726178536,
      "loss": 2.8297,
      "theoretical_loss": 3.8273323787015965,
      "tokens_seen": 619214848
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00041028084252758276,
      "loss": 2.9612,
      "theoretical_loss": 3.8272903470644755,
      "tokens_seen": 619280384
    },
    {
      "epoch": 2.01,
      "objective/train/docs_used": 773070,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.051236152648926,
      "objective/train/theoretical_loss": 3.827269333380917,
      "objective/train/tokens_used": 639773152,
      "theoretical_loss": 3.827269333380917,
      "tokens_seen": 619313152
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00041027081243731194,
      "loss": 2.9915,
      "theoretical_loss": 3.8272483211204635,
      "tokens_seen": 619345920
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004102607823470411,
      "loss": 2.84,
      "theoretical_loss": 3.8272063008681867,
      "tokens_seen": 619411456
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00041025075225677035,
      "loss": 2.7795,
      "theoretical_loss": 3.827164286306272,
      "tokens_seen": 619476992
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004102407221664995,
      "loss": 2.9163,
      "theoretical_loss": 3.8271222774333467,
      "tokens_seen": 619542528
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004102306920762287,
      "loss": 2.7408,
      "theoretical_loss": 3.8270802742480394,
      "tokens_seen": 619608064
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00041022066198595784,
      "loss": 2.7077,
      "theoretical_loss": 3.8270382767489783,
      "tokens_seen": 619673600
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004102106318956871,
      "loss": 2.7924,
      "theoretical_loss": 3.8269962849347934,
      "tokens_seen": 619739136
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00041020060180541626,
      "loss": 2.9734,
      "theoretical_loss": 3.8269542988041123,
      "tokens_seen": 619804672
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00041019057171514544,
      "loss": 2.8563,
      "theoretical_loss": 3.826912318355566,
      "tokens_seen": 619870208
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004101805416248746,
      "loss": 2.7356,
      "theoretical_loss": 3.826870343587785,
      "tokens_seen": 619935744
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004101705115346038,
      "loss": 2.94,
      "theoretical_loss": 3.8268283744993994,
      "tokens_seen": 620001280
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.000410160481444333,
      "loss": 3.0616,
      "theoretical_loss": 3.8267864110890413,
      "tokens_seen": 620066816
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004101504513540622,
      "loss": 2.8621,
      "theoretical_loss": 3.8267444533553423,
      "tokens_seen": 620132352
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00041014042126379135,
      "loss": 2.7912,
      "theoretical_loss": 3.8267025012969347,
      "tokens_seen": 620197888
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004101303911735206,
      "loss": 2.8802,
      "theoretical_loss": 3.8266605549124515,
      "tokens_seen": 620263424
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00041012036108324976,
      "loss": 2.7613,
      "theoretical_loss": 3.8266186142005254,
      "tokens_seen": 620328960
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00041011033099297894,
      "loss": 2.8051,
      "theoretical_loss": 3.82657667915979,
      "tokens_seen": 620394496
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004101003009027081,
      "loss": 2.8462,
      "theoretical_loss": 3.8265347497888804,
      "tokens_seen": 620460032
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004100902708124373,
      "loss": 2.8898,
      "theoretical_loss": 3.8264928260864304,
      "tokens_seen": 620525568
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004100802407221665,
      "loss": 2.881,
      "theoretical_loss": 3.8264509080510756,
      "tokens_seen": 620591104
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004100702106318957,
      "loss": 2.6456,
      "theoretical_loss": 3.8264089956814513,
      "tokens_seen": 620656640
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00041006018054162485,
      "loss": 2.9374,
      "theoretical_loss": 3.826367088976194,
      "tokens_seen": 620722176
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004100501504513541,
      "loss": 2.9263,
      "theoretical_loss": 3.8263251879339393,
      "tokens_seen": 620787712
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004100401203610832,
      "loss": 2.9678,
      "theoretical_loss": 3.8262832925533257,
      "tokens_seen": 620853248
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00041003009027081245,
      "loss": 2.7197,
      "theoretical_loss": 3.826241402832989,
      "tokens_seen": 620918784
    },
    {
      "epoch": 2.01,
      "objective/train/docs_used": 774439,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7738590240478516,
      "objective/train/theoretical_loss": 3.8262204600949996,
      "objective/train/tokens_used": 641411552,
      "theoretical_loss": 3.8262204600949996,
      "tokens_seen": 620951552
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00041002006018054163,
      "loss": 2.8329,
      "theoretical_loss": 3.8261995187715687,
      "tokens_seen": 620984320
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004100100300902708,
      "loss": 2.9793,
      "theoretical_loss": 3.8261576403677027,
      "tokens_seen": 621049856
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00041,
      "loss": 2.8844,
      "theoretical_loss": 3.8261157676200295,
      "tokens_seen": 621115392
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040998996990972917,
      "loss": 2.8916,
      "theoretical_loss": 3.826073900527189,
      "tokens_seen": 621180928
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040997993981945835,
      "loss": 2.8177,
      "theoretical_loss": 3.8260320390878206,
      "tokens_seen": 621246464
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004099699097291876,
      "loss": 2.5451,
      "theoretical_loss": 3.8259901833005645,
      "tokens_seen": 621312000
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004099598796389167,
      "loss": 2.862,
      "theoretical_loss": 3.825948333164062,
      "tokens_seen": 621377536
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040994984954864595,
      "loss": 2.7519,
      "theoretical_loss": 3.8259064886769547,
      "tokens_seen": 621443072
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040993981945837513,
      "loss": 2.8865,
      "theoretical_loss": 3.825864649837883,
      "tokens_seen": 621508608
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004099297893681043,
      "loss": 2.8783,
      "theoretical_loss": 3.825822816645491,
      "tokens_seen": 621574144
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004099197592778335,
      "loss": 3.0324,
      "theoretical_loss": 3.8257809890984196,
      "tokens_seen": 621639680
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004099097291875627,
      "loss": 2.9964,
      "theoretical_loss": 3.8257391671953127,
      "tokens_seen": 621705216
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004098996990972919,
      "loss": 2.9455,
      "theoretical_loss": 3.825697350934814,
      "tokens_seen": 621770752
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004098896690070211,
      "loss": 2.9301,
      "theoretical_loss": 3.8256555403155668,
      "tokens_seen": 621836288
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004098796389167503,
      "loss": 2.7822,
      "theoretical_loss": 3.8256137353362174,
      "tokens_seen": 621901824
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040986960882647945,
      "loss": 2.723,
      "theoretical_loss": 3.8255719359954083,
      "tokens_seen": 621967360
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040985957873620864,
      "loss": 2.733,
      "theoretical_loss": 3.8255301422917873,
      "tokens_seen": 622032896
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004098495486459378,
      "loss": 2.935,
      "theoretical_loss": 3.8254883542239995,
      "tokens_seen": 622098432
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040983951855566705,
      "loss": 2.8432,
      "theoretical_loss": 3.8254465717906907,
      "tokens_seen": 622163968
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004098294884653962,
      "loss": 2.7378,
      "theoretical_loss": 3.8254047949905083,
      "tokens_seen": 622229504
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004098194583751254,
      "loss": 2.7812,
      "theoretical_loss": 3.8253630238221,
      "tokens_seen": 622295040
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004098094282848546,
      "loss": 2.9145,
      "theoretical_loss": 3.825321258284113,
      "tokens_seen": 622360576
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004097993981945838,
      "loss": 3.0614,
      "theoretical_loss": 3.825279498375196,
      "tokens_seen": 622426112
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040978936810431296,
      "loss": 2.9549,
      "theoretical_loss": 3.8252377440939975,
      "tokens_seen": 622491648
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040977933801404214,
      "loss": 2.8137,
      "theoretical_loss": 3.825195995439167,
      "tokens_seen": 622557184
    },
    {
      "epoch": 2.01,
      "objective/train/docs_used": 775020,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.917367696762085,
      "objective/train/theoretical_loss": 3.825175123221218,
      "objective/train/tokens_used": 643049952,
      "theoretical_loss": 3.825175123221218,
      "tokens_seen": 622589952
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004097693079237713,
      "loss": 2.9835,
      "theoretical_loss": 3.8251542524093542,
      "tokens_seen": 622622720
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040975927783350055,
      "loss": 2.8027,
      "theoretical_loss": 3.825112515003209,
      "tokens_seen": 622688256
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004097492477432297,
      "loss": 3.0546,
      "theoretical_loss": 3.825070783219382,
      "tokens_seen": 622753792
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004097392176529589,
      "loss": 2.7498,
      "theoretical_loss": 3.825029057056524,
      "tokens_seen": 622819328
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040972918756268804,
      "loss": 2.7878,
      "theoretical_loss": 3.8249873365132876,
      "tokens_seen": 622884864
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004097191574724173,
      "loss": 2.6325,
      "theoretical_loss": 3.8249456215883235,
      "tokens_seen": 622950400
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040970912738214646,
      "loss": 2.837,
      "theoretical_loss": 3.824903912280285,
      "tokens_seen": 623015936
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040969909729187564,
      "loss": 2.8094,
      "theoretical_loss": 3.8248622085878257,
      "tokens_seen": 623081472
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004096890672016048,
      "loss": 3.0344,
      "theoretical_loss": 3.824820510509597,
      "tokens_seen": 623147008
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.000409679037111334,
      "loss": 3.0147,
      "theoretical_loss": 3.824778818044255,
      "tokens_seen": 623212544
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004096690070210632,
      "loss": 2.8037,
      "theoretical_loss": 3.824737131190452,
      "tokens_seen": 623278080
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004096589769307924,
      "loss": 2.8736,
      "theoretical_loss": 3.8246954499468444,
      "tokens_seen": 623343616
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040964894684052155,
      "loss": 2.7638,
      "theoretical_loss": 3.824653774312087,
      "tokens_seen": 623409152
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004096389167502508,
      "loss": 3.0273,
      "theoretical_loss": 3.824612104284835,
      "tokens_seen": 623474688
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040962888665997996,
      "loss": 2.8466,
      "theoretical_loss": 3.824570439863746,
      "tokens_seen": 623540224
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040961885656970914,
      "loss": 3.077,
      "theoretical_loss": 3.8245287810474746,
      "tokens_seen": 623605760
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004096088264794383,
      "loss": 2.9727,
      "theoretical_loss": 3.8244871278346793,
      "tokens_seen": 623671296
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004095987963891675,
      "loss": 2.7729,
      "theoretical_loss": 3.8244454802240173,
      "tokens_seen": 623736832
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004095887662988967,
      "loss": 2.7011,
      "theoretical_loss": 3.8244038382141463,
      "tokens_seen": 623802368
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004095787362086259,
      "loss": 3.0728,
      "theoretical_loss": 3.824362201803726,
      "tokens_seen": 623867904
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040956870611835505,
      "loss": 2.7587,
      "theoretical_loss": 3.8243205709914143,
      "tokens_seen": 623933440
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004095586760280843,
      "loss": 2.7799,
      "theoretical_loss": 3.8242789457758706,
      "tokens_seen": 623998976
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004095486459378134,
      "loss": 2.6693,
      "theoretical_loss": 3.824237326155756,
      "tokens_seen": 624064512
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040953861584754265,
      "loss": 2.8832,
      "theoretical_loss": 3.824195712129729,
      "tokens_seen": 624130048
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040952858575727183,
      "loss": 2.7424,
      "theoretical_loss": 3.824154103696452,
      "tokens_seen": 624195584
    },
    {
      "epoch": 2.01,
      "objective/train/docs_used": 776436,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9514684677124023,
      "objective/train/theoretical_loss": 3.8241333015766763,
      "objective/train/tokens_used": 644688352,
      "theoretical_loss": 3.8241333015766763,
      "tokens_seen": 624228352
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.000409518555667001,
      "loss": 2.7882,
      "theoretical_loss": 3.8241125008545853,
      "tokens_seen": 624261120
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004095085255767302,
      "loss": 2.756,
      "theoretical_loss": 3.8240709036027916,
      "tokens_seen": 624326656
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040949849548645937,
      "loss": 2.9293,
      "theoretical_loss": 3.8240293119397317,
      "tokens_seen": 624392192
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040948846539618855,
      "loss": 2.7802,
      "theoretical_loss": 3.823987725864069,
      "tokens_seen": 624457728
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004094784353059178,
      "loss": 2.8242,
      "theoretical_loss": 3.8239461453744674,
      "tokens_seen": 624523264
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004094684052156469,
      "loss": 2.8764,
      "theoretical_loss": 3.8239045704695895,
      "tokens_seen": 624588800
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040945837512537615,
      "loss": 2.6924,
      "theoretical_loss": 3.8238630011480996,
      "tokens_seen": 624654336
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040944834503510533,
      "loss": 3.0381,
      "theoretical_loss": 3.823821437408662,
      "tokens_seen": 624719872
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004094383149448345,
      "loss": 2.821,
      "theoretical_loss": 3.8237798792499422,
      "tokens_seen": 624785408
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004094282848545637,
      "loss": 2.6496,
      "theoretical_loss": 3.823738326670605,
      "tokens_seen": 624850944
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004094182547642929,
      "loss": 2.7156,
      "theoretical_loss": 3.8236967796693166,
      "tokens_seen": 624916480
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040940822467402206,
      "loss": 2.9485,
      "theoretical_loss": 3.8236552382447435,
      "tokens_seen": 624982016
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004093981945837513,
      "loss": 2.8782,
      "theoretical_loss": 3.8236137023955523,
      "tokens_seen": 625047552
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004093881644934804,
      "loss": 2.8737,
      "theoretical_loss": 3.8235721721204103,
      "tokens_seen": 625113088
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040937813440320965,
      "loss": 2.7295,
      "theoretical_loss": 3.823530647417985,
      "tokens_seen": 625178624
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004093681043129388,
      "loss": 2.9309,
      "theoretical_loss": 3.823489128286945,
      "tokens_seen": 625244160
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.000409358074222668,
      "loss": 2.9655,
      "theoretical_loss": 3.8234476147259584,
      "tokens_seen": 625309696
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004093480441323972,
      "loss": 2.8918,
      "theoretical_loss": 3.8234061067336946,
      "tokens_seen": 625375232
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004093380140421264,
      "loss": 2.7062,
      "theoretical_loss": 3.823364604308823,
      "tokens_seen": 625440768
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040932798395185556,
      "loss": 2.7036,
      "theoretical_loss": 3.823323107450014,
      "tokens_seen": 625506304
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004093179538615848,
      "loss": 2.9356,
      "theoretical_loss": 3.8232816161559375,
      "tokens_seen": 625571840
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004093079237713139,
      "loss": 2.5131,
      "theoretical_loss": 3.8232401304252646,
      "tokens_seen": 625637376
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040929789368104316,
      "loss": 2.9493,
      "theoretical_loss": 3.823198650256667,
      "tokens_seen": 625702912
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004092878635907723,
      "loss": 3.0427,
      "theoretical_loss": 3.8231571756488165,
      "tokens_seen": 625768448
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004092778335005015,
      "loss": 2.8405,
      "theoretical_loss": 3.823115706600385,
      "tokens_seen": 625833984
    },
    {
      "epoch": 2.01,
      "objective/train/docs_used": 777260,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8199868202209473,
      "objective/train/theoretical_loss": 3.8230949741605365,
      "objective/train/tokens_used": 646326752,
      "theoretical_loss": 3.8230949741605365,
      "tokens_seen": 625866752
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004092678034102307,
      "loss": 2.7164,
      "theoretical_loss": 3.823074243110045,
      "tokens_seen": 625899520
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004092577733199599,
      "loss": 2.6414,
      "theoretical_loss": 3.82303278517647,
      "tokens_seen": 625965056
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040924774322968906,
      "loss": 2.8903,
      "theoretical_loss": 3.8229913327983343,
      "tokens_seen": 626030592
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040923771313941824,
      "loss": 2.7473,
      "theoretical_loss": 3.822949885974311,
      "tokens_seen": 626096128
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004092276830491474,
      "loss": 2.7383,
      "theoretical_loss": 3.8229084447030757,
      "tokens_seen": 626161664
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040921765295887666,
      "loss": 3.0065,
      "theoretical_loss": 3.8228670089833026,
      "tokens_seen": 626227200
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004092076228686058,
      "loss": 2.6521,
      "theoretical_loss": 3.822825578813667,
      "tokens_seen": 626292736
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.000409197592778335,
      "loss": 2.8301,
      "theoretical_loss": 3.8227841541928456,
      "tokens_seen": 626358272
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040918756268806415,
      "loss": 2.7502,
      "theoretical_loss": 3.8227427351195145,
      "tokens_seen": 626423808
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004091775325977934,
      "loss": 2.9458,
      "theoretical_loss": 3.82270132159235,
      "tokens_seen": 626489344
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040916750250752257,
      "loss": 2.6212,
      "theoretical_loss": 3.82265991361003,
      "tokens_seen": 626554880
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040915747241725175,
      "loss": 2.9773,
      "theoretical_loss": 3.8226185111712327,
      "tokens_seen": 626620416
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.000409147442326981,
      "loss": 2.9269,
      "theoretical_loss": 3.822577114274635,
      "tokens_seen": 626685952
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040913741223671016,
      "loss": 3.0878,
      "theoretical_loss": 3.822535722918916,
      "tokens_seen": 626751488
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040912738214643934,
      "loss": 2.9192,
      "theoretical_loss": 3.8224943371027558,
      "tokens_seen": 626817024
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004091173520561685,
      "loss": 2.4529,
      "theoretical_loss": 3.822452956824833,
      "tokens_seen": 626882560
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004091073219658977,
      "loss": 2.703,
      "theoretical_loss": 3.822411582083827,
      "tokens_seen": 626948096
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004090972918756269,
      "loss": 2.8391,
      "theoretical_loss": 3.82237021287842,
      "tokens_seen": 627013632
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004090872617853561,
      "loss": 2.7474,
      "theoretical_loss": 3.8223288492072918,
      "tokens_seen": 627079168
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040907723169508525,
      "loss": 2.6742,
      "theoretical_loss": 3.8222874910691234,
      "tokens_seen": 627144704
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004090672016048145,
      "loss": 2.6279,
      "theoretical_loss": 3.8222461384625976,
      "tokens_seen": 627210240
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004090571715145436,
      "loss": 2.8529,
      "theoretical_loss": 3.822204791386396,
      "tokens_seen": 627275776
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040904714142427285,
      "loss": 2.8604,
      "theoretical_loss": 3.822163449839202,
      "tokens_seen": 627341312
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040903711133400203,
      "loss": 2.97,
      "theoretical_loss": 3.822122113819698,
      "tokens_seen": 627406848
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004090270812437312,
      "loss": 2.714,
      "theoretical_loss": 3.8220807833265678,
      "tokens_seen": 627472384
    },
    {
      "epoch": 2.01,
      "objective/train/docs_used": 778767,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.658456563949585,
      "objective/train/theoretical_loss": 3.8220601201519817,
      "objective/train/tokens_used": 647965152,
      "theoretical_loss": 3.8220601201519817,
      "tokens_seen": 627505152
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004090170511534604,
      "loss": 2.9567,
      "theoretical_loss": 3.8220394583584953,
      "tokens_seen": 627537920
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040900702106318957,
      "loss": 2.6408,
      "theoretical_loss": 3.821998138914166,
      "tokens_seen": 627603456
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040899699097291875,
      "loss": 2.8496,
      "theoretical_loss": 3.821956824992264,
      "tokens_seen": 627668992
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.000408986960882648,
      "loss": 2.7743,
      "theoretical_loss": 3.821915516591475,
      "tokens_seen": 627734528
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004089769307923771,
      "loss": 2.6965,
      "theoretical_loss": 3.8218742137104846,
      "tokens_seen": 627800064
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040896690070210635,
      "loss": 2.8615,
      "theoretical_loss": 3.8218329163479794,
      "tokens_seen": 627865600
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040895687061183553,
      "loss": 2.5942,
      "theoretical_loss": 3.821791624502646,
      "tokens_seen": 627931136
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004089468405215647,
      "loss": 2.8906,
      "theoretical_loss": 3.821750338173172,
      "tokens_seen": 627996672
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004089368104312939,
      "loss": 2.9984,
      "theoretical_loss": 3.8217090573582446,
      "tokens_seen": 628062208
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004089267803410231,
      "loss": 2.8825,
      "theoretical_loss": 3.821667782056552,
      "tokens_seen": 628127744
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040891675025075226,
      "loss": 2.8492,
      "theoretical_loss": 3.821626512266784,
      "tokens_seen": 628193280
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004089067201604815,
      "loss": 2.7001,
      "theoretical_loss": 3.821585247987627,
      "tokens_seen": 628258816
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004088966900702106,
      "loss": 2.8747,
      "theoretical_loss": 3.8215439892177727,
      "tokens_seen": 628324352
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040888665997993985,
      "loss": 2.9203,
      "theoretical_loss": 3.82150273595591,
      "tokens_seen": 628389888
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.000408876629889669,
      "loss": 2.8152,
      "theoretical_loss": 3.8214614882007303,
      "tokens_seen": 628455424
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004088665997993982,
      "loss": 2.818,
      "theoretical_loss": 3.8214202459509234,
      "tokens_seen": 628520960
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004088565697091274,
      "loss": 2.9598,
      "theoretical_loss": 3.8213790092051805,
      "tokens_seen": 628586496
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004088465396188566,
      "loss": 2.8876,
      "theoretical_loss": 3.821337777962194,
      "tokens_seen": 628652032
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040883650952858576,
      "loss": 2.8699,
      "theoretical_loss": 3.8212965522206552,
      "tokens_seen": 628717568
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.000408826479438315,
      "loss": 2.866,
      "theoretical_loss": 3.821255331979258,
      "tokens_seen": 628783104
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004088164493480441,
      "loss": 3.1374,
      "theoretical_loss": 3.8212141172366945,
      "tokens_seen": 628848640
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040880641925777336,
      "loss": 2.736,
      "theoretical_loss": 3.8211729079916577,
      "tokens_seen": 628914176
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004087963891675025,
      "loss": 2.7657,
      "theoretical_loss": 3.821131704242843,
      "tokens_seen": 628979712
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004087863590772317,
      "loss": 2.918,
      "theoretical_loss": 3.821090505988944,
      "tokens_seen": 629045248
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004087763289869609,
      "loss": 2.7973,
      "theoretical_loss": 3.821049313228655,
      "tokens_seen": 629110784
    },
    {
      "epoch": 2.01,
      "objective/train/docs_used": 779594,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.082977056503296,
      "objective/train/theoretical_loss": 3.821028718908207,
      "objective/train/tokens_used": 649603552,
      "theoretical_loss": 3.821028718908207,
      "tokens_seen": 629143552
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004087662988966901,
      "loss": 2.8122,
      "theoretical_loss": 3.821008125960672,
      "tokens_seen": 629176320
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040875626880641926,
      "loss": 2.943,
      "theoretical_loss": 3.8209669441836907,
      "tokens_seen": 629241856
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040874623871614844,
      "loss": 2.8359,
      "theoretical_loss": 3.8209257678964073,
      "tokens_seen": 629307392
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004087362086258776,
      "loss": 2.7174,
      "theoretical_loss": 3.8208845970975176,
      "tokens_seen": 629372928
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040872617853560686,
      "loss": 2.9775,
      "theoretical_loss": 3.82084343178572,
      "tokens_seen": 629438464
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.000408716148445336,
      "loss": 2.5463,
      "theoretical_loss": 3.8208022719597112,
      "tokens_seen": 629504000
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004087061183550652,
      "loss": 2.9226,
      "theoretical_loss": 3.820761117618189,
      "tokens_seen": 629569536
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040869608826479435,
      "loss": 2.774,
      "theoretical_loss": 3.8207199687598523,
      "tokens_seen": 629635072
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004086860581745236,
      "loss": 2.56,
      "theoretical_loss": 3.8206788253834,
      "tokens_seen": 629700608
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040867602808425277,
      "loss": 2.6942,
      "theoretical_loss": 3.8206376874875305,
      "tokens_seen": 629766144
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040866599799398195,
      "loss": 2.6446,
      "theoretical_loss": 3.8205965550709444,
      "tokens_seen": 629831680
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040865596790371113,
      "loss": 2.6938,
      "theoretical_loss": 3.8205554281323417,
      "tokens_seen": 629897216
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040864593781344036,
      "loss": 2.8075,
      "theoretical_loss": 3.8205143066704226,
      "tokens_seen": 629962752
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004086359077231695,
      "loss": 2.742,
      "theoretical_loss": 3.820473190683889,
      "tokens_seen": 630028288
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004086258776328987,
      "loss": 2.8813,
      "theoretical_loss": 3.8204320801714413,
      "tokens_seen": 630093824
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040861584754262785,
      "loss": 2.9127,
      "theoretical_loss": 3.820390975131782,
      "tokens_seen": 630159360
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004086058174523571,
      "loss": 2.9052,
      "theoretical_loss": 3.8203498755636143,
      "tokens_seen": 630224896
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040859578736208627,
      "loss": 2.7453,
      "theoretical_loss": 3.8203087814656396,
      "tokens_seen": 630290432
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040858575727181545,
      "loss": 2.8833,
      "theoretical_loss": 3.820267692836562,
      "tokens_seen": 630355968
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040857572718154463,
      "loss": 2.9168,
      "theoretical_loss": 3.8202266096750854,
      "tokens_seen": 630421504
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004085656970912738,
      "loss": 2.7973,
      "theoretical_loss": 3.8201855319799134,
      "tokens_seen": 630487040
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.000408555667001003,
      "loss": 2.6587,
      "theoretical_loss": 3.8201444597497507,
      "tokens_seen": 630552576
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040854563691073223,
      "loss": 3.037,
      "theoretical_loss": 3.8201033929833024,
      "tokens_seen": 630618112
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040853560682046136,
      "loss": 2.8116,
      "theoretical_loss": 3.8200623316792743,
      "tokens_seen": 630683648
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004085255767301906,
      "loss": 2.8274,
      "theoretical_loss": 3.820021275836372,
      "tokens_seen": 630749184
    },
    {
      "epoch": 2.01,
      "objective/train/docs_used": 780765,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5284388065338135,
      "objective/train/theoretical_loss": 3.820000749962438,
      "objective/train/tokens_used": 651241952,
      "theoretical_loss": 3.820000749962438,
      "tokens_seen": 630781952
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004085155466399197,
      "loss": 2.7464,
      "theoretical_loss": 3.8199802254533015,
      "tokens_seen": 630814720
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040850551654964895,
      "loss": 2.7889,
      "theoretical_loss": 3.8199391805287704,
      "tokens_seen": 630880256
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040849548645937814,
      "loss": 2.7819,
      "theoretical_loss": 3.8198981410614854,
      "tokens_seen": 630945792
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004084854563691073,
      "loss": 2.7129,
      "theoretical_loss": 3.819857107050155,
      "tokens_seen": 631011328
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004084754262788365,
      "loss": 2.8668,
      "theoretical_loss": 3.819816078493486,
      "tokens_seen": 631076864
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040846539618856573,
      "loss": 2.7094,
      "theoretical_loss": 3.819775055390188,
      "tokens_seen": 631142400
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040845536609829486,
      "loss": 3.036,
      "theoretical_loss": 3.8197340377389697,
      "tokens_seen": 631207936
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004084453360080241,
      "loss": 2.8579,
      "theoretical_loss": 3.81969302553854,
      "tokens_seen": 631273472
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004084353059177532,
      "loss": 2.7545,
      "theoretical_loss": 3.81965201878761,
      "tokens_seen": 631339008
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040842527582748246,
      "loss": 2.9343,
      "theoretical_loss": 3.819611017484889,
      "tokens_seen": 631404544
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040841524573721164,
      "loss": 2.9302,
      "theoretical_loss": 3.8195700216290884,
      "tokens_seen": 631470080
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004084052156469408,
      "loss": 2.6115,
      "theoretical_loss": 3.8195290312189187,
      "tokens_seen": 631535616
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040839518555667005,
      "loss": 2.9017,
      "theoretical_loss": 3.8194880462530922,
      "tokens_seen": 631601152
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004083851554663992,
      "loss": 2.8711,
      "theoretical_loss": 3.819447066730321,
      "tokens_seen": 631666688
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004083751253761284,
      "loss": 2.7859,
      "theoretical_loss": 3.8194060926493165,
      "tokens_seen": 631732224
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004083650952858576,
      "loss": 2.9598,
      "theoretical_loss": 3.819365124008793,
      "tokens_seen": 631797760
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004083550651955868,
      "loss": 2.6389,
      "theoretical_loss": 3.8193241608074637,
      "tokens_seen": 631863296
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040834503510531596,
      "loss": 2.9403,
      "theoretical_loss": 3.8192832030440416,
      "tokens_seen": 631928832
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004083350050150452,
      "loss": 2.9505,
      "theoretical_loss": 3.819242250717242,
      "tokens_seen": 631994368
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004083249749247743,
      "loss": 2.7696,
      "theoretical_loss": 3.819201303825779,
      "tokens_seen": 632059904
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040831494483450356,
      "loss": 2.9617,
      "theoretical_loss": 3.8191603623683674,
      "tokens_seen": 632125440
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004083049147442327,
      "loss": 2.7968,
      "theoretical_loss": 3.819119426343724,
      "tokens_seen": 632190976
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004082948846539619,
      "loss": 2.6314,
      "theoretical_loss": 3.8190784957505635,
      "tokens_seen": 632256512
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004082848545636911,
      "loss": 2.9328,
      "theoretical_loss": 3.819037570587603,
      "tokens_seen": 632322048
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004082748244734203,
      "loss": 2.8324,
      "theoretical_loss": 3.8189966508535598,
      "tokens_seen": 632387584
    },
    {
      "epoch": 2.01,
      "objective/train/docs_used": 781551,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.776937246322632,
      "objective/train/theoretical_loss": 3.8189761930219808,
      "objective/train/tokens_used": 652880352,
      "theoretical_loss": 3.8189761930219808,
      "tokens_seen": 632420352
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040826479438314946,
      "loss": 2.8414,
      "theoretical_loss": 3.8189557365471503,
      "tokens_seen": 632453120
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040825476429287864,
      "loss": 3.0336,
      "theoretical_loss": 3.8189148276670926,
      "tokens_seen": 632518656
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004082447342026078,
      "loss": 2.9431,
      "theoretical_loss": 3.818873924212105,
      "tokens_seen": 632584192
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040823470411233706,
      "loss": 2.8801,
      "theoretical_loss": 3.818833026180906,
      "tokens_seen": 632649728
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004082246740220662,
      "loss": 2.7998,
      "theoretical_loss": 3.8187921335722157,
      "tokens_seen": 632715264
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004082146439317954,
      "loss": 2.8741,
      "theoretical_loss": 3.8187512463847515,
      "tokens_seen": 632780800
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040820461384152455,
      "loss": 2.6709,
      "theoretical_loss": 3.8187103646172353,
      "tokens_seen": 632846336
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004081945837512538,
      "loss": 2.939,
      "theoretical_loss": 3.8186694882683865,
      "tokens_seen": 632911872
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040818455366098297,
      "loss": 2.9783,
      "theoretical_loss": 3.818628617336926,
      "tokens_seen": 632977408
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040817452357071215,
      "loss": 2.9095,
      "theoretical_loss": 3.8185877518215756,
      "tokens_seen": 633042944
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040816449348044133,
      "loss": 2.9708,
      "theoretical_loss": 3.818546891721056,
      "tokens_seen": 633108480
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040815446339017056,
      "loss": 2.9867,
      "theoretical_loss": 3.81850603703409,
      "tokens_seen": 633174016
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004081444332998997,
      "loss": 2.8313,
      "theoretical_loss": 3.8184651877594007,
      "tokens_seen": 633239552
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004081344032096289,
      "loss": 2.9659,
      "theoretical_loss": 3.8184243438957095,
      "tokens_seen": 633305088
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040812437311935805,
      "loss": 2.8979,
      "theoretical_loss": 3.8183835054417408,
      "tokens_seen": 633370624
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004081143430290873,
      "loss": 2.7882,
      "theoretical_loss": 3.818342672396219,
      "tokens_seen": 633436160
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040810431293881647,
      "loss": 2.8737,
      "theoretical_loss": 3.818301844757867,
      "tokens_seen": 633501696
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040809428284854565,
      "loss": 2.8172,
      "theoretical_loss": 3.8182610225254106,
      "tokens_seen": 633567232
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040808425275827483,
      "loss": 2.8925,
      "theoretical_loss": 3.8182202056975747,
      "tokens_seen": 633632768
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.000408074222668004,
      "loss": 2.6774,
      "theoretical_loss": 3.8181793942730846,
      "tokens_seen": 633698304
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004080641925777332,
      "loss": 2.7061,
      "theoretical_loss": 3.8181385882506667,
      "tokens_seen": 633763840
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040805416248746243,
      "loss": 3.0322,
      "theoretical_loss": 3.8180977876290467,
      "tokens_seen": 633829376
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040804413239719156,
      "loss": 2.8356,
      "theoretical_loss": 3.8180569924069525,
      "tokens_seen": 633894912
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004080341023069208,
      "loss": 2.8216,
      "theoretical_loss": 3.8180162025831104,
      "tokens_seen": 633960448
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004080240722166499,
      "loss": 2.7041,
      "theoretical_loss": 3.817975418156249,
      "tokens_seen": 634025984
    },
    {
      "epoch": 2.01,
      "objective/train/docs_used": 783036,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.2487502098083496,
      "objective/train/theoretical_loss": 3.8179550279662884,
      "objective/train/tokens_used": 654518752,
      "theoretical_loss": 3.8179550279662884,
      "tokens_seen": 634058752
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040801404212637915,
      "loss": 2.9766,
      "theoretical_loss": 3.817934639125096,
      "tokens_seen": 634091520
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040800401203610834,
      "loss": 2.7949,
      "theoretical_loss": 3.8178938654883803,
      "tokens_seen": 634157056
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004079939819458375,
      "loss": 2.6729,
      "theoretical_loss": 3.8178530972448304,
      "tokens_seen": 634222592
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004079839518555667,
      "loss": 2.7617,
      "theoretical_loss": 3.8178123343931762,
      "tokens_seen": 634288128
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040797392176529593,
      "loss": 2.6544,
      "theoretical_loss": 3.8177715769321474,
      "tokens_seen": 634353664
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040796389167502506,
      "loss": 2.9967,
      "theoretical_loss": 3.817730824860474,
      "tokens_seen": 634419200
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004079538615847543,
      "loss": 2.7682,
      "theoretical_loss": 3.817690078176888,
      "tokens_seen": 634484736
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004079438314944834,
      "loss": 2.7531,
      "theoretical_loss": 3.817649336880119,
      "tokens_seen": 634550272
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040793380140421266,
      "loss": 2.7709,
      "theoretical_loss": 3.8176086009689,
      "tokens_seen": 634615808
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040792377131394184,
      "loss": 2.7165,
      "theoretical_loss": 3.817567870441962,
      "tokens_seen": 634681344
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.000407913741223671,
      "loss": 2.8137,
      "theoretical_loss": 3.8175271452980377,
      "tokens_seen": 634746880
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004079037111334002,
      "loss": 2.9933,
      "theoretical_loss": 3.8174864255358605,
      "tokens_seen": 634812416
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004078936810431294,
      "loss": 2.8765,
      "theoretical_loss": 3.8174457111541633,
      "tokens_seen": 634877952
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040788365095285856,
      "loss": 2.8943,
      "theoretical_loss": 3.8174050021516797,
      "tokens_seen": 634943488
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004078736208625878,
      "loss": 2.979,
      "theoretical_loss": 3.8173642985271443,
      "tokens_seen": 635009024
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004078635907723169,
      "loss": 2.7522,
      "theoretical_loss": 3.817323600279291,
      "tokens_seen": 635074560
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040785356068204616,
      "loss": 2.7643,
      "theoretical_loss": 3.817282907406857,
      "tokens_seen": 635140096
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004078435305917753,
      "loss": 2.9688,
      "theoretical_loss": 3.8172422199085743,
      "tokens_seen": 635205632
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004078335005015045,
      "loss": 2.7879,
      "theoretical_loss": 3.817201537783182,
      "tokens_seen": 635271168
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004078234704112337,
      "loss": 2.8632,
      "theoretical_loss": 3.8171608610294143,
      "tokens_seen": 635336704
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004078134403209629,
      "loss": 2.8936,
      "theoretical_loss": 3.8171201896460083,
      "tokens_seen": 635402240
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040780341023069207,
      "loss": 2.8438,
      "theoretical_loss": 3.817079523631703,
      "tokens_seen": 635467776
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004077933801404213,
      "loss": 2.9047,
      "theoretical_loss": 3.8170388629852336,
      "tokens_seen": 635533312
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040778335005015043,
      "loss": 2.8102,
      "theoretical_loss": 3.8169982077053395,
      "tokens_seen": 635598848
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040777331995987966,
      "loss": 2.7813,
      "theoretical_loss": 3.816957557790759,
      "tokens_seen": 635664384
    },
    {
      "epoch": 2.01,
      "objective/train/docs_used": 783857,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6842525005340576,
      "objective/train/theoretical_loss": 3.8169372348450663,
      "objective/train/tokens_used": 656157152,
      "theoretical_loss": 3.8169372348450663,
      "tokens_seen": 635697152
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004077632898696088,
      "loss": 2.8004,
      "theoretical_loss": 3.81691691324023,
      "tokens_seen": 635729920
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.000407753259779338,
      "loss": 2.8361,
      "theoretical_loss": 3.8168762740524933,
      "tokens_seen": 635795456
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004077432296890672,
      "loss": 3.1555,
      "theoretical_loss": 3.8168356402262877,
      "tokens_seen": 635860992
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004077331995987964,
      "loss": 2.8418,
      "theoretical_loss": 3.816795011760354,
      "tokens_seen": 635926528
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040772316950852557,
      "loss": 2.7288,
      "theoretical_loss": 3.816754388653432,
      "tokens_seen": 635992064
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040771313941825475,
      "loss": 2.8514,
      "theoretical_loss": 3.816713770904263,
      "tokens_seen": 636057600
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040770310932798393,
      "loss": 2.8342,
      "theoretical_loss": 3.8166731585115894,
      "tokens_seen": 636123136
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040769307923771317,
      "loss": 2.6405,
      "theoretical_loss": 3.816632551474152,
      "tokens_seen": 636188672
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004076830491474423,
      "loss": 2.8957,
      "theoretical_loss": 3.816591949790693,
      "tokens_seen": 636254208
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040767301905717153,
      "loss": 2.6612,
      "theoretical_loss": 3.816551353459955,
      "tokens_seen": 636319744
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004076629889669007,
      "loss": 3.0216,
      "theoretical_loss": 3.816510762480683,
      "tokens_seen": 636385280
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004076529588766299,
      "loss": 2.8275,
      "theoretical_loss": 3.8164701768516185,
      "tokens_seen": 636450816
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004076429287863591,
      "loss": 2.9587,
      "theoretical_loss": 3.816429596571506,
      "tokens_seen": 636516352
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040763289869608825,
      "loss": 2.7802,
      "theoretical_loss": 3.81638902163909,
      "tokens_seen": 636581888
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004076228686058175,
      "loss": 2.867,
      "theoretical_loss": 3.8163484520531155,
      "tokens_seen": 636647424
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040761283851554667,
      "loss": 2.5608,
      "theoretical_loss": 3.8163078878123278,
      "tokens_seen": 636712960
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040760280842527585,
      "loss": 2.7673,
      "theoretical_loss": 3.8162673289154725,
      "tokens_seen": 636778496
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040759277833500503,
      "loss": 2.8313,
      "theoretical_loss": 3.8162267753612964,
      "tokens_seen": 636844032
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004075827482447342,
      "loss": 2.911,
      "theoretical_loss": 3.8161862271485445,
      "tokens_seen": 636909568
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004075727181544634,
      "loss": 2.8002,
      "theoretical_loss": 3.8161456842759645,
      "tokens_seen": 636975104
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040756268806419263,
      "loss": 2.8853,
      "theoretical_loss": 3.816105146742304,
      "tokens_seen": 637040640
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040755265797392176,
      "loss": 2.6506,
      "theoretical_loss": 3.816064614546311,
      "tokens_seen": 637106176
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.000407542627883651,
      "loss": 3.0013,
      "theoretical_loss": 3.816024087686733,
      "tokens_seen": 637171712
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004075325977933801,
      "loss": 2.8045,
      "theoretical_loss": 3.81598356616232,
      "tokens_seen": 637237248
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040752256770310935,
      "loss": 2.9861,
      "theoretical_loss": 3.8159430499718194,
      "tokens_seen": 637302784
    },
    {
      "epoch": 2.01,
      "objective/train/docs_used": 785359,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.857738494873047,
      "objective/train/theoretical_loss": 3.8159227938763953,
      "objective/train/tokens_used": 657795552,
      "theoretical_loss": 3.8159227938763953,
      "tokens_seen": 637335552
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040751253761283854,
      "loss": 2.8626,
      "theoretical_loss": 3.815902539113981,
      "tokens_seen": 637368320
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004075025075225677,
      "loss": 2.8335,
      "theoretical_loss": 3.815862033587556,
      "tokens_seen": 637433856
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004074924774322969,
      "loss": 2.9154,
      "theoretical_loss": 3.8158215333912935,
      "tokens_seen": 637499392
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040748244734202613,
      "loss": 2.9701,
      "theoretical_loss": 3.8157810385239443,
      "tokens_seen": 637564928
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040747241725175526,
      "loss": 2.7581,
      "theoretical_loss": 3.81574054898426,
      "tokens_seen": 637630464
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004074623871614845,
      "loss": 3.1263,
      "theoretical_loss": 3.8157000647709927,
      "tokens_seen": 637696000
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004074523570712136,
      "loss": 2.6475,
      "theoretical_loss": 3.8156595858828934,
      "tokens_seen": 637761536
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040744232698094286,
      "loss": 2.9252,
      "theoretical_loss": 3.815619112318715,
      "tokens_seen": 637827072
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040743229689067204,
      "loss": 2.8951,
      "theoretical_loss": 3.8155786440772106,
      "tokens_seen": 637892608
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004074222668004012,
      "loss": 3.0481,
      "theoretical_loss": 3.8155381811571325,
      "tokens_seen": 637958144
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004074122367101304,
      "loss": 2.9415,
      "theoretical_loss": 3.8154977235572356,
      "tokens_seen": 638023680
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004074022066198596,
      "loss": 2.789,
      "theoretical_loss": 3.8154572712762733,
      "tokens_seen": 638089216
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040739217652958876,
      "loss": 2.7131,
      "theoretical_loss": 3.815416824313001,
      "tokens_seen": 638154752
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.000407382146439318,
      "loss": 2.7745,
      "theoretical_loss": 3.815376382666172,
      "tokens_seen": 638220288
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004073721163490471,
      "loss": 2.9315,
      "theoretical_loss": 3.8153359463345433,
      "tokens_seen": 638285824
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040736208625877636,
      "loss": 2.6523,
      "theoretical_loss": 3.8152955153168704,
      "tokens_seen": 638351360
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004073520561685055,
      "loss": 3.1046,
      "theoretical_loss": 3.8152550896119086,
      "tokens_seen": 638416896
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004073420260782347,
      "loss": 2.9757,
      "theoretical_loss": 3.815214669218416,
      "tokens_seen": 638482432
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004073319959879639,
      "loss": 3.0314,
      "theoretical_loss": 3.8151742541351483,
      "tokens_seen": 638547968
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004073219658976931,
      "loss": 3.0917,
      "theoretical_loss": 3.8151338443608633,
      "tokens_seen": 638613504
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040731193580742227,
      "loss": 2.6262,
      "theoretical_loss": 3.8150934398943193,
      "tokens_seen": 638679040
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004073019057171515,
      "loss": 2.7144,
      "theoretical_loss": 3.8150530407342744,
      "tokens_seen": 638744576
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040729187562688063,
      "loss": 2.8772,
      "theoretical_loss": 3.815012646879487,
      "tokens_seen": 638810112
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040728184553660986,
      "loss": 2.7468,
      "theoretical_loss": 3.814972258328717,
      "tokens_seen": 638875648
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.000407271815446339,
      "loss": 2.6933,
      "theoretical_loss": 3.8149318750807235,
      "tokens_seen": 638941184
    },
    {
      "epoch": 2.01,
      "objective/train/docs_used": 786054,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.3852269649505615,
      "objective/train/theoretical_loss": 3.814911685444881,
      "objective/train/tokens_used": 659433952,
      "theoretical_loss": 3.814911685444881,
      "tokens_seen": 638973952
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004072617853560682,
      "loss": 2.5543,
      "theoretical_loss": 3.8148914971342665,
      "tokens_seen": 639006720
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004072517552657974,
      "loss": 2.8611,
      "theoretical_loss": 3.814851124488106,
      "tokens_seen": 639072256
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004072417251755266,
      "loss": 2.774,
      "theoretical_loss": 3.8148107571410037,
      "tokens_seen": 639137792
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040723169508525577,
      "loss": 2.7506,
      "theoretical_loss": 3.8147703950917204,
      "tokens_seen": 639203328
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040722166499498495,
      "loss": 2.7369,
      "theoretical_loss": 3.8147300383390172,
      "tokens_seen": 639268864
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040721163490471413,
      "loss": 2.4588,
      "theoretical_loss": 3.814689686881657,
      "tokens_seen": 639334400
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040720160481444337,
      "loss": 2.7529,
      "theoretical_loss": 3.814649340718402,
      "tokens_seen": 639399936
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004071915747241725,
      "loss": 2.7639,
      "theoretical_loss": 3.814608999848015,
      "tokens_seen": 639465472
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040718154463390173,
      "loss": 2.9076,
      "theoretical_loss": 3.814568664269259,
      "tokens_seen": 639531008
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004071715145436309,
      "loss": 2.7979,
      "theoretical_loss": 3.8145283339808986,
      "tokens_seen": 639596544
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004071614844533601,
      "loss": 2.7578,
      "theoretical_loss": 3.814488008981697,
      "tokens_seen": 639662080
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040715145436308927,
      "loss": 3.0341,
      "theoretical_loss": 3.8144476892704198,
      "tokens_seen": 639727616
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040714142427281845,
      "loss": 2.897,
      "theoretical_loss": 3.8144073748458305,
      "tokens_seen": 639793152
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040713139418254763,
      "loss": 2.9447,
      "theoretical_loss": 3.814367065706696,
      "tokens_seen": 639858688
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040712136409227687,
      "loss": 2.8507,
      "theoretical_loss": 3.814326761851781,
      "tokens_seen": 639924224
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.000407111334002006,
      "loss": 2.7963,
      "theoretical_loss": 3.8142864632798523,
      "tokens_seen": 639989760
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040710130391173523,
      "loss": 2.8708,
      "theoretical_loss": 3.8142461699896764,
      "tokens_seen": 640055296
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040709127382146436,
      "loss": 2.8233,
      "theoretical_loss": 3.8142058819800204,
      "tokens_seen": 640120832
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004070812437311936,
      "loss": 3.0831,
      "theoretical_loss": 3.8141655992496517,
      "tokens_seen": 640186368
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004070712136409228,
      "loss": 2.7594,
      "theoretical_loss": 3.814125321797338,
      "tokens_seen": 640251904
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040706118355065196,
      "loss": 2.5152,
      "theoretical_loss": 3.814085049621848,
      "tokens_seen": 640317440
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040705115346038114,
      "loss": 2.7952,
      "theoretical_loss": 3.81404478272195,
      "tokens_seen": 640382976
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004070411233701103,
      "loss": 2.7167,
      "theoretical_loss": 3.8140045210964133,
      "tokens_seen": 640448512
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004070310932798395,
      "loss": 2.8056,
      "theoretical_loss": 3.8139642647440075,
      "tokens_seen": 640514048
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040702106318956874,
      "loss": 2.6948,
      "theoretical_loss": 3.8139240136635024,
      "tokens_seen": 640579584
    },
    {
      "epoch": 2.01,
      "objective/train/docs_used": 786648,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7099239826202393,
      "objective/train/theoretical_loss": 3.8139038900998283,
      "objective/train/tokens_used": 661072352,
      "theoretical_loss": 3.8139038900998283,
      "tokens_seen": 640612352
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040701103309929786,
      "loss": 2.7378,
      "theoretical_loss": 3.8138837678536683,
      "tokens_seen": 640645120
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004070010030090271,
      "loss": 2.8674,
      "theoretical_loss": 3.813843527313276,
      "tokens_seen": 640710656
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004069909729187563,
      "loss": 2.786,
      "theoretical_loss": 3.813803292041097,
      "tokens_seen": 640776192
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040698094282848546,
      "loss": 2.9161,
      "theoretical_loss": 3.8137630620359024,
      "tokens_seen": 640841728
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040697091273821464,
      "loss": 2.8332,
      "theoretical_loss": 3.8137228372964644,
      "tokens_seen": 640907264
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004069608826479438,
      "loss": 2.754,
      "theoretical_loss": 3.8136826178215557,
      "tokens_seen": 640972800
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.000406950852557673,
      "loss": 2.7697,
      "theoretical_loss": 3.8136424036099488,
      "tokens_seen": 641038336
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040694082246740224,
      "loss": 2.7057,
      "theoretical_loss": 3.8136021946604166,
      "tokens_seen": 641103872
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040693079237713137,
      "loss": 2.8738,
      "theoretical_loss": 3.813561990971734,
      "tokens_seen": 641169408
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004069207622868606,
      "loss": 2.8105,
      "theoretical_loss": 3.8135217925426734,
      "tokens_seen": 641234944
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040691073219658973,
      "loss": 2.8888,
      "theoretical_loss": 3.81348159937201,
      "tokens_seen": 641300480
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040690070210631896,
      "loss": 2.8901,
      "theoretical_loss": 3.81344141145852,
      "tokens_seen": 641366016
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004068906720160482,
      "loss": 2.7657,
      "theoretical_loss": 3.813401228800976,
      "tokens_seen": 641431552
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004068806419257773,
      "loss": 2.6503,
      "theoretical_loss": 3.8133610513981555,
      "tokens_seen": 641497088
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040687061183550656,
      "loss": 2.725,
      "theoretical_loss": 3.8133208792488347,
      "tokens_seen": 641562624
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004068605817452357,
      "loss": 2.6084,
      "theoretical_loss": 3.8132807123517893,
      "tokens_seen": 641628160
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004068505516549649,
      "loss": 2.8921,
      "theoretical_loss": 3.813240550705797,
      "tokens_seen": 641693696
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004068405215646941,
      "loss": 2.7129,
      "theoretical_loss": 3.813200394309635,
      "tokens_seen": 641759232
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004068304914744233,
      "loss": 2.7996,
      "theoretical_loss": 3.8131602431620797,
      "tokens_seen": 641824768
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040682046138415247,
      "loss": 2.9959,
      "theoretical_loss": 3.8131200972619115,
      "tokens_seen": 641890304
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004068104312938817,
      "loss": 2.6639,
      "theoretical_loss": 3.8130799566079077,
      "tokens_seen": 641955840
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040680040120361083,
      "loss": 2.751,
      "theoretical_loss": 3.813039821198847,
      "tokens_seen": 642021376
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040679037111334006,
      "loss": 2.7787,
      "theoretical_loss": 3.8129996910335096,
      "tokens_seen": 642086912
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004067803410230692,
      "loss": 2.8948,
      "theoretical_loss": 3.8129595661106745,
      "tokens_seen": 642152448
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004067703109327984,
      "loss": 2.8215,
      "theoretical_loss": 3.8129194464291225,
      "tokens_seen": 642217984
    },
    {
      "epoch": 2.01,
      "objective/train/docs_used": 788043,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.803264856338501,
      "objective/train/theoretical_loss": 3.8128993885534466,
      "objective/train/tokens_used": 662710752,
      "theoretical_loss": 3.8128993885534466,
      "tokens_seen": 642250752
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004067602808425276,
      "loss": 2.762,
      "theoretical_loss": 3.812879331987634,
      "tokens_seen": 642283520
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004067502507522568,
      "loss": 2.7921,
      "theoretical_loss": 3.8128392227849903,
      "tokens_seen": 642349056
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040674022066198597,
      "loss": 2.869,
      "theoretical_loss": 3.8127991188199726,
      "tokens_seen": 642414592
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040673019057171515,
      "loss": 2.7139,
      "theoretical_loss": 3.812759020091362,
      "tokens_seen": 642480128
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040672016048144433,
      "loss": 2.9997,
      "theoretical_loss": 3.812718926597942,
      "tokens_seen": 642545664
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040671013039117357,
      "loss": 2.9864,
      "theoretical_loss": 3.8126788383384946,
      "tokens_seen": 642611200
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004067001003009027,
      "loss": 2.9218,
      "theoretical_loss": 3.812638755311803,
      "tokens_seen": 642676736
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040669007021063193,
      "loss": 2.8599,
      "theoretical_loss": 3.812598677516651,
      "tokens_seen": 642742272
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004066800401203611,
      "loss": 2.8809,
      "theoretical_loss": 3.812558604951822,
      "tokens_seen": 642807808
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.0004066700100300903,
      "loss": 2.9374,
      "theoretical_loss": 3.8125185376161,
      "tokens_seen": 642873344
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040665997993981947,
      "loss": 2.9225,
      "theoretical_loss": 3.8124784755082706,
      "tokens_seen": 642938880
    },
    {
      "epoch": 2.01,
      "learning_rate": 0.00040664994984954865,
      "loss": 2.8897,
      "theoretical_loss": 3.8124384186271176,
      "tokens_seen": 643004416
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040663991975927783,
      "loss": 2.67,
      "theoretical_loss": 3.812398366971428,
      "tokens_seen": 643069952
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040662988966900707,
      "loss": 2.7487,
      "theoretical_loss": 3.8123583205399862,
      "tokens_seen": 643135488
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004066198595787362,
      "loss": 2.8362,
      "theoretical_loss": 3.8123182793315804,
      "tokens_seen": 643201024
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040660982948846543,
      "loss": 3.058,
      "theoretical_loss": 3.8122782433449953,
      "tokens_seen": 643266560
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040659979939819456,
      "loss": 2.9504,
      "theoretical_loss": 3.812238212579019,
      "tokens_seen": 643332096
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004065897693079238,
      "loss": 2.9792,
      "theoretical_loss": 3.8121981870324393,
      "tokens_seen": 643397632
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.000406579739217653,
      "loss": 2.7157,
      "theoretical_loss": 3.8121581667040436,
      "tokens_seen": 643463168
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040656970912738216,
      "loss": 2.6707,
      "theoretical_loss": 3.81211815159262,
      "tokens_seen": 643528704
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040655967903711134,
      "loss": 2.7079,
      "theoretical_loss": 3.8120781416969587,
      "tokens_seen": 643594240
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004065496489468405,
      "loss": 2.9709,
      "theoretical_loss": 3.812038137015847,
      "tokens_seen": 643659776
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004065396188565697,
      "loss": 2.7942,
      "theoretical_loss": 3.811998137548075,
      "tokens_seen": 643725312
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040652958876629894,
      "loss": 3.1165,
      "theoretical_loss": 3.8119581432924337,
      "tokens_seen": 643790848
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040651955867602806,
      "loss": 2.6032,
      "theoretical_loss": 3.811918154247712,
      "tokens_seen": 643856384
    },
    {
      "epoch": 2.02,
      "objective/train/docs_used": 789663,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.804586172103882,
      "objective/train/theoretical_loss": 3.8118981616790686,
      "objective/train/tokens_used": 664349152,
      "theoretical_loss": 3.8118981616790686,
      "tokens_seen": 643889152
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004065095285857573,
      "loss": 2.8536,
      "theoretical_loss": 3.811878170412702,
      "tokens_seen": 643921920
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004064994984954865,
      "loss": 2.9554,
      "theoretical_loss": 3.811838191786193,
      "tokens_seen": 643987456
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040648946840521566,
      "loss": 2.9039,
      "theoretical_loss": 3.8117982183669787,
      "tokens_seen": 644052992
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040647943831494484,
      "loss": 2.6216,
      "theoretical_loss": 3.81175825015385,
      "tokens_seen": 644118528
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.000406469408224674,
      "loss": 2.8063,
      "theoretical_loss": 3.8117182871455997,
      "tokens_seen": 644184064
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004064593781344032,
      "loss": 2.8755,
      "theoretical_loss": 3.81167832934102,
      "tokens_seen": 644249600
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040644934804413244,
      "loss": 2.7333,
      "theoretical_loss": 3.811638376738904,
      "tokens_seen": 644315136
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040643931795386157,
      "loss": 2.8455,
      "theoretical_loss": 3.811598429338046,
      "tokens_seen": 644380672
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004064292878635908,
      "loss": 3.0777,
      "theoretical_loss": 3.8115584871372397,
      "tokens_seen": 644446208
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040641925777331993,
      "loss": 2.8559,
      "theoretical_loss": 3.811518550135279,
      "tokens_seen": 644511744
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040640922768304916,
      "loss": 3.0454,
      "theoretical_loss": 3.811478618330959,
      "tokens_seen": 644577280
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040639919759277834,
      "loss": 2.5906,
      "theoretical_loss": 3.8114386917230756,
      "tokens_seen": 644642816
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004063891675025075,
      "loss": 2.681,
      "theoretical_loss": 3.8113987703104235,
      "tokens_seen": 644708352
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004063891675025075,
      "loss": 2.8988,
      "theoretical_loss": 3.811358854091799,
      "tokens_seen": 644773888
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004063791374122367,
      "loss": 2.8432,
      "theoretical_loss": 3.8113189430659986,
      "tokens_seen": 644839424
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004063691073219659,
      "loss": 2.8233,
      "theoretical_loss": 3.811279037231819,
      "tokens_seen": 644904960
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040635907723169507,
      "loss": 2.7678,
      "theoretical_loss": 3.8112391365880574,
      "tokens_seen": 644970496
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004063490471414243,
      "loss": 2.9558,
      "theoretical_loss": 3.811199241133511,
      "tokens_seen": 645036032
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040633901705115343,
      "loss": 2.9838,
      "theoretical_loss": 3.811159350866978,
      "tokens_seen": 645101568
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040632898696088267,
      "loss": 2.69,
      "theoretical_loss": 3.8111194657872582,
      "tokens_seen": 645167104
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040631895687061185,
      "loss": 2.763,
      "theoretical_loss": 3.8110795858931485,
      "tokens_seen": 645232640
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040630892678034103,
      "loss": 2.7282,
      "theoretical_loss": 3.811039711183448,
      "tokens_seen": 645298176
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004062988966900702,
      "loss": 2.7764,
      "theoretical_loss": 3.810999841656958,
      "tokens_seen": 645363712
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004062888665997994,
      "loss": 2.7889,
      "theoretical_loss": 3.8109599773124776,
      "tokens_seen": 645429248
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040627883650952857,
      "loss": 2.6931,
      "theoretical_loss": 3.810920118148807,
      "tokens_seen": 645494784
    },
    {
      "epoch": 2.02,
      "objective/train/docs_used": 790384,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9371323585510254,
      "objective/train/theoretical_loss": 3.810900190509401,
      "objective/train/tokens_used": 665987552,
      "theoretical_loss": 3.810900190509401,
      "tokens_seen": 645527552
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004062688064192578,
      "loss": 2.9117,
      "theoretical_loss": 3.810880264164747,
      "tokens_seen": 645560320
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040625877632898693,
      "loss": 2.9301,
      "theoretical_loss": 3.8108404153590993,
      "tokens_seen": 645625856
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040624874623871617,
      "loss": 2.7162,
      "theoretical_loss": 3.810800571730665,
      "tokens_seen": 645691392
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004062387161484453,
      "loss": 2.9238,
      "theoretical_loss": 3.8107607332782463,
      "tokens_seen": 645756928
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040622868605817453,
      "loss": 3.0003,
      "theoretical_loss": 3.8107209000006455,
      "tokens_seen": 645822464
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004062186559679037,
      "loss": 3.068,
      "theoretical_loss": 3.8106810718966657,
      "tokens_seen": 645888000
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004062086258776329,
      "loss": 3.066,
      "theoretical_loss": 3.8106412489651103,
      "tokens_seen": 645953536
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004061985957873621,
      "loss": 2.8349,
      "theoretical_loss": 3.810601431204782,
      "tokens_seen": 646019072
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040618856569709126,
      "loss": 2.8544,
      "theoretical_loss": 3.810561618614485,
      "tokens_seen": 646084608
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040617853560682044,
      "loss": 2.753,
      "theoretical_loss": 3.810521811193024,
      "tokens_seen": 646150144
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040616850551654967,
      "loss": 2.7691,
      "theoretical_loss": 3.8104820089392035,
      "tokens_seen": 646215680
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004061584754262788,
      "loss": 2.8669,
      "theoretical_loss": 3.8104422118518295,
      "tokens_seen": 646281216
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040614844533600804,
      "loss": 2.9435,
      "theoretical_loss": 3.810402419929707,
      "tokens_seen": 646346752
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040613841524573727,
      "loss": 2.8705,
      "theoretical_loss": 3.810362633171642,
      "tokens_seen": 646412288
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004061283851554664,
      "loss": 2.8442,
      "theoretical_loss": 3.81032285157644,
      "tokens_seen": 646477824
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040611835506519563,
      "loss": 2.6825,
      "theoretical_loss": 3.8102830751429093,
      "tokens_seen": 646543360
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040610832497492476,
      "loss": 2.8439,
      "theoretical_loss": 3.810243303869856,
      "tokens_seen": 646608896
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.000406098294884654,
      "loss": 2.5503,
      "theoretical_loss": 3.810203537756088,
      "tokens_seen": 646674432
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004060882647943832,
      "loss": 2.5302,
      "theoretical_loss": 3.8101637768004144,
      "tokens_seen": 646739968
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040607823470411236,
      "loss": 2.8447,
      "theoretical_loss": 3.8101240210016414,
      "tokens_seen": 646805504
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040606820461384154,
      "loss": 2.9075,
      "theoretical_loss": 3.810084270358579,
      "tokens_seen": 646871040
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004060581745235707,
      "loss": 2.8127,
      "theoretical_loss": 3.810044524870036,
      "tokens_seen": 646936576
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004060481444332999,
      "loss": 2.7271,
      "theoretical_loss": 3.810004784534823,
      "tokens_seen": 647002112
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040603811434302914,
      "loss": 2.7557,
      "theoretical_loss": 3.8099650493517476,
      "tokens_seen": 647067648
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040602808425275826,
      "loss": 2.7964,
      "theoretical_loss": 3.809925319319623,
      "tokens_seen": 647133184
    },
    {
      "epoch": 2.02,
      "objective/train/docs_used": 791083,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.1197621822357178,
      "objective/train/theoretical_loss": 3.8099054562347945,
      "objective/train/tokens_used": 667625952,
      "theoretical_loss": 3.8099054562347945,
      "tokens_seen": 647165952
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004060180541624875,
      "loss": 2.8243,
      "theoretical_loss": 3.809885594437258,
      "tokens_seen": 647198720
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004060080240722167,
      "loss": 2.7903,
      "theoretical_loss": 3.8098458747034636,
      "tokens_seen": 647264256
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040599799398194586,
      "loss": 2.6526,
      "theoretical_loss": 3.809806160117053,
      "tokens_seen": 647329792
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040598796389167504,
      "loss": 2.8791,
      "theoretical_loss": 3.8097664506768365,
      "tokens_seen": 647395328
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004059779338014042,
      "loss": 2.6126,
      "theoretical_loss": 3.809726746381627,
      "tokens_seen": 647460864
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004059679037111334,
      "loss": 2.7736,
      "theoretical_loss": 3.8096870472302378,
      "tokens_seen": 647526400
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040595787362086264,
      "loss": 2.7677,
      "theoretical_loss": 3.8096473532214805,
      "tokens_seen": 647591936
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040594784353059177,
      "loss": 2.7009,
      "theoretical_loss": 3.8096076643541696,
      "tokens_seen": 647657472
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.000405937813440321,
      "loss": 2.745,
      "theoretical_loss": 3.8095679806271194,
      "tokens_seen": 647723008
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040592778335005013,
      "loss": 2.7262,
      "theoretical_loss": 3.8095283020391433,
      "tokens_seen": 647788544
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040591775325977936,
      "loss": 2.8404,
      "theoretical_loss": 3.8094886285890563,
      "tokens_seen": 647854080
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040590772316950854,
      "loss": 2.7365,
      "theoretical_loss": 3.8094489602756734,
      "tokens_seen": 647919616
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004058976930792377,
      "loss": 2.9187,
      "theoretical_loss": 3.809409297097811,
      "tokens_seen": 647985152
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004058876629889669,
      "loss": 2.6985,
      "theoretical_loss": 3.809369639054283,
      "tokens_seen": 648050688
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004058776328986961,
      "loss": 2.5365,
      "theoretical_loss": 3.809329986143907,
      "tokens_seen": 648116224
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040586760280842527,
      "loss": 2.8878,
      "theoretical_loss": 3.8092903383654995,
      "tokens_seen": 648181760
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004058575727181545,
      "loss": 2.6767,
      "theoretical_loss": 3.8092506957178776,
      "tokens_seen": 648247296
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040584754262788363,
      "loss": 2.8827,
      "theoretical_loss": 3.809211058199858,
      "tokens_seen": 648312832
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040583751253761287,
      "loss": 2.7196,
      "theoretical_loss": 3.80917142581026,
      "tokens_seen": 648378368
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040582748244734205,
      "loss": 3.0595,
      "theoretical_loss": 3.8091317985478996,
      "tokens_seen": 648443904
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040581745235707123,
      "loss": 2.8787,
      "theoretical_loss": 3.8090921764115975,
      "tokens_seen": 648509440
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004058074222668004,
      "loss": 2.9666,
      "theoretical_loss": 3.809052559400172,
      "tokens_seen": 648574976
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004057973921765296,
      "loss": 2.8784,
      "theoretical_loss": 3.809012947512442,
      "tokens_seen": 648640512
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040578736208625877,
      "loss": 2.7818,
      "theoretical_loss": 3.8089733407472273,
      "tokens_seen": 648706048
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.000405777331995988,
      "loss": 3.0789,
      "theoretical_loss": 3.8089337391033484,
      "tokens_seen": 648771584
    },
    {
      "epoch": 2.02,
      "objective/train/docs_used": 792264,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.787963628768921,
      "objective/train/theoretical_loss": 3.8089139402015415,
      "objective/train/tokens_used": 669264352,
      "theoretical_loss": 3.8089139402015415,
      "tokens_seen": 648804352
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040576730190571713,
      "loss": 2.8683,
      "theoretical_loss": 3.808894142579626,
      "tokens_seen": 648837120
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040575727181544637,
      "loss": 2.8342,
      "theoretical_loss": 3.808854551174881,
      "tokens_seen": 648902656
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004057472417251755,
      "loss": 2.7637,
      "theoretical_loss": 3.8088149648879344,
      "tokens_seen": 648968192
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040573721163490473,
      "loss": 2.8181,
      "theoretical_loss": 3.8087753837176077,
      "tokens_seen": 649033728
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004057271815446339,
      "loss": 2.8324,
      "theoretical_loss": 3.8087358076627242,
      "tokens_seen": 649099264
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004057171514543631,
      "loss": 2.9637,
      "theoretical_loss": 3.8086962367221053,
      "tokens_seen": 649164800
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004057071213640923,
      "loss": 2.9783,
      "theoretical_loss": 3.808656670894574,
      "tokens_seen": 649230336
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040569709127382146,
      "loss": 3.0008,
      "theoretical_loss": 3.8086171101789543,
      "tokens_seen": 649295872
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040568706118355064,
      "loss": 2.7163,
      "theoretical_loss": 3.808577554574069,
      "tokens_seen": 649361408
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004056770310932799,
      "loss": 2.6197,
      "theoretical_loss": 3.808538004078743,
      "tokens_seen": 649426944
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.000405667001003009,
      "loss": 2.8005,
      "theoretical_loss": 3.8084984586918,
      "tokens_seen": 649492480
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040565697091273824,
      "loss": 2.9181,
      "theoretical_loss": 3.808458918412065,
      "tokens_seen": 649558016
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004056469408224674,
      "loss": 2.7293,
      "theoretical_loss": 3.8084193832383635,
      "tokens_seen": 649623552
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004056369107321966,
      "loss": 2.7755,
      "theoretical_loss": 3.8083798531695208,
      "tokens_seen": 649689088
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004056268806419258,
      "loss": 2.9192,
      "theoretical_loss": 3.8083403282043635,
      "tokens_seen": 649754624
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040561685055165496,
      "loss": 2.5435,
      "theoretical_loss": 3.808300808341717,
      "tokens_seen": 649820160
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040560682046138414,
      "loss": 2.7715,
      "theoretical_loss": 3.8082612935804097,
      "tokens_seen": 649885696
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004055967903711134,
      "loss": 2.7429,
      "theoretical_loss": 3.808221783919267,
      "tokens_seen": 649951232
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004055867602808425,
      "loss": 2.8494,
      "theoretical_loss": 3.8081822793571174,
      "tokens_seen": 650016768
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040557673019057174,
      "loss": 2.5832,
      "theoretical_loss": 3.8081427798927887,
      "tokens_seen": 650082304
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040556670010030087,
      "loss": 2.797,
      "theoretical_loss": 3.808103285525109,
      "tokens_seen": 650147840
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004055566700100301,
      "loss": 2.9601,
      "theoretical_loss": 3.8080637962529074,
      "tokens_seen": 650213376
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004055466399197593,
      "loss": 3.0431,
      "theoretical_loss": 3.808024312075013,
      "tokens_seen": 650278912
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040553660982948846,
      "loss": 2.8572,
      "theoretical_loss": 3.807984832990255,
      "tokens_seen": 650344448
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040552657973921764,
      "loss": 2.8196,
      "theoretical_loss": 3.807945358997463,
      "tokens_seen": 650409984
    },
    {
      "epoch": 2.02,
      "objective/train/docs_used": 793032,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8657760620117188,
      "objective/train/theoretical_loss": 3.8079256239101893,
      "objective/train/tokens_used": 670902752,
      "theoretical_loss": 3.8079256239101893,
      "tokens_seen": 650442752
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004055165496489469,
      "loss": 2.8388,
      "theoretical_loss": 3.8079058900954683,
      "tokens_seen": 650475520
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.000405506519558676,
      "loss": 2.9388,
      "theoretical_loss": 3.807866426283101,
      "tokens_seen": 650541056
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040549648946840524,
      "loss": 2.7149,
      "theoretical_loss": 3.8078269675591914,
      "tokens_seen": 650606592
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040548645937813437,
      "loss": 2.6655,
      "theoretical_loss": 3.8077875139225723,
      "tokens_seen": 650672128
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004054764292878636,
      "loss": 2.708,
      "theoretical_loss": 3.807748065372075,
      "tokens_seen": 650737664
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004054663991975928,
      "loss": 3.0241,
      "theoretical_loss": 3.807708621906531,
      "tokens_seen": 650803200
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040545636910732197,
      "loss": 2.9069,
      "theoretical_loss": 3.807669183524774,
      "tokens_seen": 650868736
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040544633901705115,
      "loss": 2.7804,
      "theoretical_loss": 3.8076297502256358,
      "tokens_seen": 650934272
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040543630892678033,
      "loss": 3.0921,
      "theoretical_loss": 3.8075903220079503,
      "tokens_seen": 650999808
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004054262788365095,
      "loss": 2.917,
      "theoretical_loss": 3.807550898870552,
      "tokens_seen": 651065344
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040541624874623874,
      "loss": 2.6225,
      "theoretical_loss": 3.8075114808122734,
      "tokens_seen": 651130880
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040540621865596787,
      "loss": 2.7534,
      "theoretical_loss": 3.807472067831951,
      "tokens_seen": 651196416
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004053961885656971,
      "loss": 2.6766,
      "theoretical_loss": 3.807432659928418,
      "tokens_seen": 651261952
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004053861584754263,
      "loss": 2.7737,
      "theoretical_loss": 3.8073932571005105,
      "tokens_seen": 651327488
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040537612838515547,
      "loss": 2.8894,
      "theoretical_loss": 3.807353859347064,
      "tokens_seen": 651393024
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004053660982948847,
      "loss": 2.7208,
      "theoretical_loss": 3.8073144666669148,
      "tokens_seen": 651458560
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040535606820461383,
      "loss": 2.686,
      "theoretical_loss": 3.8072750790588987,
      "tokens_seen": 651524096
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040534603811434307,
      "loss": 2.7685,
      "theoretical_loss": 3.807235696521853,
      "tokens_seen": 651589632
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040533600802407225,
      "loss": 2.694,
      "theoretical_loss": 3.807196319054615,
      "tokens_seen": 651655168
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040532597793380143,
      "loss": 2.8357,
      "theoretical_loss": 3.807156946656022,
      "tokens_seen": 651720704
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004053159478435306,
      "loss": 2.7878,
      "theoretical_loss": 3.8071175793249123,
      "tokens_seen": 651786240
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004053059177532598,
      "loss": 2.9112,
      "theoretical_loss": 3.8070782170601234,
      "tokens_seen": 651851776
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040529588766298897,
      "loss": 2.8191,
      "theoretical_loss": 3.807038859860495,
      "tokens_seen": 651917312
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004052858575727182,
      "loss": 2.8678,
      "theoretical_loss": 3.806999507724866,
      "tokens_seen": 651982848
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040527582748244733,
      "loss": 2.8658,
      "theoretical_loss": 3.8069601606520758,
      "tokens_seen": 652048384
    },
    {
      "epoch": 2.02,
      "objective/train/docs_used": 794457,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.615185499191284,
      "objective/train/theoretical_loss": 3.8069404890138827,
      "objective/train/tokens_used": 672541152,
      "theoretical_loss": 3.8069404890138827,
      "tokens_seen": 652081152
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040526579739217657,
      "loss": 2.8081,
      "theoretical_loss": 3.8069208186409647,
      "tokens_seen": 652113920
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004052557673019057,
      "loss": 3.0599,
      "theoretical_loss": 3.8068814816903718,
      "tokens_seen": 652179456
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040524573721163493,
      "loss": 2.475,
      "theoretical_loss": 3.806842149799139,
      "tokens_seen": 652244992
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004052357071213641,
      "loss": 2.8371,
      "theoretical_loss": 3.806802822966106,
      "tokens_seen": 652310528
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004052256770310933,
      "loss": 2.9591,
      "theoretical_loss": 3.806763501190116,
      "tokens_seen": 652376064
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004052156469408225,
      "loss": 3.0144,
      "theoretical_loss": 3.8067241844700095,
      "tokens_seen": 652441600
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040520561685055166,
      "loss": 2.9075,
      "theoretical_loss": 3.8066848728046287,
      "tokens_seen": 652507136
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040519558676028084,
      "loss": 2.8039,
      "theoretical_loss": 3.8066455661928167,
      "tokens_seen": 652572672
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004051855566700101,
      "loss": 2.9402,
      "theoretical_loss": 3.806606264633416,
      "tokens_seen": 652638208
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004051755265797392,
      "loss": 2.6916,
      "theoretical_loss": 3.8065669681252707,
      "tokens_seen": 652703744
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040516549648946844,
      "loss": 2.9478,
      "theoretical_loss": 3.8065276766672236,
      "tokens_seen": 652769280
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004051554663991976,
      "loss": 2.8893,
      "theoretical_loss": 3.806488390258119,
      "tokens_seen": 652834816
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004051454363089268,
      "loss": 2.8495,
      "theoretical_loss": 3.8064491088968015,
      "tokens_seen": 652900352
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.000405135406218656,
      "loss": 2.9459,
      "theoretical_loss": 3.8064098325821156,
      "tokens_seen": 652965888
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040512537612838516,
      "loss": 2.7826,
      "theoretical_loss": 3.8063705613129075,
      "tokens_seen": 653031424
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040511534603811434,
      "loss": 2.928,
      "theoretical_loss": 3.8063312950880217,
      "tokens_seen": 653096960
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004051053159478436,
      "loss": 2.9083,
      "theoretical_loss": 3.806292033906305,
      "tokens_seen": 653162496
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004050952858575727,
      "loss": 2.9771,
      "theoretical_loss": 3.806252777766603,
      "tokens_seen": 653228032
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040508525576730194,
      "loss": 2.8971,
      "theoretical_loss": 3.806213526667763,
      "tokens_seen": 653293568
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040507522567703107,
      "loss": 2.8588,
      "theoretical_loss": 3.806174280608632,
      "tokens_seen": 653359104
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004050651955867603,
      "loss": 2.9839,
      "theoretical_loss": 3.8061350395880567,
      "tokens_seen": 653424640
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004050551654964895,
      "loss": 2.6543,
      "theoretical_loss": 3.8060958036048866,
      "tokens_seen": 653490176
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040504513540621866,
      "loss": 2.8108,
      "theoretical_loss": 3.806056572657969,
      "tokens_seen": 653555712
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040503510531594784,
      "loss": 2.7504,
      "theoretical_loss": 3.8060173467461524,
      "tokens_seen": 653621248
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004050250752256771,
      "loss": 2.9447,
      "theoretical_loss": 3.805978125868286,
      "tokens_seen": 653686784
    },
    {
      "epoch": 2.02,
      "objective/train/docs_used": 795253,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.0313971042633057,
      "objective/train/theoretical_loss": 3.805958517316725,
      "objective/train/tokens_used": 674179552,
      "theoretical_loss": 3.805958517316725,
      "tokens_seen": 653719552
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004050150451354062,
      "loss": 2.9876,
      "theoretical_loss": 3.8059389100232197,
      "tokens_seen": 653752320
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040500501504513544,
      "loss": 2.7331,
      "theoretical_loss": 3.805899699209802,
      "tokens_seen": 653817856
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040499498495486457,
      "loss": 2.6985,
      "theoretical_loss": 3.8058604934268843,
      "tokens_seen": 653883392
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004049849548645938,
      "loss": 2.6745,
      "theoretical_loss": 3.8058212926733175,
      "tokens_seen": 653948928
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.000404974924774323,
      "loss": 2.5625,
      "theoretical_loss": 3.8057820969479508,
      "tokens_seen": 654014464
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040496489468405217,
      "loss": 2.6322,
      "theoretical_loss": 3.805742906249636,
      "tokens_seen": 654080000
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040495486459378135,
      "loss": 2.7656,
      "theoretical_loss": 3.8057037205772257,
      "tokens_seen": 654145536
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040494483450351053,
      "loss": 2.8471,
      "theoretical_loss": 3.8056645399295714,
      "tokens_seen": 654211072
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004049348044132397,
      "loss": 2.807,
      "theoretical_loss": 3.805625364305526,
      "tokens_seen": 654276608
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040492477432296894,
      "loss": 2.897,
      "theoretical_loss": 3.805586193703941,
      "tokens_seen": 654342144
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040491474423269807,
      "loss": 2.5623,
      "theoretical_loss": 3.8055470281236707,
      "tokens_seen": 654407680
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004049047141424273,
      "loss": 2.6036,
      "theoretical_loss": 3.805507867563568,
      "tokens_seen": 654473216
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040489468405215643,
      "loss": 2.7559,
      "theoretical_loss": 3.805468712022488,
      "tokens_seen": 654538752
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040488465396188567,
      "loss": 2.7901,
      "theoretical_loss": 3.8054295614992832,
      "tokens_seen": 654604288
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040487462387161485,
      "loss": 2.9805,
      "theoretical_loss": 3.80539041599281,
      "tokens_seen": 654669824
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040486459378134403,
      "loss": 2.7287,
      "theoretical_loss": 3.805351275501922,
      "tokens_seen": 654735360
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004048545636910732,
      "loss": 2.7279,
      "theoretical_loss": 3.8053121400254755,
      "tokens_seen": 654800896
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040484453360080245,
      "loss": 3.0278,
      "theoretical_loss": 3.8052730095623266,
      "tokens_seen": 654866432
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004048345035105316,
      "loss": 2.7637,
      "theoretical_loss": 3.8052338841113302,
      "tokens_seen": 654931968
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004048244734202608,
      "loss": 2.7719,
      "theoretical_loss": 3.805194763671344,
      "tokens_seen": 654997504
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040481444332998994,
      "loss": 2.857,
      "theoretical_loss": 3.8051556482412248,
      "tokens_seen": 655063040
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040480441323971917,
      "loss": 2.822,
      "theoretical_loss": 3.8051165378198295,
      "tokens_seen": 655128576
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040479438314944835,
      "loss": 2.7325,
      "theoretical_loss": 3.8050774324060157,
      "tokens_seen": 655194112
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040478435305917753,
      "loss": 3.0475,
      "theoretical_loss": 3.8050383319986425,
      "tokens_seen": 655259648
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004047743229689067,
      "loss": 2.7759,
      "theoretical_loss": 3.804999236596567,
      "tokens_seen": 655325184
    },
    {
      "debugging/Self-BLEU-5": 0.4859850452541583,
      "debugging/distinct-1-grams": 0.7709278301586879,
      "debugging/distinct-2-grams": 0.9620726312090557,
      "debugging/entropy-1-grams": 5.7746798218568784,
      "debugging/entropy-2-grams": 6.637753508493868,
      "debugging/length": 600.7777777777778,
      "debugging/num_segments": 9,
      "debugging/score": 0.006362688036581238,
      "debugging/score_std": 0.004646295055099869,
      "epoch": 2.02,
      "objective/train/docs_used": 796377,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9904468059539795,
      "objective/train/theoretical_loss": 3.8049796907721594,
      "objective/train/tokens_used": 675817952,
      "theoretical_loss": 3.8049796907721594,
      "tokens_seen": 655357952
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004047642928786359,
      "loss": 2.8813,
      "theoretical_loss": 3.8049601461986486,
      "tokens_seen": 655390720
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004047542627883651,
      "loss": 2.9686,
      "theoretical_loss": 3.8049210608037463,
      "tokens_seen": 655456256
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004047442326980943,
      "loss": 2.8707,
      "theoretical_loss": 3.8048819804107197,
      "tokens_seen": 655521792
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040473420260782344,
      "loss": 2.6262,
      "theoretical_loss": 3.804842905018429,
      "tokens_seen": 655587328
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004047241725175527,
      "loss": 2.9797,
      "theoretical_loss": 3.8048038346257345,
      "tokens_seen": 655652864
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004047141424272818,
      "loss": 2.845,
      "theoretical_loss": 3.8047647692314968,
      "tokens_seen": 655718400
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040470411233701104,
      "loss": 2.7958,
      "theoretical_loss": 3.8047257088345763,
      "tokens_seen": 655783936
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004046940822467402,
      "loss": 2.9501,
      "theoretical_loss": 3.8046866534338353,
      "tokens_seen": 655849472
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004046840521564694,
      "loss": 2.7035,
      "theoretical_loss": 3.804647603028135,
      "tokens_seen": 655915008
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004046740220661986,
      "loss": 2.6593,
      "theoretical_loss": 3.804608557616338,
      "tokens_seen": 655980544
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004046639919759278,
      "loss": 2.9606,
      "theoretical_loss": 3.8045695171973066,
      "tokens_seen": 656046080
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.000404653961885657,
      "loss": 2.9506,
      "theoretical_loss": 3.8045304817699037,
      "tokens_seen": 656111616
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004046439317953862,
      "loss": 2.6727,
      "theoretical_loss": 3.804491451332993,
      "tokens_seen": 656177152
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040463390170511536,
      "loss": 2.5689,
      "theoretical_loss": 3.8044524258854375,
      "tokens_seen": 656242688
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040462387161484454,
      "loss": 2.7187,
      "theoretical_loss": 3.8044134054261014,
      "tokens_seen": 656308224
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004046138415245738,
      "loss": 2.7893,
      "theoretical_loss": 3.80437438995385,
      "tokens_seen": 656373760
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004046038114343029,
      "loss": 2.8434,
      "theoretical_loss": 3.8043353794675463,
      "tokens_seen": 656439296
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040459378134403214,
      "loss": 2.9646,
      "theoretical_loss": 3.804296373966057,
      "tokens_seen": 656504832
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040458375125376127,
      "loss": 2.9982,
      "theoretical_loss": 3.8042573734482463,
      "tokens_seen": 656570368
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004045737211634905,
      "loss": 2.9142,
      "theoretical_loss": 3.8042183779129823,
      "tokens_seen": 656635904
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004045636910732197,
      "loss": 2.8449,
      "theoretical_loss": 3.8041793873591283,
      "tokens_seen": 656701440
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040455366098294886,
      "loss": 2.8761,
      "theoretical_loss": 3.8041404017855536,
      "tokens_seen": 656766976
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040454363089267804,
      "loss": 2.7735,
      "theoretical_loss": 3.804101421191124,
      "tokens_seen": 656832512
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004045336008024073,
      "loss": 2.761,
      "theoretical_loss": 3.804062445574706,
      "tokens_seen": 656898048
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004045235707121364,
      "loss": 2.6464,
      "theoretical_loss": 3.8040234749351693,
      "tokens_seen": 656963584
    },
    {
      "epoch": 2.02,
      "objective/train/docs_used": 797121,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.958272695541382,
      "objective/train/theoretical_loss": 3.804003991481377,
      "objective/train/tokens_used": 677456352,
      "theoretical_loss": 3.804003991481377,
      "tokens_seen": 656996352
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040451354062186564,
      "loss": 2.8226,
      "theoretical_loss": 3.8039845092713804,
      "tokens_seen": 657029120
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040450351053159477,
      "loss": 2.9393,
      "theoretical_loss": 3.8039455485822087,
      "tokens_seen": 657094656
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.000404493480441324,
      "loss": 2.8633,
      "theoretical_loss": 3.8039065928665226,
      "tokens_seen": 657160192
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004044834503510532,
      "loss": 2.9581,
      "theoretical_loss": 3.8038676421231914,
      "tokens_seen": 657225728
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040447342026078237,
      "loss": 2.7917,
      "theoretical_loss": 3.8038286963510846,
      "tokens_seen": 657291264
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040446339017051155,
      "loss": 2.6505,
      "theoretical_loss": 3.8037897555490723,
      "tokens_seen": 657356800
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040445336008024073,
      "loss": 3.0437,
      "theoretical_loss": 3.803750819716025,
      "tokens_seen": 657422336
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004044433299899699,
      "loss": 2.9178,
      "theoretical_loss": 3.8037118888508132,
      "tokens_seen": 657487872
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040443329989969915,
      "loss": 2.7833,
      "theoretical_loss": 3.803672962952308,
      "tokens_seen": 657553408
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040442326980942827,
      "loss": 2.8205,
      "theoretical_loss": 3.803634042019381,
      "tokens_seen": 657618944
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004044132397191575,
      "loss": 2.6532,
      "theoretical_loss": 3.8035951260509036,
      "tokens_seen": 657684480
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040440320962888663,
      "loss": 2.9728,
      "theoretical_loss": 3.8035562150457487,
      "tokens_seen": 657750016
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040439317953861587,
      "loss": 2.8446,
      "theoretical_loss": 3.803517309002788,
      "tokens_seen": 657815552
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040438314944834505,
      "loss": 2.6177,
      "theoretical_loss": 3.8034784079208945,
      "tokens_seen": 657881088
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040437311935807423,
      "loss": 2.8787,
      "theoretical_loss": 3.803439511798943,
      "tokens_seen": 657946624
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004043630892678034,
      "loss": 3.0163,
      "theoretical_loss": 3.8034006206358044,
      "tokens_seen": 658012160
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040435305917753265,
      "loss": 2.7094,
      "theoretical_loss": 3.803361734430355,
      "tokens_seen": 658077696
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004043430290872618,
      "loss": 2.9232,
      "theoretical_loss": 3.8033228531814682,
      "tokens_seen": 658143232
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.000404332998996991,
      "loss": 2.9224,
      "theoretical_loss": 3.803283976888019,
      "tokens_seen": 658208768
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040432296890672014,
      "loss": 2.6132,
      "theoretical_loss": 3.803245105548883,
      "tokens_seen": 658274304
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040431293881644937,
      "loss": 2.7078,
      "theoretical_loss": 3.803206239162935,
      "tokens_seen": 658339840
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040430290872617855,
      "loss": 2.6425,
      "theoretical_loss": 3.803167377729051,
      "tokens_seen": 658405376
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040429287863590773,
      "loss": 2.7553,
      "theoretical_loss": 3.803128521246107,
      "tokens_seen": 658470912
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004042828485456369,
      "loss": 2.9433,
      "theoretical_loss": 3.80308966971298,
      "tokens_seen": 658536448
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004042728184553661,
      "loss": 2.9494,
      "theoretical_loss": 3.8030508231285465,
      "tokens_seen": 658601984
    },
    {
      "epoch": 2.02,
      "objective/train/docs_used": 798701,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.928380250930786,
      "objective/train/theoretical_loss": 3.8030314016917397,
      "objective/train/tokens_used": 679094752,
      "theoretical_loss": 3.8030314016917397,
      "tokens_seen": 658634752
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004042627883650953,
      "loss": 3.0227,
      "theoretical_loss": 3.803011981491685,
      "tokens_seen": 658667520
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004042527582748245,
      "loss": 2.8351,
      "theoretical_loss": 3.802973144801272,
      "tokens_seen": 658733056
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040424272818455364,
      "loss": 2.837,
      "theoretical_loss": 3.802934313056186,
      "tokens_seen": 658798592
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004042326980942829,
      "loss": 2.9081,
      "theoretical_loss": 3.802895486255305,
      "tokens_seen": 658864128
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.000404222668004012,
      "loss": 2.8801,
      "theoretical_loss": 3.802856664397508,
      "tokens_seen": 658929664
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040421263791374124,
      "loss": 2.8748,
      "theoretical_loss": 3.8028178474816743,
      "tokens_seen": 658995200
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004042026078234704,
      "loss": 2.9186,
      "theoretical_loss": 3.802779035506684,
      "tokens_seen": 659060736
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004041925777331996,
      "loss": 2.8017,
      "theoretical_loss": 3.802740228471416,
      "tokens_seen": 659126272
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004041825476429288,
      "loss": 2.7335,
      "theoretical_loss": 3.8027014263747505,
      "tokens_seen": 659191808
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.000404172517552658,
      "loss": 2.7138,
      "theoretical_loss": 3.802662629215569,
      "tokens_seen": 659257344
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040416248746238714,
      "loss": 2.843,
      "theoretical_loss": 3.802623836992752,
      "tokens_seen": 659322880
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004041524573721164,
      "loss": 2.9381,
      "theoretical_loss": 3.8025850497051805,
      "tokens_seen": 659388416
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004041424272818455,
      "loss": 3.0615,
      "theoretical_loss": 3.802546267351737,
      "tokens_seen": 659453952
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040413239719157474,
      "loss": 2.6555,
      "theoretical_loss": 3.802507489931303,
      "tokens_seen": 659519488
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004041223671013039,
      "loss": 2.6922,
      "theoretical_loss": 3.802468717442761,
      "tokens_seen": 659585024
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004041123370110331,
      "loss": 2.8747,
      "theoretical_loss": 3.802429949884994,
      "tokens_seen": 659650560
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004041023069207623,
      "loss": 2.6937,
      "theoretical_loss": 3.8023911872568847,
      "tokens_seen": 659716096
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040409227683049147,
      "loss": 2.9162,
      "theoretical_loss": 3.8023524295573177,
      "tokens_seen": 659781632
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040408224674022065,
      "loss": 2.7625,
      "theoretical_loss": 3.8023136767851753,
      "tokens_seen": 659847168
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004040722166499499,
      "loss": 2.3571,
      "theoretical_loss": 3.802274928939343,
      "tokens_seen": 659912704
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.000404062186559679,
      "loss": 2.8614,
      "theoretical_loss": 3.802236186018705,
      "tokens_seen": 659978240
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040405215646940824,
      "loss": 2.8778,
      "theoretical_loss": 3.802197448022147,
      "tokens_seen": 660043776
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040404212637913737,
      "loss": 2.7498,
      "theoretical_loss": 3.802158714948553,
      "tokens_seen": 660109312
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004040320962888666,
      "loss": 2.893,
      "theoretical_loss": 3.8021199867968094,
      "tokens_seen": 660174848
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004040220661985958,
      "loss": 2.9762,
      "theoretical_loss": 3.8020812635658023,
      "tokens_seen": 660240384
    },
    {
      "epoch": 2.02,
      "objective/train/docs_used": 799454,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4462287425994873,
      "objective/train/theoretical_loss": 3.8020619037952272,
      "objective/train/tokens_used": 680733152,
      "theoretical_loss": 3.8020619037952272,
      "tokens_seen": 660273152
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040401203610832497,
      "loss": 2.8877,
      "theoretical_loss": 3.802042545254418,
      "tokens_seen": 660305920
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040400200601805415,
      "loss": 2.8215,
      "theoretical_loss": 3.802003831861544,
      "tokens_seen": 660371456
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004039919759277834,
      "loss": 2.798,
      "theoretical_loss": 3.8019651233860663,
      "tokens_seen": 660436992
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004039819458375125,
      "loss": 2.9349,
      "theoretical_loss": 3.8019264198268736,
      "tokens_seen": 660502528
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040397191574724175,
      "loss": 2.7946,
      "theoretical_loss": 3.801887721182853,
      "tokens_seen": 660568064
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004039618856569709,
      "loss": 2.9562,
      "theoretical_loss": 3.8018490274528918,
      "tokens_seen": 660633600
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004039518555667001,
      "loss": 2.9122,
      "theoretical_loss": 3.801810338635881,
      "tokens_seen": 660699136
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004039418254764293,
      "loss": 2.9185,
      "theoretical_loss": 3.8017716547307074,
      "tokens_seen": 660764672
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040393179538615847,
      "loss": 2.7965,
      "theoretical_loss": 3.801732975736262,
      "tokens_seen": 660830208
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040392176529588765,
      "loss": 3.0356,
      "theoretical_loss": 3.801694301651433,
      "tokens_seen": 660895744
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040391173520561683,
      "loss": 2.8722,
      "theoretical_loss": 3.8016556324751116,
      "tokens_seen": 660961280
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040390170511534607,
      "loss": 2.9781,
      "theoretical_loss": 3.801616968206188,
      "tokens_seen": 661026816
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040389167502507525,
      "loss": 2.853,
      "theoretical_loss": 3.8015783088435526,
      "tokens_seen": 661092352
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040388164493480443,
      "loss": 3.0213,
      "theoretical_loss": 3.8015396543860964,
      "tokens_seen": 661157888
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004038716148445336,
      "loss": 2.7481,
      "theoretical_loss": 3.8015010048327116,
      "tokens_seen": 661223424
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040386158475426285,
      "loss": 2.7774,
      "theoretical_loss": 3.8014623601822892,
      "tokens_seen": 661288960
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.000403851554663992,
      "loss": 2.7178,
      "theoretical_loss": 3.8014237204337222,
      "tokens_seen": 661354496
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004038415245737212,
      "loss": 2.7672,
      "theoretical_loss": 3.8013850855859026,
      "tokens_seen": 661420032
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040383149448345034,
      "loss": 2.8802,
      "theoretical_loss": 3.801346455637724,
      "tokens_seen": 661485568
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040382146439317957,
      "loss": 2.8351,
      "theoretical_loss": 3.801307830588079,
      "tokens_seen": 661551104
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040381143430290875,
      "loss": 2.8868,
      "theoretical_loss": 3.801269210435862,
      "tokens_seen": 661616640
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040380140421263794,
      "loss": 2.7866,
      "theoretical_loss": 3.801230595179966,
      "tokens_seen": 661682176
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004037913741223671,
      "loss": 2.7692,
      "theoretical_loss": 3.801191984819286,
      "tokens_seen": 661747712
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004037813440320963,
      "loss": 2.7309,
      "theoretical_loss": 3.8011533793527166,
      "tokens_seen": 661813248
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004037713139418255,
      "loss": 2.7817,
      "theoretical_loss": 3.8011147787791533,
      "tokens_seen": 661878784
    },
    {
      "epoch": 2.02,
      "objective/train/docs_used": 800678,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.2656052112579346,
      "objective/train/theoretical_loss": 3.8010954803269037,
      "objective/train/tokens_used": 682371552,
      "theoretical_loss": 3.8010954803269037,
      "tokens_seen": 661911552
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004037612838515547,
      "loss": 2.8415,
      "theoretical_loss": 3.801076183097491,
      "tokens_seen": 661944320
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040375125376128384,
      "loss": 2.8146,
      "theoretical_loss": 3.801037592306626,
      "tokens_seen": 662009856
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004037412236710131,
      "loss": 2.8478,
      "theoretical_loss": 3.800999006405454,
      "tokens_seen": 662075392
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004037311935807422,
      "loss": 2.8708,
      "theoretical_loss": 3.8009604253928715,
      "tokens_seen": 662140928
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040372116349047144,
      "loss": 2.9692,
      "theoretical_loss": 3.800921849267776,
      "tokens_seen": 662206464
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004037111334002006,
      "loss": 3.0543,
      "theoretical_loss": 3.8008832780290636,
      "tokens_seen": 662272000
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004037011033099298,
      "loss": 2.8201,
      "theoretical_loss": 3.8008447116756336,
      "tokens_seen": 662337536
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.000403691073219659,
      "loss": 2.8622,
      "theoretical_loss": 3.800806150206382,
      "tokens_seen": 662403072
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004036810431293882,
      "loss": 2.8962,
      "theoretical_loss": 3.800767593620209,
      "tokens_seen": 662468608
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040367101303911734,
      "loss": 3.0683,
      "theoretical_loss": 3.800729041916012,
      "tokens_seen": 662534144
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004036609829488466,
      "loss": 3.0585,
      "theoretical_loss": 3.8006904950926907,
      "tokens_seen": 662599680
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004036509528585757,
      "loss": 3.0067,
      "theoretical_loss": 3.800651953149144,
      "tokens_seen": 662665216
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040364092276830494,
      "loss": 3.0874,
      "theoretical_loss": 3.800613416084272,
      "tokens_seen": 662730752
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004036308926780341,
      "loss": 2.905,
      "theoretical_loss": 3.8005748838969744,
      "tokens_seen": 662796288
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004036208625877633,
      "loss": 2.9418,
      "theoretical_loss": 3.8005363565861527,
      "tokens_seen": 662861824
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004036108324974925,
      "loss": 2.7549,
      "theoretical_loss": 3.800497834150706,
      "tokens_seen": 662927360
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040360080240722167,
      "loss": 2.8905,
      "theoretical_loss": 3.800459316589537,
      "tokens_seen": 662992896
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040359077231695085,
      "loss": 2.752,
      "theoretical_loss": 3.800420803901546,
      "tokens_seen": 663058432
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004035807422266801,
      "loss": 2.7383,
      "theoretical_loss": 3.8003822960856364,
      "tokens_seen": 663123968
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004035707121364092,
      "loss": 2.9399,
      "theoretical_loss": 3.8003437931407094,
      "tokens_seen": 663189504
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040356068204613844,
      "loss": 2.9639,
      "theoretical_loss": 3.8003052950656673,
      "tokens_seen": 663255040
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040355065195586757,
      "loss": 2.9999,
      "theoretical_loss": 3.800266801859414,
      "tokens_seen": 663320576
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004035406218655968,
      "loss": 2.8012,
      "theoretical_loss": 3.8002283135208517,
      "tokens_seen": 663386112
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.000403530591775326,
      "loss": 2.7796,
      "theoretical_loss": 3.800189830048885,
      "tokens_seen": 663451648
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040352056168505517,
      "loss": 2.8371,
      "theoretical_loss": 3.800151351442418,
      "tokens_seen": 663517184
    },
    {
      "epoch": 2.02,
      "objective/train/docs_used": 801382,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9103448390960693,
      "objective/train/theoretical_loss": 3.800132113963404,
      "objective/train/tokens_used": 684009952,
      "theoretical_loss": 3.800132113963404,
      "tokens_seen": 663549952
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040351053159478435,
      "loss": 2.8541,
      "theoretical_loss": 3.8001128777003546,
      "tokens_seen": 663582720
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004035005015045136,
      "loss": 2.8874,
      "theoretical_loss": 3.800074408821599,
      "tokens_seen": 663648256
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004034904714142427,
      "loss": 2.8394,
      "theoretical_loss": 3.8000359448050576,
      "tokens_seen": 663713792
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040348044132397195,
      "loss": 2.8841,
      "theoretical_loss": 3.7999974856496346,
      "tokens_seen": 663779328
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004034704112337011,
      "loss": 2.6108,
      "theoretical_loss": 3.799959031354237,
      "tokens_seen": 663844864
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004034603811434303,
      "loss": 2.8604,
      "theoretical_loss": 3.7999205819177693,
      "tokens_seen": 663910400
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004034503510531595,
      "loss": 2.6844,
      "theoretical_loss": 3.79988213733914,
      "tokens_seen": 663975936
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040344032096288867,
      "loss": 2.7968,
      "theoretical_loss": 3.7998436976172543,
      "tokens_seen": 664041472
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040343029087261785,
      "loss": 2.923,
      "theoretical_loss": 3.79980526275102,
      "tokens_seen": 664107008
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040342026078234703,
      "loss": 2.7212,
      "theoretical_loss": 3.799766832739345,
      "tokens_seen": 664172544
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004034102306920762,
      "loss": 2.8013,
      "theoretical_loss": 3.799728407581137,
      "tokens_seen": 664238080
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040340020060180545,
      "loss": 2.8651,
      "theoretical_loss": 3.799689987275304,
      "tokens_seen": 664303616
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004033901705115346,
      "loss": 3.1115,
      "theoretical_loss": 3.799651571820755,
      "tokens_seen": 664369152
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004033801404212638,
      "loss": 2.7985,
      "theoretical_loss": 3.7996131612163984,
      "tokens_seen": 664434688
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.000403370110330993,
      "loss": 2.779,
      "theoretical_loss": 3.799574755461144,
      "tokens_seen": 664500224
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004033600802407222,
      "loss": 3.0251,
      "theoretical_loss": 3.7995363545539016,
      "tokens_seen": 664565760
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040335005015045136,
      "loss": 2.9,
      "theoretical_loss": 3.799497958493581,
      "tokens_seen": 664631296
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040334002006018054,
      "loss": 2.8359,
      "theoretical_loss": 3.7994595672790927,
      "tokens_seen": 664696832
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004033299899699097,
      "loss": 2.8659,
      "theoretical_loss": 3.7994211809093468,
      "tokens_seen": 664762368
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040331995987963895,
      "loss": 3.0194,
      "theoretical_loss": 3.799382799383255,
      "tokens_seen": 664827904
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004033099297893681,
      "loss": 2.9732,
      "theoretical_loss": 3.7993444226997295,
      "tokens_seen": 664893440
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004032998996990973,
      "loss": 2.7585,
      "theoretical_loss": 3.7993060508576804,
      "tokens_seen": 664958976
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040328986960882644,
      "loss": 2.8927,
      "theoretical_loss": 3.799267683856021,
      "tokens_seen": 665024512
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004032798395185557,
      "loss": 2.9975,
      "theoretical_loss": 3.799229321693664,
      "tokens_seen": 665090048
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040326980942828486,
      "loss": 2.7975,
      "theoretical_loss": 3.7991909643695214,
      "tokens_seen": 665155584
    },
    {
      "epoch": 2.02,
      "objective/train/docs_used": 802714,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.3324520587921143,
      "objective/train/theoretical_loss": 3.7991717875214412,
      "objective/train/tokens_used": 685648352,
      "theoretical_loss": 3.7991717875214412,
      "tokens_seen": 665188352
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040325977933801404,
      "loss": 2.7329,
      "theoretical_loss": 3.7991526118825067,
      "tokens_seen": 665221120
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004032497492477432,
      "loss": 2.9358,
      "theoretical_loss": 3.799114264231534,
      "tokens_seen": 665286656
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004032397191574724,
      "loss": 2.9407,
      "theoretical_loss": 3.7990759214155156,
      "tokens_seen": 665352192
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004032296890672016,
      "loss": 2.8405,
      "theoretical_loss": 3.799037583433368,
      "tokens_seen": 665417728
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004032196589769308,
      "loss": 3.0254,
      "theoretical_loss": 3.7989992502840044,
      "tokens_seen": 665483264
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040320962888665995,
      "loss": 2.6699,
      "theoretical_loss": 3.79896092196634,
      "tokens_seen": 665548800
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004031995987963892,
      "loss": 2.8213,
      "theoretical_loss": 3.7989225984792903,
      "tokens_seen": 665614336
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040318956870611836,
      "loss": 2.8812,
      "theoretical_loss": 3.7988842798217703,
      "tokens_seen": 665679872
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040317953861584754,
      "loss": 2.8593,
      "theoretical_loss": 3.7988459659926965,
      "tokens_seen": 665745408
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004031695085255767,
      "loss": 2.8102,
      "theoretical_loss": 3.798807656990986,
      "tokens_seen": 665810944
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004031594784353059,
      "loss": 3.0188,
      "theoretical_loss": 3.798769352815554,
      "tokens_seen": 665876480
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040314944834503514,
      "loss": 2.7791,
      "theoretical_loss": 3.7987310534653194,
      "tokens_seen": 665942016
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004031394182547643,
      "loss": 2.8322,
      "theoretical_loss": 3.7986927589391977,
      "tokens_seen": 666007552
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004031293881644935,
      "loss": 2.9257,
      "theoretical_loss": 3.798654469236107,
      "tokens_seen": 666073088
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004031193580742227,
      "loss": 2.8127,
      "theoretical_loss": 3.798616184354967,
      "tokens_seen": 666138624
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040310932798395187,
      "loss": 2.847,
      "theoretical_loss": 3.798577904294694,
      "tokens_seen": 666204160
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040309929789368105,
      "loss": 3.0944,
      "theoretical_loss": 3.7985396290542086,
      "tokens_seen": 666269696
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004030892678034103,
      "loss": 2.6814,
      "theoretical_loss": 3.798501358632429,
      "tokens_seen": 666335232
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004030792377131394,
      "loss": 2.9891,
      "theoretical_loss": 3.798463093028275,
      "tokens_seen": 666400768
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040306920762286864,
      "loss": 2.7525,
      "theoretical_loss": 3.798424832240666,
      "tokens_seen": 666466304
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040305917753259777,
      "loss": 3.0883,
      "theoretical_loss": 3.7983865762685225,
      "tokens_seen": 666531840
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.000403049147442327,
      "loss": 2.946,
      "theoretical_loss": 3.7983483251107657,
      "tokens_seen": 666597376
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004030391173520562,
      "loss": 2.8419,
      "theoretical_loss": 3.798310078766315,
      "tokens_seen": 666662912
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040302908726178537,
      "loss": 2.8564,
      "theoretical_loss": 3.7982718372340933,
      "tokens_seen": 666728448
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040301905717151455,
      "loss": 2.7861,
      "theoretical_loss": 3.7982336005130213,
      "tokens_seen": 666793984
    },
    {
      "epoch": 2.02,
      "objective/train/docs_used": 803703,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.0033750534057617,
      "objective/train/theoretical_loss": 3.798214483956329,
      "objective/train/tokens_used": 687286752,
      "theoretical_loss": 3.798214483956329,
      "tokens_seen": 666826752
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004030090270812438,
      "loss": 2.9464,
      "theoretical_loss": 3.7981953686020207,
      "tokens_seen": 666859520
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004029989969909729,
      "loss": 3.1007,
      "theoretical_loss": 3.798157141500014,
      "tokens_seen": 666925056
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040298896690070215,
      "loss": 2.8264,
      "theoretical_loss": 3.7981189192059244,
      "tokens_seen": 666990592
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004029789368104313,
      "loss": 2.8476,
      "theoretical_loss": 3.7980807017186744,
      "tokens_seen": 667056128
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004029689067201605,
      "loss": 3.1105,
      "theoretical_loss": 3.7980424890371873,
      "tokens_seen": 667121664
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004029588766298897,
      "loss": 2.6239,
      "theoretical_loss": 3.7980042811603867,
      "tokens_seen": 667187200
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040294884653961887,
      "loss": 2.8954,
      "theoretical_loss": 3.797966078087197,
      "tokens_seen": 667252736
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040293881644934805,
      "loss": 3.0282,
      "theoretical_loss": 3.797927879816542,
      "tokens_seen": 667318272
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040292878635907723,
      "loss": 3.0588,
      "theoretical_loss": 3.7978896863473466,
      "tokens_seen": 667383808
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004029187562688064,
      "loss": 2.8539,
      "theoretical_loss": 3.7978514976785362,
      "tokens_seen": 667449344
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040290872617853565,
      "loss": 2.8734,
      "theoretical_loss": 3.797813313809036,
      "tokens_seen": 667514880
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004028986960882648,
      "loss": 2.8289,
      "theoretical_loss": 3.7977751347377717,
      "tokens_seen": 667580416
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.000402888665997994,
      "loss": 2.7012,
      "theoretical_loss": 3.7977369604636695,
      "tokens_seen": 667645952
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004028786359077232,
      "loss": 2.9201,
      "theoretical_loss": 3.7976987909856557,
      "tokens_seen": 667711488
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004028686058174524,
      "loss": 2.8353,
      "theoretical_loss": 3.7976606263026573,
      "tokens_seen": 667777024
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040285857572718156,
      "loss": 3.0119,
      "theoretical_loss": 3.7976224664136007,
      "tokens_seen": 667842560
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040284854563691074,
      "loss": 2.9295,
      "theoretical_loss": 3.7975843113174146,
      "tokens_seen": 667908096
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004028385155466399,
      "loss": 2.653,
      "theoretical_loss": 3.797546161013026,
      "tokens_seen": 667973632
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040282848545636915,
      "loss": 2.9208,
      "theoretical_loss": 3.7975080154993632,
      "tokens_seen": 668039168
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004028184553660983,
      "loss": 2.9167,
      "theoretical_loss": 3.7974698747753552,
      "tokens_seen": 668104704
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004028084252758275,
      "loss": 2.961,
      "theoretical_loss": 3.7974317388399297,
      "tokens_seen": 668170240
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040279839518555664,
      "loss": 2.9478,
      "theoretical_loss": 3.797393607692017,
      "tokens_seen": 668235776
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004027883650952859,
      "loss": 2.8548,
      "theoretical_loss": 3.7973554813305466,
      "tokens_seen": 668301312
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040277833500501506,
      "loss": 2.9796,
      "theoretical_loss": 3.797317359754448,
      "tokens_seen": 668366848
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040276830491474424,
      "loss": 2.7677,
      "theoretical_loss": 3.797279242962651,
      "tokens_seen": 668432384
    },
    {
      "epoch": 2.02,
      "objective/train/docs_used": 804287,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6115474700927734,
      "objective/train/theoretical_loss": 3.7972601863605315,
      "objective/train/tokens_used": 688925152,
      "theoretical_loss": 3.7972601863605315,
      "tokens_seen": 668465152
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004027582748244734,
      "loss": 2.9188,
      "theoretical_loss": 3.797241130954087,
      "tokens_seen": 668497920
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004027482447342026,
      "loss": 2.7482,
      "theoretical_loss": 3.7972030237276866,
      "tokens_seen": 668563456
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004027382146439318,
      "loss": 2.6248,
      "theoretical_loss": 3.7971649212823806,
      "tokens_seen": 668628992
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.000402728184553661,
      "loss": 2.8523,
      "theoretical_loss": 3.7971268236171016,
      "tokens_seen": 668694528
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040271815446339015,
      "loss": 2.7225,
      "theoretical_loss": 3.7970887307307812,
      "tokens_seen": 668760064
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004027081243731194,
      "loss": 2.7711,
      "theoretical_loss": 3.7970506426223514,
      "tokens_seen": 668825600
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040269809428284856,
      "loss": 3.0768,
      "theoretical_loss": 3.797012559290745,
      "tokens_seen": 668891136
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040268806419257774,
      "loss": 3.112,
      "theoretical_loss": 3.796974480734894,
      "tokens_seen": 668956672
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004026780341023069,
      "loss": 2.8936,
      "theoretical_loss": 3.7969364069537344,
      "tokens_seen": 669022208
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004026680040120361,
      "loss": 2.8176,
      "theoretical_loss": 3.796898337946197,
      "tokens_seen": 669087744
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004026579739217653,
      "loss": 2.8419,
      "theoretical_loss": 3.796860273711217,
      "tokens_seen": 669153280
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004026479438314945,
      "loss": 2.9462,
      "theoretical_loss": 3.7968222142477295,
      "tokens_seen": 669218816
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040263791374122365,
      "loss": 2.8592,
      "theoretical_loss": 3.796784159554668,
      "tokens_seen": 669284352
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004026278836509529,
      "loss": 2.7023,
      "theoretical_loss": 3.796746109630967,
      "tokens_seen": 669349888
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.000402617853560682,
      "loss": 2.7926,
      "theoretical_loss": 3.796708064475564,
      "tokens_seen": 669415424
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040260782347041125,
      "loss": 2.7279,
      "theoretical_loss": 3.796670024087393,
      "tokens_seen": 669480960
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040259779338014043,
      "loss": 3.0721,
      "theoretical_loss": 3.796631988465391,
      "tokens_seen": 669546496
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004025877632898696,
      "loss": 2.7945,
      "theoretical_loss": 3.796593957608494,
      "tokens_seen": 669612032
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004025777331995988,
      "loss": 2.9095,
      "theoretical_loss": 3.7965559315156385,
      "tokens_seen": 669677568
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040256770310932797,
      "loss": 2.977,
      "theoretical_loss": 3.796517910185762,
      "tokens_seen": 669743104
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040255767301905715,
      "loss": 2.8426,
      "theoretical_loss": 3.7964798936178017,
      "tokens_seen": 669808640
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004025476429287864,
      "loss": 2.8808,
      "theoretical_loss": 3.7964418818106953,
      "tokens_seen": 669874176
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004025376128385155,
      "loss": 2.8402,
      "theoretical_loss": 3.7964038747633815,
      "tokens_seen": 669939712
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040252758274824475,
      "loss": 2.9317,
      "theoretical_loss": 3.796365872474798,
      "tokens_seen": 670005248
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040251755265797393,
      "loss": 2.8884,
      "theoretical_loss": 3.7963278749438842,
      "tokens_seen": 670070784
    },
    {
      "epoch": 2.02,
      "objective/train/docs_used": 805654,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.389000654220581,
      "objective/train/theoretical_loss": 3.796308877962222,
      "objective/train/tokens_used": 690563552,
      "theoretical_loss": 3.796308877962222,
      "tokens_seen": 670103552
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004025075225677031,
      "loss": 3.0705,
      "theoretical_loss": 3.796289882169579,
      "tokens_seen": 670136320
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004024974924774323,
      "loss": 2.9853,
      "theoretical_loss": 3.7962518941508216,
      "tokens_seen": 670201856
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004024874623871615,
      "loss": 2.7051,
      "theoretical_loss": 3.796213910886552,
      "tokens_seen": 670267392
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040247743229689066,
      "loss": 2.6793,
      "theoretical_loss": 3.79617593237571,
      "tokens_seen": 670332928
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004024674022066199,
      "loss": 2.7152,
      "theoretical_loss": 3.796137958617237,
      "tokens_seen": 670398464
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.000402457372116349,
      "loss": 2.7582,
      "theoretical_loss": 3.796099989610073,
      "tokens_seen": 670464000
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040244734202607825,
      "loss": 3.1524,
      "theoretical_loss": 3.7960620253531596,
      "tokens_seen": 670529536
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004024373119358074,
      "loss": 3.0175,
      "theoretical_loss": 3.7960240658454385,
      "tokens_seen": 670595072
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004024272818455366,
      "loss": 2.637,
      "theoretical_loss": 3.795986111085851,
      "tokens_seen": 670660608
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004024172517552658,
      "loss": 2.8622,
      "theoretical_loss": 3.7959481610733397,
      "tokens_seen": 670726144
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.000402407221664995,
      "loss": 2.5987,
      "theoretical_loss": 3.7959102158068463,
      "tokens_seen": 670791680
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004023971915747242,
      "loss": 3.1051,
      "theoretical_loss": 3.795872275285315,
      "tokens_seen": 670857216
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004023871614844534,
      "loss": 2.5387,
      "theoretical_loss": 3.7958343395076883,
      "tokens_seen": 670922752
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004023771313941826,
      "loss": 2.6301,
      "theoretical_loss": 3.7957964084729094,
      "tokens_seen": 670988288
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040236710130391176,
      "loss": 3.0434,
      "theoretical_loss": 3.795758482179923,
      "tokens_seen": 671053824
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040235707121364094,
      "loss": 2.9399,
      "theoretical_loss": 3.7957205606276725,
      "tokens_seen": 671119360
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004023470411233701,
      "loss": 2.9368,
      "theoretical_loss": 3.795682643815103,
      "tokens_seen": 671184896
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040233701103309935,
      "loss": 3.0092,
      "theoretical_loss": 3.7956447317411595,
      "tokens_seen": 671250432
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004023269809428285,
      "loss": 2.6606,
      "theoretical_loss": 3.7956068244047865,
      "tokens_seen": 671315968
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004023169508525577,
      "loss": 2.905,
      "theoretical_loss": 3.79556892180493,
      "tokens_seen": 671381504
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040230692076228684,
      "loss": 2.768,
      "theoretical_loss": 3.795531023940536,
      "tokens_seen": 671447040
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004022968906720161,
      "loss": 3.1211,
      "theoretical_loss": 3.7954931308105513,
      "tokens_seen": 671512576
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040228686058174526,
      "loss": 3.0504,
      "theoretical_loss": 3.7954552424139214,
      "tokens_seen": 671578112
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040227683049147444,
      "loss": 2.8887,
      "theoretical_loss": 3.795417358749594,
      "tokens_seen": 671643648
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004022668004012036,
      "loss": 2.6697,
      "theoretical_loss": 3.795379479816516,
      "tokens_seen": 671709184
    },
    {
      "epoch": 2.02,
      "objective/train/docs_used": 806421,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.0061473846435547,
      "objective/train/theoretical_loss": 3.7953605421238663,
      "objective/train/tokens_used": 692201952,
      "theoretical_loss": 3.7953605421238663,
      "tokens_seen": 671741952
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004022567703109328,
      "loss": 2.9267,
      "theoretical_loss": 3.7953416056136344,
      "tokens_seen": 671774720
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.000402246740220662,
      "loss": 2.9145,
      "theoretical_loss": 3.795303736139899,
      "tokens_seen": 671840256
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004022367101303912,
      "loss": 2.9675,
      "theoretical_loss": 3.795265871394256,
      "tokens_seen": 671905792
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040222668004012035,
      "loss": 3.1682,
      "theoretical_loss": 3.795228011375655,
      "tokens_seen": 671971328
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004022166499498496,
      "loss": 2.7738,
      "theoretical_loss": 3.795190156083045,
      "tokens_seen": 672036864
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040220661985957876,
      "loss": 2.8969,
      "theoretical_loss": 3.7951523055153755,
      "tokens_seen": 672102400
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040219658976930794,
      "loss": 2.9553,
      "theoretical_loss": 3.795114459671595,
      "tokens_seen": 672167936
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004021865596790371,
      "loss": 2.8685,
      "theoretical_loss": 3.795076618550654,
      "tokens_seen": 672233472
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004021765295887663,
      "loss": 2.8885,
      "theoretical_loss": 3.795038782151504,
      "tokens_seen": 672299008
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004021664994984955,
      "loss": 2.9009,
      "theoretical_loss": 3.795000950473094,
      "tokens_seen": 672364544
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004021564694082247,
      "loss": 3.0639,
      "theoretical_loss": 3.7949631235143753,
      "tokens_seen": 672430080
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040214643931795385,
      "loss": 2.7665,
      "theoretical_loss": 3.7949253012743,
      "tokens_seen": 672495616
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004021364092276831,
      "loss": 2.8488,
      "theoretical_loss": 3.7948874837518183,
      "tokens_seen": 672561152
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004021263791374122,
      "loss": 2.8415,
      "theoretical_loss": 3.794849670945884,
      "tokens_seen": 672626688
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040211634904714145,
      "loss": 2.9815,
      "theoretical_loss": 3.7948118628554477,
      "tokens_seen": 672692224
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040210631895687063,
      "loss": 2.9568,
      "theoretical_loss": 3.794774059479463,
      "tokens_seen": 672757760
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004020962888665998,
      "loss": 2.7326,
      "theoretical_loss": 3.7947362608168826,
      "tokens_seen": 672823296
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.000402086258776329,
      "loss": 3.0833,
      "theoretical_loss": 3.79469846686666,
      "tokens_seen": 672888832
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040207622868605817,
      "loss": 3.0286,
      "theoretical_loss": 3.794660677627748,
      "tokens_seen": 672954368
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040206619859578735,
      "loss": 2.8386,
      "theoretical_loss": 3.7946228930991017,
      "tokens_seen": 673019904
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004020561685055166,
      "loss": 2.9388,
      "theoretical_loss": 3.794585113279675,
      "tokens_seen": 673085440
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004020461384152457,
      "loss": 3.028,
      "theoretical_loss": 3.7945473381684227,
      "tokens_seen": 673150976
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040203610832497495,
      "loss": 2.7476,
      "theoretical_loss": 3.7945095677642993,
      "tokens_seen": 673216512
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040202607823470413,
      "loss": 2.9742,
      "theoretical_loss": 3.7944718020662602,
      "tokens_seen": 673282048
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004020160481444333,
      "loss": 3.0702,
      "theoretical_loss": 3.794434041073261,
      "tokens_seen": 673347584
    },
    {
      "epoch": 2.02,
      "objective/train/docs_used": 807207,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9605650901794434,
      "objective/train/theoretical_loss": 3.794415162340826,
      "objective/train/tokens_used": 693840352,
      "theoretical_loss": 3.794415162340826,
      "tokens_seen": 673380352
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004020060180541625,
      "loss": 2.7631,
      "theoretical_loss": 3.7943962847842587,
      "tokens_seen": 673413120
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004019959879638917,
      "loss": 3.1873,
      "theoretical_loss": 3.794358533198208,
      "tokens_seen": 673478656
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040198595787362086,
      "loss": 3.0331,
      "theoretical_loss": 3.7943207863140667,
      "tokens_seen": 673544192
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004019759277833501,
      "loss": 2.8757,
      "theoretical_loss": 3.7942830441307915,
      "tokens_seen": 673609728
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004019658976930792,
      "loss": 2.8293,
      "theoretical_loss": 3.794245306647339,
      "tokens_seen": 673675264
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040195586760280845,
      "loss": 2.7213,
      "theoretical_loss": 3.7942075738626677,
      "tokens_seen": 673740800
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004019458375125376,
      "loss": 2.8015,
      "theoretical_loss": 3.794169845775736,
      "tokens_seen": 673806336
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004019358074222668,
      "loss": 2.7856,
      "theoretical_loss": 3.7941321223855002,
      "tokens_seen": 673871872
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.000401925777331996,
      "loss": 2.8947,
      "theoretical_loss": 3.794094403690921,
      "tokens_seen": 673937408
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004019157472417252,
      "loss": 2.7252,
      "theoretical_loss": 3.794056689690956,
      "tokens_seen": 674002944
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040190571715145436,
      "loss": 2.9241,
      "theoretical_loss": 3.794018980384566,
      "tokens_seen": 674068480
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004018956870611836,
      "loss": 2.9409,
      "theoretical_loss": 3.7939812757707094,
      "tokens_seen": 674134016
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004018856569709127,
      "loss": 2.9654,
      "theoretical_loss": 3.7939435758483464,
      "tokens_seen": 674199552
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040187562688064196,
      "loss": 2.6226,
      "theoretical_loss": 3.793905880616437,
      "tokens_seen": 674265088
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004018655967903711,
      "loss": 2.7045,
      "theoretical_loss": 3.793868190073943,
      "tokens_seen": 674330624
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004018555667001003,
      "loss": 2.8452,
      "theoretical_loss": 3.793830504219824,
      "tokens_seen": 674396160
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004018455366098295,
      "loss": 2.8149,
      "theoretical_loss": 3.793792823053042,
      "tokens_seen": 674461696
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004018355065195587,
      "loss": 2.997,
      "theoretical_loss": 3.7937551465725585,
      "tokens_seen": 674527232
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040182547642928786,
      "loss": 2.7987,
      "theoretical_loss": 3.7937174747773357,
      "tokens_seen": 674592768
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040181544633901704,
      "loss": 2.7863,
      "theoretical_loss": 3.7936798076663347,
      "tokens_seen": 674658304
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004018054162487462,
      "loss": 2.7427,
      "theoretical_loss": 3.79364214523852,
      "tokens_seen": 674723840
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040179538615847546,
      "loss": 2.8898,
      "theoretical_loss": 3.7936044874928534,
      "tokens_seen": 674789376
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004017853560682046,
      "loss": 2.7465,
      "theoretical_loss": 3.7935668344282982,
      "tokens_seen": 674854912
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004017753259779338,
      "loss": 2.7848,
      "theoretical_loss": 3.793529186043818,
      "tokens_seen": 674920448
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040176529588766295,
      "loss": 2.7028,
      "theoretical_loss": 3.7934915423383764,
      "tokens_seen": 674985984
    },
    {
      "epoch": 2.02,
      "objective/train/docs_used": 807950,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.584026336669922,
      "objective/train/theoretical_loss": 3.7934727222399722,
      "objective/train/tokens_used": 695478752,
      "theoretical_loss": 3.7934727222399722,
      "tokens_seen": 675018752
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004017552657973922,
      "loss": 2.759,
      "theoretical_loss": 3.793453903310939,
      "tokens_seen": 675051520
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040174523570712137,
      "loss": 2.9461,
      "theoretical_loss": 3.793416268960469,
      "tokens_seen": 675117056
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040173520561685055,
      "loss": 2.9082,
      "theoretical_loss": 3.7933786392859323,
      "tokens_seen": 675182592
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040172517552657973,
      "loss": 2.8488,
      "theoretical_loss": 3.793341014286293,
      "tokens_seen": 675248128
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040171514543630896,
      "loss": 3.0354,
      "theoretical_loss": 3.7933033939605183,
      "tokens_seen": 675313664
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004017051153460381,
      "loss": 2.7168,
      "theoretical_loss": 3.793265778307572,
      "tokens_seen": 675379200
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004016950852557673,
      "loss": 2.8791,
      "theoretical_loss": 3.7932281673264225,
      "tokens_seen": 675444736
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040168505516549645,
      "loss": 2.5665,
      "theoretical_loss": 3.7931905610160355,
      "tokens_seen": 675510272
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004016750250752257,
      "loss": 3.0928,
      "theoretical_loss": 3.793152959375377,
      "tokens_seen": 675575808
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040166499498495487,
      "loss": 2.8882,
      "theoretical_loss": 3.7931153624034155,
      "tokens_seen": 675641344
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040165496489468405,
      "loss": 2.9232,
      "theoretical_loss": 3.793077770099118,
      "tokens_seen": 675706880
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004016449348044133,
      "loss": 3.0883,
      "theoretical_loss": 3.7930401824614526,
      "tokens_seen": 675772416
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0004016349047141424,
      "loss": 2.8628,
      "theoretical_loss": 3.7930025994893866,
      "tokens_seen": 675837952
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040162487462387165,
      "loss": 3.0159,
      "theoretical_loss": 3.79296502118189,
      "tokens_seen": 675903488
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.00040161484453360083,
      "loss": 2.916,
      "theoretical_loss": 3.792927447537931,
      "tokens_seen": 675969024
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00040160481444333,
      "loss": 2.9463,
      "theoretical_loss": 3.792889878556479,
      "tokens_seen": 676034560
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0004015947843530592,
      "loss": 2.9628,
      "theoretical_loss": 3.7928523142365025,
      "tokens_seen": 676100096
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00040158475426278837,
      "loss": 3.0279,
      "theoretical_loss": 3.7928147545769724,
      "tokens_seen": 676165632
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00040157472417251755,
      "loss": 2.4866,
      "theoretical_loss": 3.792777199576859,
      "tokens_seen": 676231168
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0004015646940822468,
      "loss": 2.6683,
      "theoretical_loss": 3.792739649235132,
      "tokens_seen": 676296704
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0004015546639919759,
      "loss": 2.8666,
      "theoretical_loss": 3.792702103550763,
      "tokens_seen": 676362240
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00040154463390170515,
      "loss": 3.0335,
      "theoretical_loss": 3.792664562522723,
      "tokens_seen": 676427776
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00040153460381143433,
      "loss": 3.1476,
      "theoretical_loss": 3.7926270261499826,
      "tokens_seen": 676493312
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0004015245737211635,
      "loss": 2.7187,
      "theoretical_loss": 3.792589494431515,
      "tokens_seen": 676558848
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0004015145436308927,
      "loss": 2.9815,
      "theoretical_loss": 3.792551967366291,
      "tokens_seen": 676624384
    },
    {
      "epoch": 2.03,
      "objective/train/docs_used": 809057,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.946103096008301,
      "objective/train/theoretical_loss": 3.7925332055783247,
      "objective/train/tokens_used": 697117152,
      "theoretical_loss": 3.7925332055783247,
      "tokens_seen": 676657152
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0004015045135406219,
      "loss": 3.0284,
      "theoretical_loss": 3.7925144449532837,
      "tokens_seen": 676689920
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00040149448345035106,
      "loss": 2.8247,
      "theoretical_loss": 3.7924769271914665,
      "tokens_seen": 676755456
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0004014844533600803,
      "loss": 2.9618,
      "theoretical_loss": 3.7924394140798117,
      "tokens_seen": 676820992
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0004014744232698094,
      "loss": 2.795,
      "theoretical_loss": 3.792401905617293,
      "tokens_seen": 676886528
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00040146439317953865,
      "loss": 3.0326,
      "theoretical_loss": 3.792364401802884,
      "tokens_seen": 676952064
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0004014543630892678,
      "loss": 3.1664,
      "theoretical_loss": 3.7923269026355593,
      "tokens_seen": 677017600
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.000401444332998997,
      "loss": 2.8116,
      "theoretical_loss": 3.7922894081142924,
      "tokens_seen": 677083136
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0004014343029087262,
      "loss": 2.6791,
      "theoretical_loss": 3.792251918238059,
      "tokens_seen": 677148672
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0004014242728184554,
      "loss": 2.8246,
      "theoretical_loss": 3.7922144330058334,
      "tokens_seen": 677214208
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00040141424272818456,
      "loss": 2.7107,
      "theoretical_loss": 3.7921769524165923,
      "tokens_seen": 677279744
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0004014042126379138,
      "loss": 2.9204,
      "theoretical_loss": 3.7921394764693095,
      "tokens_seen": 677345280
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0004013941825476429,
      "loss": 2.6977,
      "theoretical_loss": 3.7921020051629624,
      "tokens_seen": 677410816
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00040138415245737216,
      "loss": 2.658,
      "theoretical_loss": 3.7920645384965272,
      "tokens_seen": 677476352
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0004013741223671013,
      "loss": 2.928,
      "theoretical_loss": 3.7920270764689805,
      "tokens_seen": 677541888
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0004013640922768305,
      "loss": 2.8152,
      "theoretical_loss": 3.791989619079299,
      "tokens_seen": 677607424
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0004013540621865597,
      "loss": 2.8409,
      "theoretical_loss": 3.7919521663264604,
      "tokens_seen": 677672960
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0004013440320962889,
      "loss": 3.0376,
      "theoretical_loss": 3.7919147182094424,
      "tokens_seen": 677738496
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00040133400200601806,
      "loss": 2.7009,
      "theoretical_loss": 3.7918772747272227,
      "tokens_seen": 677804032
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00040132397191574724,
      "loss": 2.8787,
      "theoretical_loss": 3.79183983587878,
      "tokens_seen": 677869568
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0004013139418254764,
      "loss": 2.9441,
      "theoretical_loss": 3.7918024016630922,
      "tokens_seen": 677935104
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00040130391173520566,
      "loss": 2.6351,
      "theoretical_loss": 3.7917649720791395,
      "tokens_seen": 678000640
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0004012938816449348,
      "loss": 2.7128,
      "theoretical_loss": 3.7917275471259,
      "tokens_seen": 678066176
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.000401283851554664,
      "loss": 2.9271,
      "theoretical_loss": 3.791690126802354,
      "tokens_seen": 678131712
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00040127382146439315,
      "loss": 2.9554,
      "theoretical_loss": 3.7916527111074814,
      "tokens_seen": 678197248
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0004012637913741224,
      "loss": 2.9009,
      "theoretical_loss": 3.791615300040262,
      "tokens_seen": 678262784
    },
    {
      "epoch": 2.03,
      "objective/train/docs_used": 809644,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9782016277313232,
      "objective/train/theoretical_loss": 3.791596596241704,
      "objective/train/tokens_used": 698755552,
      "theoretical_loss": 3.791596596241704,
      "tokens_seen": 678295552
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00040125376128385157,
      "loss": 2.9855,
      "theoretical_loss": 3.7915778935996767,
      "tokens_seen": 678328320
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00040124373119358075,
      "loss": 3.0623,
      "theoretical_loss": 3.7915404917847066,
      "tokens_seen": 678393856
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00040123370110330993,
      "loss": 2.8576,
      "theoretical_loss": 3.791503094594333,
      "tokens_seen": 678459392
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00040122367101303916,
      "loss": 2.8939,
      "theoretical_loss": 3.791465702027537,
      "tokens_seen": 678524928
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0004012136409227683,
      "loss": 2.9493,
      "theoretical_loss": 3.7914283140833005,
      "tokens_seen": 678590464
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0004012036108324975,
      "loss": 2.5703,
      "theoretical_loss": 3.791390930760606,
      "tokens_seen": 678656000
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00040119358074222665,
      "loss": 2.9312,
      "theoretical_loss": 3.791353552058436,
      "tokens_seen": 678721536
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0004011835506519559,
      "loss": 2.6408,
      "theoretical_loss": 3.791316177975773,
      "tokens_seen": 678787072
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00040117352056168507,
      "loss": 2.7553,
      "theoretical_loss": 3.7912788085116005,
      "tokens_seen": 678852608
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00040116349047141425,
      "loss": 2.8303,
      "theoretical_loss": 3.7912414436649016,
      "tokens_seen": 678918144
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00040115346038114343,
      "loss": 3.0389,
      "theoretical_loss": 3.7912040834346605,
      "tokens_seen": 678983680
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0004011434302908726,
      "loss": 2.7881,
      "theoretical_loss": 3.7911667278198617,
      "tokens_seen": 679049216
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0004011334002006018,
      "loss": 2.6978,
      "theoretical_loss": 3.791129376819489,
      "tokens_seen": 679114752
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00040112337011033103,
      "loss": 2.9489,
      "theoretical_loss": 3.7910920304325275,
      "tokens_seen": 679180288
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00040111334002006016,
      "loss": 2.7048,
      "theoretical_loss": 3.7910546886579617,
      "tokens_seen": 679245824
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0004011033099297894,
      "loss": 2.9393,
      "theoretical_loss": 3.791017351494778,
      "tokens_seen": 679311360
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0004010932798395185,
      "loss": 2.9715,
      "theoretical_loss": 3.790980018941961,
      "tokens_seen": 679376896
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00040108324974924775,
      "loss": 2.8581,
      "theoretical_loss": 3.790942690998498,
      "tokens_seen": 679442432
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00040107321965897693,
      "loss": 2.5724,
      "theoretical_loss": 3.790905367663374,
      "tokens_seen": 679507968
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0004010631895687061,
      "loss": 2.8063,
      "theoretical_loss": 3.790868048935577,
      "tokens_seen": 679573504
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0004010531594784353,
      "loss": 3.0307,
      "theoretical_loss": 3.7908307348140933,
      "tokens_seen": 679639040
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00040104312938816453,
      "loss": 2.8895,
      "theoretical_loss": 3.7907934252979105,
      "tokens_seen": 679704576
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00040103309929789366,
      "loss": 2.9646,
      "theoretical_loss": 3.790756120386016,
      "tokens_seen": 679770112
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0004010230692076229,
      "loss": 2.8548,
      "theoretical_loss": 3.7907188200773985,
      "tokens_seen": 679835648
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.000401013039117352,
      "loss": 2.9081,
      "theoretical_loss": 3.7906815243710454,
      "tokens_seen": 679901184
    },
    {
      "epoch": 2.03,
      "objective/train/docs_used": 811111,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.683133363723755,
      "objective/train/theoretical_loss": 3.7906628782434018,
      "objective/train/tokens_used": 700393952,
      "theoretical_loss": 3.7906628782434018,
      "tokens_seen": 679933952
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00040100300902708126,
      "loss": 2.9855,
      "theoretical_loss": 3.7906442332659456,
      "tokens_seen": 679966720
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00040099297893681044,
      "loss": 2.7769,
      "theoretical_loss": 3.790606946761088,
      "tokens_seen": 680032256
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0004009829488465396,
      "loss": 2.7631,
      "theoretical_loss": 3.790569664855462,
      "tokens_seen": 680097792
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0004009729187562688,
      "loss": 3.0075,
      "theoretical_loss": 3.7905323875480574,
      "tokens_seen": 680163328
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.000400962888665998,
      "loss": 2.6681,
      "theoretical_loss": 3.7904951148378636,
      "tokens_seen": 680228864
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00040095285857572716,
      "loss": 3.1031,
      "theoretical_loss": 3.7904578467238714,
      "tokens_seen": 680294400
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0004009428284854564,
      "loss": 2.8202,
      "theoretical_loss": 3.790420583205071,
      "tokens_seen": 680359936
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0004009327983951855,
      "loss": 2.8716,
      "theoretical_loss": 3.7903833242804534,
      "tokens_seen": 680425472
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00040092276830491476,
      "loss": 2.6202,
      "theoretical_loss": 3.790346069949009,
      "tokens_seen": 680491008
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0004009127382146439,
      "loss": 2.8913,
      "theoretical_loss": 3.790308820209731,
      "tokens_seen": 680556544
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0004009027081243731,
      "loss": 3.0475,
      "theoretical_loss": 3.7902715750616087,
      "tokens_seen": 680622080
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00040089267803410236,
      "loss": 2.9008,
      "theoretical_loss": 3.7902343345036367,
      "tokens_seen": 680687616
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0004008826479438315,
      "loss": 2.9714,
      "theoretical_loss": 3.790197098534806,
      "tokens_seen": 680753152
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0004008726178535607,
      "loss": 2.8291,
      "theoretical_loss": 3.7901598671541104,
      "tokens_seen": 680818688
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0004008625877632899,
      "loss": 3.0035,
      "theoretical_loss": 3.7901226403605417,
      "tokens_seen": 680884224
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0004008525576730191,
      "loss": 2.8466,
      "theoretical_loss": 3.7900854181530943,
      "tokens_seen": 680949760
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00040084252758274826,
      "loss": 3.0043,
      "theoretical_loss": 3.7900482005307614,
      "tokens_seen": 681015296
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00040083249749247744,
      "loss": 2.7097,
      "theoretical_loss": 3.7900109874925376,
      "tokens_seen": 681080832
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0004008224674022066,
      "loss": 2.713,
      "theoretical_loss": 3.7899737790374166,
      "tokens_seen": 681146368
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00040081243731193586,
      "loss": 3.0035,
      "theoretical_loss": 3.7899365751643934,
      "tokens_seen": 681211904
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.000400802407221665,
      "loss": 2.9085,
      "theoretical_loss": 3.7898993758724635,
      "tokens_seen": 681277440
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0004007923771313942,
      "loss": 2.908,
      "theoretical_loss": 3.789862181160621,
      "tokens_seen": 681342976
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00040078234704112335,
      "loss": 2.8204,
      "theoretical_loss": 3.7898249910278627,
      "tokens_seen": 681408512
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0004007723169508526,
      "loss": 2.8935,
      "theoretical_loss": 3.7897878054731837,
      "tokens_seen": 681474048
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00040076228686058177,
      "loss": 3.0723,
      "theoretical_loss": 3.7897506244955808,
      "tokens_seen": 681539584
    },
    {
      "epoch": 2.03,
      "objective/train/docs_used": 811509,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8229663372039795,
      "objective/train/theoretical_loss": 3.789732035722869,
      "objective/train/tokens_used": 702032352,
      "theoretical_loss": 3.789732035722869,
      "tokens_seen": 681572352
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00040075225677031095,
      "loss": 2.7856,
      "theoretical_loss": 3.7897134480940506,
      "tokens_seen": 681605120
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00040074222668004013,
      "loss": 2.6028,
      "theoretical_loss": 3.7896762762675893,
      "tokens_seen": 681670656
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00040073219658976936,
      "loss": 2.8811,
      "theoretical_loss": 3.7896391090151944,
      "tokens_seen": 681736192
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0004007221664994985,
      "loss": 2.9482,
      "theoretical_loss": 3.789601946335864,
      "tokens_seen": 681801728
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0004007121364092277,
      "loss": 2.8387,
      "theoretical_loss": 3.7895647882285957,
      "tokens_seen": 681867264
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00040070210631895685,
      "loss": 2.8092,
      "theoretical_loss": 3.789527634692387,
      "tokens_seen": 681932800
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0004006920762286861,
      "loss": 2.9132,
      "theoretical_loss": 3.7894904857262373,
      "tokens_seen": 681998336
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00040068204613841527,
      "loss": 3.1188,
      "theoretical_loss": 3.7894533413291445,
      "tokens_seen": 682063872
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00040067201604814445,
      "loss": 3.0301,
      "theoretical_loss": 3.7894162015001087,
      "tokens_seen": 682129408
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00040066198595787363,
      "loss": 2.6942,
      "theoretical_loss": 3.789379066238128,
      "tokens_seen": 682194944
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0004006519558676028,
      "loss": 3.0303,
      "theoretical_loss": 3.789341935542204,
      "tokens_seen": 682260480
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.000400641925777332,
      "loss": 2.9624,
      "theoretical_loss": 3.789304809411335,
      "tokens_seen": 682326016
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00040063189568706123,
      "loss": 2.893,
      "theoretical_loss": 3.789267687844522,
      "tokens_seen": 682391552
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00040062186559679036,
      "loss": 2.8456,
      "theoretical_loss": 3.7892305708407656,
      "tokens_seen": 682457088
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0004006118355065196,
      "loss": 2.7705,
      "theoretical_loss": 3.789193458399067,
      "tokens_seen": 682522624
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0004006018054162487,
      "loss": 3.0149,
      "theoretical_loss": 3.789156350518428,
      "tokens_seen": 682588160
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00040059177532597795,
      "loss": 2.8807,
      "theoretical_loss": 3.7891192471978488,
      "tokens_seen": 682653696
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00040058174523570713,
      "loss": 2.9405,
      "theoretical_loss": 3.789082148436332,
      "tokens_seen": 682719232
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0004005717151454363,
      "loss": 2.8325,
      "theoretical_loss": 3.7890450542328806,
      "tokens_seen": 682784768
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0004005616850551655,
      "loss": 2.9391,
      "theoretical_loss": 3.7890079645864967,
      "tokens_seen": 682850304
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00040055165496489473,
      "loss": 3.1444,
      "theoretical_loss": 3.7889708794961825,
      "tokens_seen": 682915840
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00040054162487462386,
      "loss": 2.9871,
      "theoretical_loss": 3.788933798960942,
      "tokens_seen": 682981376
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0004005315947843531,
      "loss": 2.8059,
      "theoretical_loss": 3.7888967229797785,
      "tokens_seen": 683046912
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0004005215646940822,
      "loss": 2.5791,
      "theoretical_loss": 3.7888596515516952,
      "tokens_seen": 683112448
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00040051153460381146,
      "loss": 3.0333,
      "theoretical_loss": 3.7888225846756978,
      "tokens_seen": 683177984
    },
    {
      "epoch": 2.03,
      "objective/train/docs_used": 812647,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.369586229324341,
      "objective/train/theoretical_loss": 3.7888040529444194,
      "objective/train/tokens_used": 703670752,
      "theoretical_loss": 3.7888040529444194,
      "tokens_seen": 683210752
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00040050150451354064,
      "loss": 2.8012,
      "theoretical_loss": 3.788785522350789,
      "tokens_seen": 683243520
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0004004914744232698,
      "loss": 3.2464,
      "theoretical_loss": 3.7887484645759746,
      "tokens_seen": 683309056
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.000400481444332999,
      "loss": 2.9848,
      "theoretical_loss": 3.788711411350259,
      "tokens_seen": 683374592
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0004004714142427282,
      "loss": 2.8967,
      "theoretical_loss": 3.7886743626726487,
      "tokens_seen": 683440128
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00040046138415245736,
      "loss": 2.9512,
      "theoretical_loss": 3.788637318542148,
      "tokens_seen": 683505664
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0004004513540621866,
      "loss": 2.9431,
      "theoretical_loss": 3.7886002789577633,
      "tokens_seen": 683571200
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0004004413239719157,
      "loss": 2.8512,
      "theoretical_loss": 3.788563243918502,
      "tokens_seen": 683636736
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00040043129388164496,
      "loss": 2.7822,
      "theoretical_loss": 3.7885262134233697,
      "tokens_seen": 683702272
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0004004212637913741,
      "loss": 2.977,
      "theoretical_loss": 3.7884891874713738,
      "tokens_seen": 683767808
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0004004112337011033,
      "loss": 2.7165,
      "theoretical_loss": 3.7884521660615205,
      "tokens_seen": 683833344
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0004004012036108325,
      "loss": 2.6355,
      "theoretical_loss": 3.7884151491928186,
      "tokens_seen": 683898880
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0004003911735205617,
      "loss": 2.8597,
      "theoretical_loss": 3.788378136864276,
      "tokens_seen": 683964416
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00040038114343029087,
      "loss": 2.9315,
      "theoretical_loss": 3.7883411290749005,
      "tokens_seen": 684029952
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0004003711133400201,
      "loss": 2.7077,
      "theoretical_loss": 3.7883041258237,
      "tokens_seen": 684095488
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00040036108324974923,
      "loss": 2.912,
      "theoretical_loss": 3.7882671271096844,
      "tokens_seen": 684161024
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00040035105315947846,
      "loss": 2.9169,
      "theoretical_loss": 3.7882301329318624,
      "tokens_seen": 684226560
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0004003410230692076,
      "loss": 2.8847,
      "theoretical_loss": 3.788193143289243,
      "tokens_seen": 684292096
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0004003309929789368,
      "loss": 2.8739,
      "theoretical_loss": 3.7881561581808363,
      "tokens_seen": 684357632
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.000400320962888666,
      "loss": 2.8555,
      "theoretical_loss": 3.788119177605653,
      "tokens_seen": 684423168
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0004003109327983952,
      "loss": 2.7767,
      "theoretical_loss": 3.7880822015627023,
      "tokens_seen": 684488704
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00040030090270812437,
      "loss": 2.8128,
      "theoretical_loss": 3.788045230050996,
      "tokens_seen": 684554240
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00040029087261785355,
      "loss": 2.9026,
      "theoretical_loss": 3.788008263069544,
      "tokens_seen": 684619776
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00040028084252758273,
      "loss": 2.7174,
      "theoretical_loss": 3.7879713006173583,
      "tokens_seen": 684685312
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00040027081243731197,
      "loss": 2.9723,
      "theoretical_loss": 3.787934342693451,
      "tokens_seen": 684750848
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0004002607823470411,
      "loss": 3.017,
      "theoretical_loss": 3.787897389296833,
      "tokens_seen": 684816384
    },
    {
      "epoch": 2.03,
      "objective/train/docs_used": 812952,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.962841510772705,
      "objective/train/theoretical_loss": 3.787878914295949,
      "objective/train/tokens_used": 705309152,
      "theoretical_loss": 3.787878914295949,
      "tokens_seen": 684849152
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00040025075225677033,
      "loss": 2.9822,
      "theoretical_loss": 3.787860440426517,
      "tokens_seen": 684881920
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00040024072216649946,
      "loss": 3.1501,
      "theoretical_loss": 3.787823496081515,
      "tokens_seen": 684947456
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0004002306920762287,
      "loss": 3.0707,
      "theoretical_loss": 3.787786556260841,
      "tokens_seen": 685012992
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00040022066198595787,
      "loss": 2.8825,
      "theoretical_loss": 3.7877496209635075,
      "tokens_seen": 685078528
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00040021063189568705,
      "loss": 3.1118,
      "theoretical_loss": 3.7877126901885276,
      "tokens_seen": 685144064
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00040020060180541623,
      "loss": 3.1401,
      "theoretical_loss": 3.7876757639349163,
      "tokens_seen": 685209600
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00040019057171514547,
      "loss": 3.1658,
      "theoretical_loss": 3.787638842201686,
      "tokens_seen": 685275136
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0004001805416248746,
      "loss": 3.1845,
      "theoretical_loss": 3.7876019249878525,
      "tokens_seen": 685340672
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00040017051153460383,
      "loss": 2.8411,
      "theoretical_loss": 3.78756501229243,
      "tokens_seen": 685406208
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00040016048144433296,
      "loss": 3.0347,
      "theoretical_loss": 3.7875281041144335,
      "tokens_seen": 685471744
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0004001504513540622,
      "loss": 3.0841,
      "theoretical_loss": 3.787491200452878,
      "tokens_seen": 685537280
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00040014042126379143,
      "loss": 2.7852,
      "theoretical_loss": 3.78745430130678,
      "tokens_seen": 685602816
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00040013039117352056,
      "loss": 2.7576,
      "theoretical_loss": 3.7874174066751545,
      "tokens_seen": 685668352
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0004001203610832498,
      "loss": 3.0064,
      "theoretical_loss": 3.7873805165570182,
      "tokens_seen": 685733888
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0004001103309929789,
      "loss": 2.9897,
      "theoretical_loss": 3.7873436309513884,
      "tokens_seen": 685799424
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00040010030090270815,
      "loss": 3.0036,
      "theoretical_loss": 3.7873067498572808,
      "tokens_seen": 685864960
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00040009027081243733,
      "loss": 2.9462,
      "theoretical_loss": 3.787269873273713,
      "tokens_seen": 685930496
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0004000802407221665,
      "loss": 3.017,
      "theoretical_loss": 3.787233001199703,
      "tokens_seen": 685996032
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0004000702106318957,
      "loss": 3.1799,
      "theoretical_loss": 3.7871961336342674,
      "tokens_seen": 686061568
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00040006018054162493,
      "loss": 3.0584,
      "theoretical_loss": 3.787159270576425,
      "tokens_seen": 686127104
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00040005015045135406,
      "loss": 2.837,
      "theoretical_loss": 3.787122412025195,
      "tokens_seen": 686192640
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0004000401203610833,
      "loss": 3.0278,
      "theoretical_loss": 3.7870855579795952,
      "tokens_seen": 686258176
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0004000300902708124,
      "loss": 2.9139,
      "theoretical_loss": 3.787048708438644,
      "tokens_seen": 686323712
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00040002006018054166,
      "loss": 3.0253,
      "theoretical_loss": 3.7870118634013625,
      "tokens_seen": 686389248
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00040001003009027084,
      "loss": 2.9288,
      "theoretical_loss": 3.786975022866769,
      "tokens_seen": 686454784
    },
    {
      "epoch": 2.03,
      "objective/train/docs_used": 812952,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6860439777374268,
      "objective/train/theoretical_loss": 3.786956604287674,
      "objective/train/tokens_used": 706947552,
      "theoretical_loss": 3.786956604287674,
      "tokens_seen": 686487552
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0004,
      "loss": 3.0279,
      "theoretical_loss": 3.786938186833884,
      "tokens_seen": 686520320
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003999899699097292,
      "loss": 3.0499,
      "theoretical_loss": 3.7869013553017274,
      "tokens_seen": 686585856
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003999799398194584,
      "loss": 3.1523,
      "theoretical_loss": 3.7868645282693203,
      "tokens_seen": 686651392
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039996990972918756,
      "loss": 2.8675,
      "theoretical_loss": 3.786827705735683,
      "tokens_seen": 686716928
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003999598796389168,
      "loss": 2.8969,
      "theoretical_loss": 3.7867908876998366,
      "tokens_seen": 686782464
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003999498495486459,
      "loss": 3.0232,
      "theoretical_loss": 3.786754074160803,
      "tokens_seen": 686848000
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039993981945837516,
      "loss": 3.1544,
      "theoretical_loss": 3.7867172651176038,
      "tokens_seen": 686913536
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003999297893681043,
      "loss": 3.3209,
      "theoretical_loss": 3.7866804605692614,
      "tokens_seen": 686979072
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003999197592778335,
      "loss": 2.9485,
      "theoretical_loss": 3.786643660514798,
      "tokens_seen": 687044608
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003999097291875627,
      "loss": 2.9027,
      "theoretical_loss": 3.786606864953236,
      "tokens_seen": 687110144
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003998996990972919,
      "loss": 3.0381,
      "theoretical_loss": 3.786570073883599,
      "tokens_seen": 687175680
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039988966900702107,
      "loss": 3.0156,
      "theoretical_loss": 3.78653328730491,
      "tokens_seen": 687241216
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003998796389167503,
      "loss": 3.2087,
      "theoretical_loss": 3.7864965052161925,
      "tokens_seen": 687306752
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039986960882647943,
      "loss": 3.0047,
      "theoretical_loss": 3.7864597276164704,
      "tokens_seen": 687372288
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039985957873620866,
      "loss": 3.0319,
      "theoretical_loss": 3.7864229545047685,
      "tokens_seen": 687437824
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003998495486459378,
      "loss": 3.2716,
      "theoretical_loss": 3.7863861858801107,
      "tokens_seen": 687503360
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.000399839518555667,
      "loss": 2.8685,
      "theoretical_loss": 3.786349421741522,
      "tokens_seen": 687568896
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003998294884653962,
      "loss": 2.9713,
      "theoretical_loss": 3.7863126620880276,
      "tokens_seen": 687634432
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003998194583751254,
      "loss": 2.9644,
      "theoretical_loss": 3.786275906918653,
      "tokens_seen": 687699968
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039980942828485457,
      "loss": 2.9327,
      "theoretical_loss": 3.7862391562324245,
      "tokens_seen": 687765504
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039979939819458375,
      "loss": 2.8463,
      "theoretical_loss": 3.786202410028367,
      "tokens_seen": 687831040
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039978936810431293,
      "loss": 3.209,
      "theoretical_loss": 3.7861656683055074,
      "tokens_seen": 687896576
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039977933801404217,
      "loss": 3.247,
      "theoretical_loss": 3.7861289310628727,
      "tokens_seen": 687962112
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003997693079237713,
      "loss": 3.0825,
      "theoretical_loss": 3.7860921982994897,
      "tokens_seen": 688027648
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039975927783350053,
      "loss": 2.9732,
      "theoretical_loss": 3.786055470014386,
      "tokens_seen": 688093184
    },
    {
      "epoch": 2.03,
      "objective/train/docs_used": 813708,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.2107160091400146,
      "objective/train/theoretical_loss": 3.786037107550884,
      "objective/train/tokens_used": 708585952,
      "theoretical_loss": 3.786037107550884,
      "tokens_seen": 688125952
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039974924774322966,
      "loss": 3.1854,
      "theoretical_loss": 3.7860187462065875,
      "tokens_seen": 688158720
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003997392176529589,
      "loss": 3.036,
      "theoretical_loss": 3.7859820268751245,
      "tokens_seen": 688224256
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039972918756268807,
      "loss": 2.9536,
      "theoretical_loss": 3.785945312019024,
      "tokens_seen": 688289792
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039971915747241725,
      "loss": 2.7898,
      "theoretical_loss": 3.7859086016373134,
      "tokens_seen": 688355328
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039970912738214643,
      "loss": 2.9671,
      "theoretical_loss": 3.785871895729024,
      "tokens_seen": 688420864
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039969909729187567,
      "loss": 2.9539,
      "theoretical_loss": 3.785835194293183,
      "tokens_seen": 688486400
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003996890672016048,
      "loss": 2.9139,
      "theoretical_loss": 3.7857984973288206,
      "tokens_seen": 688551936
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039967903711133403,
      "loss": 3.1913,
      "theoretical_loss": 3.7857618048349657,
      "tokens_seen": 688617472
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039966900702106316,
      "loss": 3.1764,
      "theoretical_loss": 3.7857251168106494,
      "tokens_seen": 688683008
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003996589769307924,
      "loss": 2.9506,
      "theoretical_loss": 3.7856884332549017,
      "tokens_seen": 688748544
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003996489468405216,
      "loss": 3.1994,
      "theoretical_loss": 3.7856517541667523,
      "tokens_seen": 688814080
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039963891675025076,
      "loss": 3.0963,
      "theoretical_loss": 3.7856150795452335,
      "tokens_seen": 688879616
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039962888665997994,
      "loss": 2.9912,
      "theoretical_loss": 3.7855784093893754,
      "tokens_seen": 688945152
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003996188565697091,
      "loss": 3.1565,
      "theoretical_loss": 3.78554174369821,
      "tokens_seen": 689010688
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003996088264794383,
      "loss": 2.9481,
      "theoretical_loss": 3.7855050824707694,
      "tokens_seen": 689076224
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039959879638916754,
      "loss": 2.9037,
      "theoretical_loss": 3.7854684257060853,
      "tokens_seen": 689141760
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039958876629889666,
      "loss": 3.0422,
      "theoretical_loss": 3.7854317734031904,
      "tokens_seen": 689207296
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003995787362086259,
      "loss": 3.0962,
      "theoretical_loss": 3.7853951255611165,
      "tokens_seen": 689272832
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003995687061183551,
      "loss": 2.9324,
      "theoretical_loss": 3.785358482178898,
      "tokens_seen": 689338368
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039955867602808426,
      "loss": 3.0348,
      "theoretical_loss": 3.7853218432555678,
      "tokens_seen": 689403904
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039954864593781344,
      "loss": 2.9487,
      "theoretical_loss": 3.7852852087901594,
      "tokens_seen": 689469440
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003995386158475426,
      "loss": 3.1092,
      "theoretical_loss": 3.7852485787817063,
      "tokens_seen": 689534976
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003995285857572718,
      "loss": 2.9527,
      "theoretical_loss": 3.7852119532292434,
      "tokens_seen": 689600512
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039951855566700104,
      "loss": 3.163,
      "theoretical_loss": 3.785175332131805,
      "tokens_seen": 689666048
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039950852557673017,
      "loss": 2.9854,
      "theoretical_loss": 3.7851387154884257,
      "tokens_seen": 689731584
    },
    {
      "epoch": 2.03,
      "objective/train/docs_used": 815148,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.4033315181732178,
      "objective/train/theoretical_loss": 3.7851204088367068,
      "objective/train/tokens_used": 710224352,
      "theoretical_loss": 3.7851204088367068,
      "tokens_seen": 689764352
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003994984954864594,
      "loss": 3.257,
      "theoretical_loss": 3.785102103298141,
      "tokens_seen": 689797120
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039948846539618853,
      "loss": 3.0081,
      "theoretical_loss": 3.7850654955599863,
      "tokens_seen": 689862656
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039947843530591776,
      "loss": 2.9746,
      "theoretical_loss": 3.7850288922729973,
      "tokens_seen": 689928192
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039946840521564694,
      "loss": 2.9742,
      "theoretical_loss": 3.78499229343621,
      "tokens_seen": 689993728
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003994583751253761,
      "loss": 3.0874,
      "theoretical_loss": 3.784955699048661,
      "tokens_seen": 690059264
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003994483450351053,
      "loss": 3.1328,
      "theoretical_loss": 3.784919109109386,
      "tokens_seen": 690124800
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003994383149448345,
      "loss": 2.89,
      "theoretical_loss": 3.784882523617423,
      "tokens_seen": 690190336
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039942828485456367,
      "loss": 2.852,
      "theoretical_loss": 3.784845942571809,
      "tokens_seen": 690255872
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003994182547642929,
      "loss": 3.2457,
      "theoretical_loss": 3.784809365971581,
      "tokens_seen": 690321408
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039940822467402203,
      "loss": 2.9565,
      "theoretical_loss": 3.784772793815777,
      "tokens_seen": 690386944
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039939819458375127,
      "loss": 2.902,
      "theoretical_loss": 3.784736226103436,
      "tokens_seen": 690452480
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003993881644934805,
      "loss": 3.042,
      "theoretical_loss": 3.784699662833596,
      "tokens_seen": 690518016
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039937813440320963,
      "loss": 3.0018,
      "theoretical_loss": 3.784663104005295,
      "tokens_seen": 690583552
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039936810431293886,
      "loss": 2.8559,
      "theoretical_loss": 3.784626549617573,
      "tokens_seen": 690649088
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.000399358074222668,
      "loss": 3.0771,
      "theoretical_loss": 3.7845899996694685,
      "tokens_seen": 690714624
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003993480441323972,
      "loss": 3.186,
      "theoretical_loss": 3.7845534541600223,
      "tokens_seen": 690780160
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003993380140421264,
      "loss": 2.9598,
      "theoretical_loss": 3.7845169130882734,
      "tokens_seen": 690845696
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003993279839518556,
      "loss": 3.0254,
      "theoretical_loss": 3.7844803764532617,
      "tokens_seen": 690911232
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039931795386158477,
      "loss": 2.7558,
      "theoretical_loss": 3.784443844254029,
      "tokens_seen": 690976768
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039930792377131395,
      "loss": 3.0946,
      "theoretical_loss": 3.784407316489615,
      "tokens_seen": 691042304
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039929789368104313,
      "loss": 2.9042,
      "theoretical_loss": 3.784370793159061,
      "tokens_seen": 691107840
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039928786359077237,
      "loss": 3.007,
      "theoretical_loss": 3.784334274261409,
      "tokens_seen": 691173376
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003992778335005015,
      "loss": 3.001,
      "theoretical_loss": 3.7842977597957006,
      "tokens_seen": 691238912
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039926780341023073,
      "loss": 3.1521,
      "theoretical_loss": 3.784261249760977,
      "tokens_seen": 691304448
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039925777331995986,
      "loss": 3.0055,
      "theoretical_loss": 3.784224744156282,
      "tokens_seen": 691369984
    },
    {
      "epoch": 2.03,
      "objective/train/docs_used": 815781,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.672840118408203,
      "objective/train/theoretical_loss": 3.7842064930148953,
      "objective/train/tokens_used": 711862752,
      "theoretical_loss": 3.7842064930148953,
      "tokens_seen": 691402752
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003992477432296891,
      "loss": 3.1546,
      "theoretical_loss": 3.784188242980657,
      "tokens_seen": 691435520
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039923771313941827,
      "loss": 2.9453,
      "theoretical_loss": 3.784151746233145,
      "tokens_seen": 691501056
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039922768304914745,
      "loss": 2.872,
      "theoretical_loss": 3.78411525391279,
      "tokens_seen": 691566592
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039921765295887663,
      "loss": 3.0222,
      "theoretical_loss": 3.7840787660186344,
      "tokens_seen": 691632128
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039920762286860587,
      "loss": 2.9783,
      "theoretical_loss": 3.7840422825497235,
      "tokens_seen": 691697664
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.000399197592778335,
      "loss": 2.9824,
      "theoretical_loss": 3.7840058035050994,
      "tokens_seen": 691763200
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039918756268806423,
      "loss": 3.1181,
      "theoretical_loss": 3.7839693288838085,
      "tokens_seen": 691828736
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039917753259779336,
      "loss": 3.0017,
      "theoretical_loss": 3.783932858684895,
      "tokens_seen": 691894272
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003991675025075226,
      "loss": 2.8385,
      "theoretical_loss": 3.783896392907402,
      "tokens_seen": 691959808
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003991574724172518,
      "loss": 3.1507,
      "theoretical_loss": 3.7838599315503774,
      "tokens_seen": 692025344
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039914744232698096,
      "loss": 2.9381,
      "theoretical_loss": 3.7838234746128654,
      "tokens_seen": 692090880
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039913741223671014,
      "loss": 3.2148,
      "theoretical_loss": 3.7837870220939127,
      "tokens_seen": 692156416
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003991273821464393,
      "loss": 3.0144,
      "theoretical_loss": 3.783750573992565,
      "tokens_seen": 692221952
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003991173520561685,
      "loss": 3.1139,
      "theoretical_loss": 3.7837141303078683,
      "tokens_seen": 692287488
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039910732196589774,
      "loss": 3.0469,
      "theoretical_loss": 3.7836776910388696,
      "tokens_seen": 692353024
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039909729187562686,
      "loss": 2.8595,
      "theoretical_loss": 3.783641256184617,
      "tokens_seen": 692418560
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003990872617853561,
      "loss": 3.0581,
      "theoretical_loss": 3.7836048257441566,
      "tokens_seen": 692484096
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003990772316950853,
      "loss": 3.0968,
      "theoretical_loss": 3.7835683997165366,
      "tokens_seen": 692549632
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039906720160481446,
      "loss": 2.6969,
      "theoretical_loss": 3.7835319781008048,
      "tokens_seen": 692615168
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039905717151454364,
      "loss": 2.973,
      "theoretical_loss": 3.78349556089601,
      "tokens_seen": 692680704
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003990471414242728,
      "loss": 2.8064,
      "theoretical_loss": 3.7834591481011994,
      "tokens_seen": 692746240
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.000399037111334002,
      "loss": 3.1452,
      "theoretical_loss": 3.783422739715424,
      "tokens_seen": 692811776
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039902708124373124,
      "loss": 3.2073,
      "theoretical_loss": 3.7833863357377306,
      "tokens_seen": 692877312
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039901705115346037,
      "loss": 2.7702,
      "theoretical_loss": 3.7833499361671703,
      "tokens_seen": 692942848
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003990070210631896,
      "loss": 2.927,
      "theoretical_loss": 3.7833135410027925,
      "tokens_seen": 693008384
    },
    {
      "epoch": 2.03,
      "objective/train/docs_used": 817009,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7722556591033936,
      "objective/train/theoretical_loss": 3.7832953450726245,
      "objective/train/tokens_used": 713501152,
      "theoretical_loss": 3.7832953450726245,
      "tokens_seen": 693041152
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039899699097291873,
      "loss": 3.0062,
      "theoretical_loss": 3.783277150243646,
      "tokens_seen": 693073920
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039898696088264796,
      "loss": 3.0559,
      "theoretical_loss": 3.7832407638887826,
      "tokens_seen": 693139456
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039897693079237714,
      "loss": 3.0706,
      "theoretical_loss": 3.783204381937253,
      "tokens_seen": 693204992
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003989669007021063,
      "loss": 2.9877,
      "theoretical_loss": 3.7831680043881066,
      "tokens_seen": 693270528
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003989568706118355,
      "loss": 2.8072,
      "theoretical_loss": 3.7831316312403964,
      "tokens_seen": 693336064
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003989468405215647,
      "loss": 2.7801,
      "theoretical_loss": 3.7830952624931724,
      "tokens_seen": 693401600
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039893681043129387,
      "loss": 3.0159,
      "theoretical_loss": 3.7830588981454873,
      "tokens_seen": 693467136
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003989267803410231,
      "loss": 2.9373,
      "theoretical_loss": 3.783022538196393,
      "tokens_seen": 693532672
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039891675025075223,
      "loss": 2.9548,
      "theoretical_loss": 3.782986182644941,
      "tokens_seen": 693598208
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039890672016048147,
      "loss": 3.2316,
      "theoretical_loss": 3.7829498314901855,
      "tokens_seen": 693663744
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039889669007021065,
      "loss": 3.1027,
      "theoretical_loss": 3.7829134847311785,
      "tokens_seen": 693729280
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039888665997993983,
      "loss": 3.2061,
      "theoretical_loss": 3.782877142366974,
      "tokens_seen": 693794816
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.000398876629889669,
      "loss": 3.1499,
      "theoretical_loss": 3.782840804396624,
      "tokens_seen": 693860352
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003988665997993982,
      "loss": 3.1008,
      "theoretical_loss": 3.782804470819184,
      "tokens_seen": 693925888
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039885656970912737,
      "loss": 3.0239,
      "theoretical_loss": 3.782768141633708,
      "tokens_seen": 693991424
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003988465396188566,
      "loss": 3.1596,
      "theoretical_loss": 3.7827318168392488,
      "tokens_seen": 694056960
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039883650952858573,
      "loss": 2.9416,
      "theoretical_loss": 3.782695496434863,
      "tokens_seen": 694122496
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039882647943831497,
      "loss": 2.9495,
      "theoretical_loss": 3.782659180419605,
      "tokens_seen": 694188032
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003988164493480441,
      "loss": 2.9938,
      "theoretical_loss": 3.7826228687925294,
      "tokens_seen": 694253568
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039880641925777333,
      "loss": 2.9565,
      "theoretical_loss": 3.7825865615526926,
      "tokens_seen": 694319104
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003987963891675025,
      "loss": 3.0452,
      "theoretical_loss": 3.7825502586991506,
      "tokens_seen": 694384640
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003987863590772317,
      "loss": 2.9098,
      "theoretical_loss": 3.782513960230959,
      "tokens_seen": 694450176
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003987763289869609,
      "loss": 2.9125,
      "theoretical_loss": 3.7824776661471744,
      "tokens_seen": 694515712
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039876629889669006,
      "loss": 2.9881,
      "theoretical_loss": 3.7824413764468536,
      "tokens_seen": 694581248
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039875626880641924,
      "loss": 3.0315,
      "theoretical_loss": 3.7824050911290543,
      "tokens_seen": 694646784
    },
    {
      "epoch": 2.03,
      "objective/train/docs_used": 817800,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.937741279602051,
      "objective/train/theoretical_loss": 3.782386950113305,
      "objective/train/tokens_used": 715139552,
      "theoretical_loss": 3.782386950113305,
      "tokens_seen": 694679552
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039874623871614847,
      "loss": 2.7712,
      "theoretical_loss": 3.782368810192833,
      "tokens_seen": 694712320
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003987362086258776,
      "loss": 3.0905,
      "theoretical_loss": 3.782332533637248,
      "tokens_seen": 694777856
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039872617853560683,
      "loss": 3.1504,
      "theoretical_loss": 3.7822962614613567,
      "tokens_seen": 694843392
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.000398716148445336,
      "loss": 3.1779,
      "theoretical_loss": 3.7822599936642174,
      "tokens_seen": 694908928
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003987061183550652,
      "loss": 2.9834,
      "theoretical_loss": 3.7822237302448887,
      "tokens_seen": 694974464
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003986960882647944,
      "loss": 2.9804,
      "theoretical_loss": 3.78218747120243,
      "tokens_seen": 695040000
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039868605817452356,
      "loss": 2.9168,
      "theoretical_loss": 3.7821512165358993,
      "tokens_seen": 695105536
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039867602808425274,
      "loss": 3.0235,
      "theoretical_loss": 3.782114966244357,
      "tokens_seen": 695171072
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.000398665997993982,
      "loss": 2.9155,
      "theoretical_loss": 3.7820787203268624,
      "tokens_seen": 695236608
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003986559679037111,
      "loss": 3.0621,
      "theoretical_loss": 3.7820424787824747,
      "tokens_seen": 695302144
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039864593781344034,
      "loss": 2.9857,
      "theoretical_loss": 3.7820062416102553,
      "tokens_seen": 695367680
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003986359077231695,
      "loss": 2.9637,
      "theoretical_loss": 3.7819700088092647,
      "tokens_seen": 695433216
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003986258776328987,
      "loss": 2.8773,
      "theoretical_loss": 3.781933780378563,
      "tokens_seen": 695498752
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039861584754262794,
      "loss": 2.8766,
      "theoretical_loss": 3.781897556317212,
      "tokens_seen": 695564288
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039860581745235706,
      "loss": 2.9318,
      "theoretical_loss": 3.781861336624273,
      "tokens_seen": 695629824
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003985957873620863,
      "loss": 2.9717,
      "theoretical_loss": 3.7818251212988074,
      "tokens_seen": 695695360
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003985857572718155,
      "loss": 2.9584,
      "theoretical_loss": 3.781788910339877,
      "tokens_seen": 695760896
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039857572718154466,
      "loss": 2.743,
      "theoretical_loss": 3.781752703746545,
      "tokens_seen": 695826432
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039856569709127384,
      "loss": 3.0222,
      "theoretical_loss": 3.7817165015178733,
      "tokens_seen": 695891968
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.000398555667001003,
      "loss": 3.1952,
      "theoretical_loss": 3.7816803036529247,
      "tokens_seen": 695957504
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003985456369107322,
      "loss": 2.9163,
      "theoretical_loss": 3.781644110150763,
      "tokens_seen": 696023040
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039853560682046144,
      "loss": 3.0045,
      "theoretical_loss": 3.781607921010451,
      "tokens_seen": 696088576
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039852557673019057,
      "loss": 2.875,
      "theoretical_loss": 3.781571736231053,
      "tokens_seen": 696154112
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003985155466399198,
      "loss": 3.1498,
      "theoretical_loss": 3.7815355558116326,
      "tokens_seen": 696219648
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039850551654964893,
      "loss": 3.0813,
      "theoretical_loss": 3.7814993797512546,
      "tokens_seen": 696285184
    },
    {
      "epoch": 2.03,
      "objective/train/docs_used": 823060,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.890718460083008,
      "objective/train/theoretical_loss": 3.7814812933554136,
      "objective/train/tokens_used": 716777952,
      "theoretical_loss": 3.7814812933554136,
      "tokens_seen": 696317952
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039849548645937816,
      "loss": 2.9396,
      "theoretical_loss": 3.781463208048983,
      "tokens_seen": 696350720
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039848545636910734,
      "loss": 2.9903,
      "theoretical_loss": 3.781427040703883,
      "tokens_seen": 696416256
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003984754262788365,
      "loss": 2.9655,
      "theoretical_loss": 3.781390877715019,
      "tokens_seen": 696481792
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003984653961885657,
      "loss": 2.9875,
      "theoretical_loss": 3.781354719081458,
      "tokens_seen": 696547328
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003984553660982949,
      "loss": 2.7848,
      "theoretical_loss": 3.7813185648022647,
      "tokens_seen": 696612864
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039844533600802407,
      "loss": 2.7708,
      "theoretical_loss": 3.7812824148765056,
      "tokens_seen": 696678400
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003984353059177533,
      "loss": 2.9602,
      "theoretical_loss": 3.781246269303247,
      "tokens_seen": 696743936
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039842527582748243,
      "loss": 2.9896,
      "theoretical_loss": 3.7812101280815553,
      "tokens_seen": 696809472
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039841524573721167,
      "loss": 2.9217,
      "theoretical_loss": 3.7811739912104976,
      "tokens_seen": 696875008
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039840521564694085,
      "loss": 2.9529,
      "theoretical_loss": 3.781137858689141,
      "tokens_seen": 696940544
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039839518555667003,
      "loss": 3.0855,
      "theoretical_loss": 3.781101730516553,
      "tokens_seen": 697006080
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003983851554663992,
      "loss": 3.0658,
      "theoretical_loss": 3.7810656066918016,
      "tokens_seen": 697071616
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003983751253761284,
      "loss": 2.76,
      "theoretical_loss": 3.7810294872139547,
      "tokens_seen": 697137152
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039836509528585757,
      "loss": 2.8135,
      "theoretical_loss": 3.7809933720820803,
      "tokens_seen": 697202688
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003983550651955868,
      "loss": 3.0346,
      "theoretical_loss": 3.7809572612952476,
      "tokens_seen": 697268224
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039834503510531593,
      "loss": 2.9071,
      "theoretical_loss": 3.7809211548525257,
      "tokens_seen": 697333760
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039833500501504517,
      "loss": 3.0486,
      "theoretical_loss": 3.7808850527529825,
      "tokens_seen": 697399296
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003983249749247743,
      "loss": 2.8335,
      "theoretical_loss": 3.7808489549956894,
      "tokens_seen": 697464832
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039831494483450353,
      "loss": 3.1494,
      "theoretical_loss": 3.7808128615797143,
      "tokens_seen": 697530368
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003983049147442327,
      "loss": 2.9519,
      "theoretical_loss": 3.780776772504129,
      "tokens_seen": 697595904
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003982948846539619,
      "loss": 2.9174,
      "theoretical_loss": 3.7807406877680023,
      "tokens_seen": 697661440
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003982848545636911,
      "loss": 2.8904,
      "theoretical_loss": 3.7807046073704065,
      "tokens_seen": 697726976
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039827482447342026,
      "loss": 2.9245,
      "theoretical_loss": 3.780668531310411,
      "tokens_seen": 697792512
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039826479438314944,
      "loss": 3.0444,
      "theoretical_loss": 3.7806324595870877,
      "tokens_seen": 697858048
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039825476429287867,
      "loss": 2.9878,
      "theoretical_loss": 3.7805963921995085,
      "tokens_seen": 697923584
    },
    {
      "epoch": 2.03,
      "objective/train/docs_used": 828245,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.0669054985046387,
      "objective/train/theoretical_loss": 3.7805783601313325,
      "objective/train/tokens_used": 718416352,
      "theoretical_loss": 3.7805783601313325,
      "tokens_seen": 697956352
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003982447342026078,
      "loss": 3.0409,
      "theoretical_loss": 3.7805603291467444,
      "tokens_seen": 697989120
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039823470411233703,
      "loss": 3.0377,
      "theoretical_loss": 3.7805242704278683,
      "tokens_seen": 698054656
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003982246740220662,
      "loss": 2.7453,
      "theoretical_loss": 3.7804882160419515,
      "tokens_seen": 698120192
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003982146439317954,
      "loss": 3.0153,
      "theoretical_loss": 3.780452165988067,
      "tokens_seen": 698185728
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003982046138415246,
      "loss": 2.937,
      "theoretical_loss": 3.780416120265289,
      "tokens_seen": 698251264
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039819458375125376,
      "loss": 2.9108,
      "theoretical_loss": 3.780380078872689,
      "tokens_seen": 698316800
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039818455366098294,
      "loss": 2.7331,
      "theoretical_loss": 3.7803440418093417,
      "tokens_seen": 698382336
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003981745235707122,
      "loss": 3.0641,
      "theoretical_loss": 3.78030800907432,
      "tokens_seen": 698447872
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003981644934804413,
      "loss": 2.9069,
      "theoretical_loss": 3.780271980666699,
      "tokens_seen": 698513408
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039815446339017054,
      "loss": 2.8881,
      "theoretical_loss": 3.780235956585552,
      "tokens_seen": 698578944
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039814443329989966,
      "loss": 3.1442,
      "theoretical_loss": 3.7801999368299537,
      "tokens_seen": 698644480
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003981344032096289,
      "loss": 2.9341,
      "theoretical_loss": 3.78016392139898,
      "tokens_seen": 698710016
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003981243731193581,
      "loss": 2.6563,
      "theoretical_loss": 3.7801279102917063,
      "tokens_seen": 698775552
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039811434302908726,
      "loss": 3.1938,
      "theoretical_loss": 3.780091903507206,
      "tokens_seen": 698841088
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039810431293881644,
      "loss": 3.0582,
      "theoretical_loss": 3.7800559010445567,
      "tokens_seen": 698906624
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003980942828485457,
      "loss": 2.7882,
      "theoretical_loss": 3.7800199029028345,
      "tokens_seen": 698972160
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003980842527582748,
      "loss": 2.9789,
      "theoretical_loss": 3.779983909081115,
      "tokens_seen": 699037696
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039807422266800404,
      "loss": 2.9775,
      "theoretical_loss": 3.7799479195784746,
      "tokens_seen": 699103232
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039806419257773317,
      "loss": 3.027,
      "theoretical_loss": 3.779911934393991,
      "tokens_seen": 699168768
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003980541624874624,
      "loss": 3.0297,
      "theoretical_loss": 3.7798759535267417,
      "tokens_seen": 699234304
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003980441323971916,
      "loss": 2.949,
      "theoretical_loss": 3.779839976975803,
      "tokens_seen": 699299840
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039803410230692077,
      "loss": 2.8488,
      "theoretical_loss": 3.779804004740253,
      "tokens_seen": 699365376
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039802407221664995,
      "loss": 2.9867,
      "theoretical_loss": 3.7797680368191706,
      "tokens_seen": 699430912
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039801404212637913,
      "loss": 2.9055,
      "theoretical_loss": 3.7797320732116337,
      "tokens_seen": 699496448
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003980040120361083,
      "loss": 3.1272,
      "theoretical_loss": 3.7796961139167204,
      "tokens_seen": 699561984
    },
    {
      "epoch": 2.03,
      "objective/train/docs_used": 833268,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7306745052337646,
      "objective/train/theoretical_loss": 3.7796781358862104,
      "objective/train/tokens_used": 720054752,
      "theoretical_loss": 3.7796781358862104,
      "tokens_seen": 699594752
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039799398194583754,
      "loss": 2.9271,
      "theoretical_loss": 3.7796601589335106,
      "tokens_seen": 699627520
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039798395185556667,
      "loss": 2.8494,
      "theoretical_loss": 3.7796242082610823,
      "tokens_seen": 699693056
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003979739217652959,
      "loss": 2.9823,
      "theoretical_loss": 3.7795882618985157,
      "tokens_seen": 699758592
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039796389167502503,
      "loss": 2.9894,
      "theoretical_loss": 3.779552319844891,
      "tokens_seen": 699824128
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039795386158475427,
      "loss": 2.8972,
      "theoretical_loss": 3.779516382099287,
      "tokens_seen": 699889664
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039794383149448345,
      "loss": 2.8741,
      "theoretical_loss": 3.7794804486607845,
      "tokens_seen": 699955200
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039793380140421263,
      "loss": 2.907,
      "theoretical_loss": 3.779444519528465,
      "tokens_seen": 700020736
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003979237713139418,
      "loss": 2.9134,
      "theoretical_loss": 3.779408594701408,
      "tokens_seen": 700086272
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039791374122367105,
      "loss": 2.7774,
      "theoretical_loss": 3.779372674178696,
      "tokens_seen": 700151808
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003979037111334002,
      "loss": 3.0367,
      "theoretical_loss": 3.7793367579594097,
      "tokens_seen": 700217344
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003978936810431294,
      "loss": 2.8712,
      "theoretical_loss": 3.7793008460426307,
      "tokens_seen": 700282880
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003978836509528586,
      "loss": 2.9051,
      "theoretical_loss": 3.779264938427441,
      "tokens_seen": 700348416
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039787362086258777,
      "loss": 3.1493,
      "theoretical_loss": 3.7792290351129236,
      "tokens_seen": 700413952
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.000397863590772317,
      "loss": 2.8041,
      "theoretical_loss": 3.7791931360981605,
      "tokens_seen": 700479488
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039785356068204613,
      "loss": 3.0465,
      "theoretical_loss": 3.7791572413822343,
      "tokens_seen": 700545024
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039784353059177537,
      "loss": 2.8724,
      "theoretical_loss": 3.7791213509642287,
      "tokens_seen": 700610560
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003978335005015045,
      "loss": 2.8372,
      "theoretical_loss": 3.7790854648432273,
      "tokens_seen": 700676096
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039782347041123373,
      "loss": 3.0213,
      "theoretical_loss": 3.779049583018313,
      "tokens_seen": 700741632
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003978134403209629,
      "loss": 3.0222,
      "theoretical_loss": 3.7790137054885697,
      "tokens_seen": 700807168
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003978034102306921,
      "loss": 2.9475,
      "theoretical_loss": 3.7789778322530827,
      "tokens_seen": 700872704
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003977933801404213,
      "loss": 2.7373,
      "theoretical_loss": 3.7789419633109365,
      "tokens_seen": 700938240
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039778335005015046,
      "loss": 2.9236,
      "theoretical_loss": 3.7789060986612144,
      "tokens_seen": 701003776
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039777331995987964,
      "loss": 2.8003,
      "theoretical_loss": 3.7788702383030035,
      "tokens_seen": 701069312
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039776328986960887,
      "loss": 2.7636,
      "theoretical_loss": 3.7788343822353876,
      "tokens_seen": 701134848
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.000397753259779338,
      "loss": 2.7646,
      "theoretical_loss": 3.778798530457453,
      "tokens_seen": 701200384
    },
    {
      "epoch": 2.03,
      "objective/train/docs_used": 838443,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.2617812156677246,
      "objective/train/theoretical_loss": 3.7787806061768308,
      "objective/train/tokens_used": 721693152,
      "theoretical_loss": 3.7787806061768308,
      "tokens_seen": 701233152
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039774322968906723,
      "loss": 2.84,
      "theoretical_loss": 3.7787626829682863,
      "tokens_seen": 701265920
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003977331995987964,
      "loss": 2.831,
      "theoretical_loss": 3.778726839766972,
      "tokens_seen": 701331456
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003977231695085256,
      "loss": 2.8498,
      "theoretical_loss": 3.7786910008525987,
      "tokens_seen": 701396992
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003977131394182548,
      "loss": 2.8381,
      "theoretical_loss": 3.778655166224252,
      "tokens_seen": 701462528
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039770310932798396,
      "loss": 2.9054,
      "theoretical_loss": 3.778619335881019,
      "tokens_seen": 701528064
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039769307923771314,
      "loss": 3.1357,
      "theoretical_loss": 3.7785835098219867,
      "tokens_seen": 701593600
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003976830491474424,
      "loss": 2.7488,
      "theoretical_loss": 3.7785476880462436,
      "tokens_seen": 701659136
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003976730190571715,
      "loss": 3.1839,
      "theoretical_loss": 3.778511870552877,
      "tokens_seen": 701724672
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039766298896690074,
      "loss": 3.026,
      "theoretical_loss": 3.778476057340975,
      "tokens_seen": 701790208
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039765295887662987,
      "loss": 2.8715,
      "theoretical_loss": 3.778440248409627,
      "tokens_seen": 701855744
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003976429287863591,
      "loss": 2.9057,
      "theoretical_loss": 3.7784044437579207,
      "tokens_seen": 701921280
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003976328986960883,
      "loss": 2.9479,
      "theoretical_loss": 3.7783686433849457,
      "tokens_seen": 701986816
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039762286860581746,
      "loss": 2.8235,
      "theoretical_loss": 3.778332847289791,
      "tokens_seen": 702052352
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039761283851554664,
      "loss": 2.9134,
      "theoretical_loss": 3.7782970554715467,
      "tokens_seen": 702117888
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003976028084252759,
      "loss": 2.9239,
      "theoretical_loss": 3.7782612679293015,
      "tokens_seen": 702183424
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.000397592778335005,
      "loss": 2.7292,
      "theoretical_loss": 3.778225484662147,
      "tokens_seen": 702248960
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039758274824473424,
      "loss": 2.8728,
      "theoretical_loss": 3.7781897056691722,
      "tokens_seen": 702314496
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039757271815446337,
      "loss": 3.129,
      "theoretical_loss": 3.778153930949469,
      "tokens_seen": 702380032
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003975626880641926,
      "loss": 2.8147,
      "theoretical_loss": 3.7781181605021277,
      "tokens_seen": 702445568
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003975526579739218,
      "loss": 3.1122,
      "theoretical_loss": 3.7780823943262396,
      "tokens_seen": 702511104
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039754262788365097,
      "loss": 2.9351,
      "theoretical_loss": 3.7780466324208968,
      "tokens_seen": 702576640
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039753259779338015,
      "loss": 2.5674,
      "theoretical_loss": 3.77801087478519,
      "tokens_seen": 702642176
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039752256770310933,
      "loss": 3.0548,
      "theoretical_loss": 3.7779751214182125,
      "tokens_seen": 702707712
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003975125376128385,
      "loss": 2.8329,
      "theoretical_loss": 3.777939372319056,
      "tokens_seen": 702773248
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039750250752256774,
      "loss": 2.9937,
      "theoretical_loss": 3.777903627486813,
      "tokens_seen": 702838784
    },
    {
      "epoch": 2.03,
      "objective/train/docs_used": 839055,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7638256549835205,
      "objective/train/theoretical_loss": 3.777885756670501,
      "objective/train/tokens_used": 723331552,
      "theoretical_loss": 3.777885756670501,
      "tokens_seen": 702871552
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039749247743229687,
      "loss": 2.9961,
      "theoretical_loss": 3.7778678869205766,
      "tokens_seen": 702904320
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003974824473420261,
      "loss": 2.8954,
      "theoretical_loss": 3.7778321506194406,
      "tokens_seen": 702969856
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039747241725175523,
      "loss": 2.8431,
      "theoretical_loss": 3.7777964185824975,
      "tokens_seen": 703035392
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039746238716148447,
      "loss": 2.9681,
      "theoretical_loss": 3.7777606908088415,
      "tokens_seen": 703100928
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039745235707121365,
      "loss": 3.0087,
      "theoretical_loss": 3.7777249672975666,
      "tokens_seen": 703166464
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039744232698094283,
      "loss": 3.2902,
      "theoretical_loss": 3.7776892480477673,
      "tokens_seen": 703232000
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.000397432296890672,
      "loss": 3.011,
      "theoretical_loss": 3.7776535330585377,
      "tokens_seen": 703297536
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039742226680040125,
      "loss": 2.8944,
      "theoretical_loss": 3.777617822328973,
      "tokens_seen": 703363072
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003974122367101304,
      "loss": 2.918,
      "theoretical_loss": 3.7775821158581686,
      "tokens_seen": 703428608
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003974022066198596,
      "loss": 2.7816,
      "theoretical_loss": 3.777546413645219,
      "tokens_seen": 703494144
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039739217652958874,
      "loss": 3.0054,
      "theoretical_loss": 3.777510715689221,
      "tokens_seen": 703559680
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039738214643931797,
      "loss": 2.9605,
      "theoretical_loss": 3.77747502198927,
      "tokens_seen": 703625216
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039737211634904715,
      "loss": 2.8161,
      "theoretical_loss": 3.7774393325444615,
      "tokens_seen": 703690752
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039736208625877633,
      "loss": 2.9456,
      "theoretical_loss": 3.7774036473538937,
      "tokens_seen": 703756288
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003973520561685055,
      "loss": 3.0258,
      "theoretical_loss": 3.777367966416662,
      "tokens_seen": 703821824
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003973420260782347,
      "loss": 2.979,
      "theoretical_loss": 3.777332289731864,
      "tokens_seen": 703887360
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003973319959879639,
      "loss": 2.9257,
      "theoretical_loss": 3.777296617298597,
      "tokens_seen": 703952896
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003973219658976931,
      "loss": 2.7951,
      "theoretical_loss": 3.7772609491159588,
      "tokens_seen": 704018432
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039731193580742224,
      "loss": 2.9469,
      "theoretical_loss": 3.7772252851830466,
      "tokens_seen": 704083968
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003973019057171515,
      "loss": 3.1696,
      "theoretical_loss": 3.777189625498959,
      "tokens_seen": 704149504
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003972918756268806,
      "loss": 2.9136,
      "theoretical_loss": 3.7771539700627956,
      "tokens_seen": 704215040
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039728184553660984,
      "loss": 2.8302,
      "theoretical_loss": 3.7771183188736535,
      "tokens_seen": 704280576
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.000397271815446339,
      "loss": 2.9683,
      "theoretical_loss": 3.777082671930632,
      "tokens_seen": 704346112
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003972617853560682,
      "loss": 2.8791,
      "theoretical_loss": 3.77704702923283,
      "tokens_seen": 704411648
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003972517552657974,
      "loss": 2.969,
      "theoretical_loss": 3.7770113907793483,
      "tokens_seen": 704477184
    },
    {
      "epoch": 2.03,
      "objective/train/docs_used": 840230,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.1280670166015625,
      "objective/train/theoretical_loss": 3.776993573143946,
      "objective/train/tokens_used": 724969952,
      "theoretical_loss": 3.776993573143946,
      "tokens_seen": 704509952
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003972417251755266,
      "loss": 2.7157,
      "theoretical_loss": 3.776975756569286,
      "tokens_seen": 704542720
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039723169508525574,
      "loss": 2.8356,
      "theoretical_loss": 3.776940126601743,
      "tokens_seen": 704608256
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.000397221664994985,
      "loss": 2.8503,
      "theoretical_loss": 3.7769045008758204,
      "tokens_seen": 704673792
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003972116349047141,
      "loss": 2.9024,
      "theoretical_loss": 3.7768688793906184,
      "tokens_seen": 704739328
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039720160481444334,
      "loss": 3.0062,
      "theoretical_loss": 3.776833262145237,
      "tokens_seen": 704804864
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003971915747241725,
      "loss": 3.1082,
      "theoretical_loss": 3.776797649138779,
      "tokens_seen": 704870400
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003971815446339017,
      "loss": 2.9018,
      "theoretical_loss": 3.776762040370345,
      "tokens_seen": 704935936
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003971715145436309,
      "loss": 2.7978,
      "theoretical_loss": 3.776726435839037,
      "tokens_seen": 705001472
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039716148445336007,
      "loss": 2.8779,
      "theoretical_loss": 3.776690835543956,
      "tokens_seen": 705067008
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039715145436308925,
      "loss": 2.9387,
      "theoretical_loss": 3.7766552394842057,
      "tokens_seen": 705132544
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003971414242728185,
      "loss": 2.6552,
      "theoretical_loss": 3.7766196476588885,
      "tokens_seen": 705198080
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039713139418254766,
      "loss": 3.0808,
      "theoretical_loss": 3.7765840600671066,
      "tokens_seen": 705263616
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039712136409227684,
      "loss": 2.9143,
      "theoretical_loss": 3.7765484767079633,
      "tokens_seen": 705329152
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003971113340020061,
      "loss": 3.0465,
      "theoretical_loss": 3.776512897580562,
      "tokens_seen": 705394688
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003971013039117352,
      "loss": 2.8223,
      "theoretical_loss": 3.7764773226840065,
      "tokens_seen": 705460224
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039709127382146444,
      "loss": 2.892,
      "theoretical_loss": 3.7764417520174005,
      "tokens_seen": 705525760
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039708124373119357,
      "loss": 2.7686,
      "theoretical_loss": 3.7764061855798485,
      "tokens_seen": 705591296
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003970712136409228,
      "loss": 2.8557,
      "theoretical_loss": 3.776370623370455,
      "tokens_seen": 705656832
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.000397061183550652,
      "loss": 2.9014,
      "theoretical_loss": 3.7763350653883236,
      "tokens_seen": 705722368
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039705115346038117,
      "loss": 2.8048,
      "theoretical_loss": 3.776299511632561,
      "tokens_seen": 705787904
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039704112337011035,
      "loss": 2.9953,
      "theoretical_loss": 3.7762639621022718,
      "tokens_seen": 705853440
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039703109327983953,
      "loss": 3.0841,
      "theoretical_loss": 3.7762284167965614,
      "tokens_seen": 705918976
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003970210631895687,
      "loss": 2.9048,
      "theoretical_loss": 3.7761928757145355,
      "tokens_seen": 705984512
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039701103309929794,
      "loss": 2.7224,
      "theoretical_loss": 3.7761573388553007,
      "tokens_seen": 706050048
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039700100300902707,
      "loss": 2.8874,
      "theoretical_loss": 3.776121806217963,
      "tokens_seen": 706115584
    },
    {
      "epoch": 2.03,
      "objective/train/docs_used": 840818,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.0678188800811768,
      "objective/train/theoretical_loss": 3.7761040414822267,
      "objective/train/tokens_used": 726608352,
      "theoretical_loss": 3.7761040414822267,
      "tokens_seen": 706148352
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003969909729187563,
      "loss": 2.97,
      "theoretical_loss": 3.776086277801629,
      "tokens_seen": 706181120
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039698094282848543,
      "loss": 2.8323,
      "theoretical_loss": 3.776050753605406,
      "tokens_seen": 706246656
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039697091273821467,
      "loss": 2.9463,
      "theoretical_loss": 3.776015233628401,
      "tokens_seen": 706312192
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039696088264794385,
      "loss": 3.0658,
      "theoretical_loss": 3.7759797178697214,
      "tokens_seen": 706377728
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039695085255767303,
      "loss": 3.0671,
      "theoretical_loss": 3.775944206328475,
      "tokens_seen": 706443264
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003969408224674022,
      "loss": 2.9199,
      "theoretical_loss": 3.77590869900377,
      "tokens_seen": 706508800
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039693079237713145,
      "loss": 2.8948,
      "theoretical_loss": 3.7758731958947145,
      "tokens_seen": 706574336
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003969207622868606,
      "loss": 2.8665,
      "theoretical_loss": 3.7758376970004166,
      "tokens_seen": 706639872
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003969107321965898,
      "loss": 2.8945,
      "theoretical_loss": 3.775802202319986,
      "tokens_seen": 706705408
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039690070210631894,
      "loss": 3.097,
      "theoretical_loss": 3.7757667118525315,
      "tokens_seen": 706770944
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039689067201604817,
      "loss": 2.9045,
      "theoretical_loss": 3.7757312255971622,
      "tokens_seen": 706836480
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039688064192577735,
      "loss": 2.7058,
      "theoretical_loss": 3.775695743552988,
      "tokens_seen": 706902016
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039687061183550653,
      "loss": 3.0764,
      "theoretical_loss": 3.775660265719118,
      "tokens_seen": 706967552
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003968605817452357,
      "loss": 2.9457,
      "theoretical_loss": 3.7756247920946633,
      "tokens_seen": 707033088
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003968505516549649,
      "loss": 2.9319,
      "theoretical_loss": 3.7755893226787345,
      "tokens_seen": 707098624
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003968405215646941,
      "loss": 2.9837,
      "theoretical_loss": 3.7755538574704417,
      "tokens_seen": 707164160
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003968304914744233,
      "loss": 3.0032,
      "theoretical_loss": 3.7755183964688968,
      "tokens_seen": 707229696
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039682046138415244,
      "loss": 2.7542,
      "theoretical_loss": 3.7754829396732097,
      "tokens_seen": 707295232
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003968104312938817,
      "loss": 2.9074,
      "theoretical_loss": 3.7754474870824923,
      "tokens_seen": 707360768
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003968004012036108,
      "loss": 2.716,
      "theoretical_loss": 3.7754120386958574,
      "tokens_seen": 707426304
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039679037111334004,
      "loss": 2.7301,
      "theoretical_loss": 3.775376594512416,
      "tokens_seen": 707491840
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003967803410230692,
      "loss": 3.0049,
      "theoretical_loss": 3.7753411545312803,
      "tokens_seen": 707557376
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003967703109327984,
      "loss": 3.0008,
      "theoretical_loss": 3.7753057187515644,
      "tokens_seen": 707622912
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003967602808425276,
      "loss": 3.1812,
      "theoretical_loss": 3.7752702871723796,
      "tokens_seen": 707688448
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003967502507522568,
      "loss": 2.9954,
      "theoretical_loss": 3.77523485979284,
      "tokens_seen": 707753984
    },
    {
      "epoch": 2.03,
      "objective/train/docs_used": 842171,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9438395500183105,
      "objective/train/theoretical_loss": 3.77521714767766,
      "objective/train/tokens_used": 728246752,
      "theoretical_loss": 3.77521714767766,
      "tokens_seen": 707786752
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039674022066198594,
      "loss": 2.8499,
      "theoretical_loss": 3.7751994366120583,
      "tokens_seen": 707819520
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003967301905717152,
      "loss": 2.9356,
      "theoretical_loss": 3.7751640176291485,
      "tokens_seen": 707885056
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003967201604814443,
      "loss": 3.1708,
      "theoretical_loss": 3.7751286028432247,
      "tokens_seen": 707950592
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039671013039117354,
      "loss": 2.9933,
      "theoretical_loss": 3.7750931922534017,
      "tokens_seen": 708016128
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003967001003009027,
      "loss": 2.985,
      "theoretical_loss": 3.7750577858587926,
      "tokens_seen": 708081664
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003966900702106319,
      "loss": 2.7421,
      "theoretical_loss": 3.775022383658513,
      "tokens_seen": 708147200
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003966800401203611,
      "loss": 3.0504,
      "theoretical_loss": 3.774986985651678,
      "tokens_seen": 708212736
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039667001003009027,
      "loss": 2.8545,
      "theoretical_loss": 3.774951591837402,
      "tokens_seen": 708278272
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039665997993981945,
      "loss": 2.7481,
      "theoretical_loss": 3.7749162022148024,
      "tokens_seen": 708343808
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003966499498495487,
      "loss": 2.7234,
      "theoretical_loss": 3.774880816782993,
      "tokens_seen": 708409344
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003966399197592778,
      "loss": 2.9445,
      "theoretical_loss": 3.774845435541091,
      "tokens_seen": 708474880
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039662988966900704,
      "loss": 2.8104,
      "theoretical_loss": 3.7748100584882134,
      "tokens_seen": 708540416
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039661985957873617,
      "loss": 2.9856,
      "theoretical_loss": 3.774774685623475,
      "tokens_seen": 708605952
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003966098294884654,
      "loss": 2.8665,
      "theoretical_loss": 3.774739316945994,
      "tokens_seen": 708671488
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003965997993981946,
      "loss": 2.8147,
      "theoretical_loss": 3.774703952454888,
      "tokens_seen": 708737024
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039658976930792377,
      "loss": 3.0254,
      "theoretical_loss": 3.7746685921492737,
      "tokens_seen": 708802560
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.00039657973921765295,
      "loss": 3.0298,
      "theoretical_loss": 3.7746332360282677,
      "tokens_seen": 708868096
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003965697091273822,
      "loss": 3.0365,
      "theoretical_loss": 3.77459788409099,
      "tokens_seen": 708933632
    },
    {
      "epoch": 2.03,
      "learning_rate": 0.0003965596790371113,
      "loss": 3.0917,
      "theoretical_loss": 3.7745625363365582,
      "tokens_seen": 708999168
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039654964894684055,
      "loss": 3.0597,
      "theoretical_loss": 3.7745271927640904,
      "tokens_seen": 709064704
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003965396188565697,
      "loss": 2.8429,
      "theoretical_loss": 3.7744918533727057,
      "tokens_seen": 709130240
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003965295887662989,
      "loss": 2.9297,
      "theoretical_loss": 3.7744565181615233,
      "tokens_seen": 709195776
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003965195586760281,
      "loss": 2.9209,
      "theoretical_loss": 3.7744211871296613,
      "tokens_seen": 709261312
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039650952858575727,
      "loss": 2.7967,
      "theoretical_loss": 3.774385860276241,
      "tokens_seen": 709326848
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039649949849548645,
      "loss": 3.0973,
      "theoretical_loss": 3.7743505376003816,
      "tokens_seen": 709392384
    },
    {
      "epoch": 2.04,
      "objective/train/docs_used": 842767,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9267566204071045,
      "objective/train/theoretical_loss": 3.774332877828762,
      "objective/train/tokens_used": 729885152,
      "theoretical_loss": 3.774332877828762,
      "tokens_seen": 709425152
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039648946840521563,
      "loss": 2.8282,
      "theoretical_loss": 3.7743152191012026,
      "tokens_seen": 709457920
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003964794383149448,
      "loss": 2.88,
      "theoretical_loss": 3.7742799047778255,
      "tokens_seen": 709523456
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039646940822467405,
      "loss": 2.8465,
      "theoretical_loss": 3.77424459462937,
      "tokens_seen": 709588992
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003964593781344032,
      "loss": 2.949,
      "theoretical_loss": 3.774209288654957,
      "tokens_seen": 709654528
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003964493480441324,
      "loss": 2.9575,
      "theoretical_loss": 3.7741739868537088,
      "tokens_seen": 709720064
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003964393179538616,
      "loss": 2.8422,
      "theoretical_loss": 3.7741386892247455,
      "tokens_seen": 709785600
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003964292878635908,
      "loss": 3.0997,
      "theoretical_loss": 3.77410339576719,
      "tokens_seen": 709851136
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039641925777331996,
      "loss": 2.8329,
      "theoretical_loss": 3.774068106480163,
      "tokens_seen": 709916672
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039640922768304914,
      "loss": 2.9717,
      "theoretical_loss": 3.774032821362788,
      "tokens_seen": 709982208
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003963991975927783,
      "loss": 2.9319,
      "theoretical_loss": 3.773997540414187,
      "tokens_seen": 710047744
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039638916750250755,
      "loss": 2.8197,
      "theoretical_loss": 3.773962263633482,
      "tokens_seen": 710113280
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039637913741223673,
      "loss": 2.7053,
      "theoretical_loss": 3.7739269910197972,
      "tokens_seen": 710178816
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003963691073219659,
      "loss": 2.9991,
      "theoretical_loss": 3.7738917225722557,
      "tokens_seen": 710244352
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003963590772316951,
      "loss": 2.995,
      "theoretical_loss": 3.7738564582899805,
      "tokens_seen": 710309888
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003963490471414243,
      "loss": 3.0643,
      "theoretical_loss": 3.7738211981720964,
      "tokens_seen": 710375424
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003963390170511535,
      "loss": 3.0085,
      "theoretical_loss": 3.7737859422177262,
      "tokens_seen": 710440960
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039632898696088264,
      "loss": 2.9135,
      "theoretical_loss": 3.773750690425995,
      "tokens_seen": 710506496
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003963189568706119,
      "loss": 2.8237,
      "theoretical_loss": 3.7737154427960276,
      "tokens_seen": 710572032
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.000396308926780341,
      "loss": 2.9719,
      "theoretical_loss": 3.7736801993269484,
      "tokens_seen": 710637568
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039629889669007024,
      "loss": 2.8327,
      "theoretical_loss": 3.7736449600178834,
      "tokens_seen": 710703104
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003962888665997994,
      "loss": 2.8824,
      "theoretical_loss": 3.773609724867957,
      "tokens_seen": 710768640
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003962788365095286,
      "loss": 2.8354,
      "theoretical_loss": 3.7735744938762954,
      "tokens_seen": 710834176
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003962688064192578,
      "loss": 3.2813,
      "theoretical_loss": 3.7735392670420254,
      "tokens_seen": 710899712
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.000396258776328987,
      "loss": 2.8482,
      "theoretical_loss": 3.7735040443642713,
      "tokens_seen": 710965248
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039624874623871614,
      "loss": 2.9067,
      "theoretical_loss": 3.773468825842161,
      "tokens_seen": 711030784
    },
    {
      "epoch": 2.04,
      "objective/train/docs_used": 844300,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.3437087535858154,
      "objective/train/theoretical_loss": 3.7734512181391993,
      "objective/train/tokens_used": 731523552,
      "theoretical_loss": 3.7734512181391993,
      "tokens_seen": 711063552
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003962387161484454,
      "loss": 3.113,
      "theoretical_loss": 3.773433611474821,
      "tokens_seen": 711096320
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003962286860581745,
      "loss": 2.9326,
      "theoretical_loss": 3.773398401261378,
      "tokens_seen": 711161856
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039621865596790374,
      "loss": 3.1375,
      "theoretical_loss": 3.7733631952009596,
      "tokens_seen": 711227392
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003962086258776329,
      "loss": 2.7491,
      "theoretical_loss": 3.773327993292693,
      "tokens_seen": 711292928
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003961985957873621,
      "loss": 2.9599,
      "theoretical_loss": 3.7732927955357063,
      "tokens_seen": 711358464
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003961885656970913,
      "loss": 2.7884,
      "theoretical_loss": 3.773257601929128,
      "tokens_seen": 711424000
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039617853560682047,
      "loss": 2.8457,
      "theoretical_loss": 3.7732224124720855,
      "tokens_seen": 711489536
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039616850551654965,
      "loss": 2.9076,
      "theoretical_loss": 3.7731872271637075,
      "tokens_seen": 711555072
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003961584754262789,
      "loss": 2.8034,
      "theoretical_loss": 3.7731520460031236,
      "tokens_seen": 711620608
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.000396148445336008,
      "loss": 2.7777,
      "theoretical_loss": 3.7731168689894625,
      "tokens_seen": 711686144
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039613841524573724,
      "loss": 2.9276,
      "theoretical_loss": 3.7730816961218534,
      "tokens_seen": 711751680
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039612838515546637,
      "loss": 2.7608,
      "theoretical_loss": 3.7730465273994263,
      "tokens_seen": 711817216
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003961183550651956,
      "loss": 2.8722,
      "theoretical_loss": 3.773011362821311,
      "tokens_seen": 711882752
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003961083249749248,
      "loss": 2.7365,
      "theoretical_loss": 3.772976202386637,
      "tokens_seen": 711948288
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039609829488465397,
      "loss": 2.8973,
      "theoretical_loss": 3.772941046094536,
      "tokens_seen": 712013824
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039608826479438315,
      "loss": 2.7796,
      "theoretical_loss": 3.772905893944138,
      "tokens_seen": 712079360
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003960782347041124,
      "loss": 2.7532,
      "theoretical_loss": 3.7728707459345734,
      "tokens_seen": 712144896
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003960682046138415,
      "loss": 2.973,
      "theoretical_loss": 3.7728356020649745,
      "tokens_seen": 712210432
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039605817452357075,
      "loss": 3.0538,
      "theoretical_loss": 3.772800462334472,
      "tokens_seen": 712275968
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003960481444332999,
      "loss": 2.759,
      "theoretical_loss": 3.7727653267421983,
      "tokens_seen": 712341504
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003960381143430291,
      "loss": 2.8805,
      "theoretical_loss": 3.772730195287285,
      "tokens_seen": 712407040
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003960280842527583,
      "loss": 2.8023,
      "theoretical_loss": 3.772695067968865,
      "tokens_seen": 712472576
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039601805416248747,
      "loss": 2.9331,
      "theoretical_loss": 3.7726599447860694,
      "tokens_seen": 712538112
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039600802407221665,
      "loss": 2.9116,
      "theoretical_loss": 3.772624825738032,
      "tokens_seen": 712603648
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039599799398194583,
      "loss": 3.1991,
      "theoretical_loss": 3.772589710823886,
      "tokens_seen": 712669184
    },
    {
      "epoch": 2.04,
      "objective/train/docs_used": 844668,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.137606143951416,
      "objective/train/theoretical_loss": 3.772572154916751,
      "objective/train/tokens_used": 733161952,
      "theoretical_loss": 3.772572154916751,
      "tokens_seen": 712701952
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.000395987963891675,
      "loss": 3.0682,
      "theoretical_loss": 3.772554600042764,
      "tokens_seen": 712734720
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039597793380140425,
      "loss": 2.923,
      "theoretical_loss": 3.7725194933938004,
      "tokens_seen": 712800256
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003959679037111334,
      "loss": 3.009,
      "theoretical_loss": 3.7724843908761283,
      "tokens_seen": 712865792
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003959578736208626,
      "loss": 2.9241,
      "theoretical_loss": 3.772449292488883,
      "tokens_seen": 712931328
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003959478435305918,
      "loss": 2.7148,
      "theoretical_loss": 3.772414198231197,
      "tokens_seen": 712996864
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.000395937813440321,
      "loss": 2.6896,
      "theoretical_loss": 3.7723791081022062,
      "tokens_seen": 713062400
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039592778335005016,
      "loss": 2.7487,
      "theoretical_loss": 3.7723440221010454,
      "tokens_seen": 713127936
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039591775325977934,
      "loss": 2.9316,
      "theoretical_loss": 3.7723089402268495,
      "tokens_seen": 713193472
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003959077231695085,
      "loss": 2.8535,
      "theoretical_loss": 3.772273862478754,
      "tokens_seen": 713259008
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039589769307923775,
      "loss": 2.7874,
      "theoretical_loss": 3.7722387888558946,
      "tokens_seen": 713324544
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003958876629889669,
      "loss": 2.7757,
      "theoretical_loss": 3.7722037193574067,
      "tokens_seen": 713390080
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003958776328986961,
      "loss": 2.6952,
      "theoretical_loss": 3.7721686539824275,
      "tokens_seen": 713455616
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039586760280842524,
      "loss": 2.8004,
      "theoretical_loss": 3.7721335927300927,
      "tokens_seen": 713521152
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003958575727181545,
      "loss": 2.8383,
      "theoretical_loss": 3.772098535599539,
      "tokens_seen": 713586688
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039584754262788366,
      "loss": 3.1454,
      "theoretical_loss": 3.772063482589904,
      "tokens_seen": 713652224
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039583751253761284,
      "loss": 2.7299,
      "theoretical_loss": 3.7720284337003243,
      "tokens_seen": 713717760
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.000395827482447342,
      "loss": 2.9994,
      "theoretical_loss": 3.7719933889299377,
      "tokens_seen": 713783296
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003958174523570712,
      "loss": 2.7725,
      "theoretical_loss": 3.771958348277882,
      "tokens_seen": 713848832
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003958074222668004,
      "loss": 2.7891,
      "theoretical_loss": 3.7719233117432944,
      "tokens_seen": 713914368
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003957973921765296,
      "loss": 2.8777,
      "theoretical_loss": 3.771888279325314,
      "tokens_seen": 713979904
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039578736208625875,
      "loss": 2.9531,
      "theoretical_loss": 3.771853251023079,
      "tokens_seen": 714045440
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.000395777331995988,
      "loss": 2.9526,
      "theoretical_loss": 3.7718182268357285,
      "tokens_seen": 714110976
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039576730190571716,
      "loss": 2.879,
      "theoretical_loss": 3.7717832067624015,
      "tokens_seen": 714176512
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039575727181544634,
      "loss": 2.9837,
      "theoretical_loss": 3.771748190802237,
      "tokens_seen": 714242048
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003957472417251755,
      "loss": 2.878,
      "theoretical_loss": 3.771713178954375,
      "tokens_seen": 714307584
    },
    {
      "epoch": 2.04,
      "objective/train/docs_used": 846114,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8042542934417725,
      "objective/train/theoretical_loss": 3.771695674572288,
      "objective/train/tokens_used": 734800352,
      "theoretical_loss": 3.771695674572288,
      "tokens_seen": 714340352
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003957372116349047,
      "loss": 2.8705,
      "theoretical_loss": 3.7716781712179546,
      "tokens_seen": 714373120
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003957271815446339,
      "loss": 2.8951,
      "theoretical_loss": 3.7716431675921167,
      "tokens_seen": 714438656
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003957171514543631,
      "loss": 3.003,
      "theoretical_loss": 3.7716081680760007,
      "tokens_seen": 714504192
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039570712136409225,
      "loss": 2.734,
      "theoretical_loss": 3.7715731726687487,
      "tokens_seen": 714569728
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003956970912738215,
      "loss": 2.9267,
      "theoretical_loss": 3.7715381813694995,
      "tokens_seen": 714635264
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003956870611835506,
      "loss": 2.9854,
      "theoretical_loss": 3.771503194177396,
      "tokens_seen": 714700800
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039567703109327985,
      "loss": 2.7856,
      "theoretical_loss": 3.771468211091579,
      "tokens_seen": 714766336
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039566700100300903,
      "loss": 2.7446,
      "theoretical_loss": 3.77143323211119,
      "tokens_seen": 714831872
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003956569709127382,
      "loss": 2.6806,
      "theoretical_loss": 3.7713982572353704,
      "tokens_seen": 714897408
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003956469408224674,
      "loss": 2.8043,
      "theoretical_loss": 3.771363286463263,
      "tokens_seen": 714962944
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039563691073219657,
      "loss": 3.0227,
      "theoretical_loss": 3.77132831979401,
      "tokens_seen": 715028480
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003956268806419258,
      "loss": 2.8135,
      "theoretical_loss": 3.7712933572267544,
      "tokens_seen": 715094016
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.000395616850551655,
      "loss": 2.5729,
      "theoretical_loss": 3.7712583987606387,
      "tokens_seen": 715159552
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039560682046138417,
      "loss": 2.8661,
      "theoretical_loss": 3.7712234443948063,
      "tokens_seen": 715225088
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039559679037111335,
      "loss": 2.8424,
      "theoretical_loss": 3.7711884941284004,
      "tokens_seen": 715290624
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003955867602808426,
      "loss": 2.8453,
      "theoretical_loss": 3.771153547960565,
      "tokens_seen": 715356160
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003955767301905717,
      "loss": 3.137,
      "theoretical_loss": 3.771118605890444,
      "tokens_seen": 715421696
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039556670010030095,
      "loss": 3.0339,
      "theoretical_loss": 3.771083667917181,
      "tokens_seen": 715487232
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003955566700100301,
      "loss": 2.7859,
      "theoretical_loss": 3.7710487340399217,
      "tokens_seen": 715552768
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003955466399197593,
      "loss": 2.7065,
      "theoretical_loss": 3.7710138042578096,
      "tokens_seen": 715618304
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003955366098294885,
      "loss": 2.9122,
      "theoretical_loss": 3.7709788785699905,
      "tokens_seen": 715683840
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039552657973921767,
      "loss": 2.8986,
      "theoretical_loss": 3.770943956975609,
      "tokens_seen": 715749376
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039551654964894685,
      "loss": 2.7145,
      "theoretical_loss": 3.770909039473811,
      "tokens_seen": 715814912
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039550651955867603,
      "loss": 2.7466,
      "theoretical_loss": 3.770874126063742,
      "tokens_seen": 715880448
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003954964894684052,
      "loss": 2.7592,
      "theoretical_loss": 3.770839216744548,
      "tokens_seen": 715945984
    },
    {
      "epoch": 2.04,
      "objective/train/docs_used": 846880,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.995128631591797,
      "objective/train/theoretical_loss": 3.7708217636187626,
      "objective/train/tokens_used": 736438752,
      "theoretical_loss": 3.7708217636187626,
      "tokens_seen": 715978752
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039548645937813445,
      "loss": 2.6837,
      "theoretical_loss": 3.770804311515376,
      "tokens_seen": 716011520
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003954764292878636,
      "loss": 3.0525,
      "theoretical_loss": 3.7707694103753715,
      "tokens_seen": 716077056
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003954663991975928,
      "loss": 2.9757,
      "theoretical_loss": 3.770734513323682,
      "tokens_seen": 716142592
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.000395456369107322,
      "loss": 2.7955,
      "theoretical_loss": 3.770699620359454,
      "tokens_seen": 716208128
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003954463390170512,
      "loss": 2.8557,
      "theoretical_loss": 3.770664731481835,
      "tokens_seen": 716273664
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039543630892678036,
      "loss": 2.8646,
      "theoretical_loss": 3.7706298466899724,
      "tokens_seen": 716339200
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039542627883650954,
      "loss": 2.9426,
      "theoretical_loss": 3.7705949659830145,
      "tokens_seen": 716404736
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003954162487462387,
      "loss": 2.7996,
      "theoretical_loss": 3.7705600893601083,
      "tokens_seen": 716470272
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039540621865596795,
      "loss": 3.0203,
      "theoretical_loss": 3.7705252168204035,
      "tokens_seen": 716535808
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003953961885656971,
      "loss": 2.9499,
      "theoretical_loss": 3.770490348363048,
      "tokens_seen": 716601344
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003953861584754263,
      "loss": 3.0872,
      "theoretical_loss": 3.77045548398719,
      "tokens_seen": 716666880
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039537612838515544,
      "loss": 2.6637,
      "theoretical_loss": 3.7704206236919795,
      "tokens_seen": 716732416
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003953660982948847,
      "loss": 2.8815,
      "theoretical_loss": 3.770385767476566,
      "tokens_seen": 716797952
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039535606820461386,
      "loss": 2.943,
      "theoretical_loss": 3.7703509153400976,
      "tokens_seen": 716863488
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039534603811434304,
      "loss": 2.8796,
      "theoretical_loss": 3.7703160672817257,
      "tokens_seen": 716929024
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003953360080240722,
      "loss": 2.9777,
      "theoretical_loss": 3.7702812233006,
      "tokens_seen": 716994560
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003953259779338014,
      "loss": 2.858,
      "theoretical_loss": 3.77024638339587,
      "tokens_seen": 717060096
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003953159478435306,
      "loss": 2.8236,
      "theoretical_loss": 3.7702115475666877,
      "tokens_seen": 717125632
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003953059177532598,
      "loss": 2.9503,
      "theoretical_loss": 3.7701767158122026,
      "tokens_seen": 717191168
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039529588766298895,
      "loss": 2.8212,
      "theoretical_loss": 3.7701418881315667,
      "tokens_seen": 717256704
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003952858575727182,
      "loss": 2.9965,
      "theoretical_loss": 3.7701070645239314,
      "tokens_seen": 717322240
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039527582748244736,
      "loss": 2.8431,
      "theoretical_loss": 3.7700722449884476,
      "tokens_seen": 717387776
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039526579739217654,
      "loss": 2.8866,
      "theoretical_loss": 3.7700374295242676,
      "tokens_seen": 717453312
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003952557673019057,
      "loss": 2.9852,
      "theoretical_loss": 3.770002618130544,
      "tokens_seen": 717518848
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003952457372116349,
      "loss": 2.7195,
      "theoretical_loss": 3.769967810806428,
      "tokens_seen": 717584384
    },
    {
      "epoch": 2.04,
      "objective/train/docs_used": 848381,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.755312204360962,
      "objective/train/theoretical_loss": 3.7699504086702085,
      "objective/train/tokens_used": 738077152,
      "theoretical_loss": 3.7699504086702085,
      "tokens_seen": 717617152
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003952357071213641,
      "loss": 2.9789,
      "theoretical_loss": 3.7699330075510735,
      "tokens_seen": 717649920
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003952256770310933,
      "loss": 2.678,
      "theoretical_loss": 3.7698982083636325,
      "tokens_seen": 717715456
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039521564694082245,
      "loss": 2.9702,
      "theoretical_loss": 3.7698634132432587,
      "tokens_seen": 717780992
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003952056168505517,
      "loss": 2.7129,
      "theoretical_loss": 3.7698286221891055,
      "tokens_seen": 717846528
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003951955867602808,
      "loss": 2.9426,
      "theoretical_loss": 3.769793835200326,
      "tokens_seen": 717912064
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039518555667001005,
      "loss": 2.8837,
      "theoretical_loss": 3.7697590522760747,
      "tokens_seen": 717977600
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039517552657973923,
      "loss": 2.89,
      "theoretical_loss": 3.7697242734155054,
      "tokens_seen": 718043136
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003951654964894684,
      "loss": 2.9438,
      "theoretical_loss": 3.7696894986177725,
      "tokens_seen": 718108672
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003951554663991976,
      "loss": 2.692,
      "theoretical_loss": 3.7696547278820307,
      "tokens_seen": 718174208
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039514543630892677,
      "loss": 3.0629,
      "theoretical_loss": 3.769619961207435,
      "tokens_seen": 718239744
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039513540621865595,
      "loss": 2.6968,
      "theoretical_loss": 3.7695851985931412,
      "tokens_seen": 718305280
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003951253761283852,
      "loss": 2.6478,
      "theoretical_loss": 3.7695504400383033,
      "tokens_seen": 718370816
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003951153460381143,
      "loss": 2.6335,
      "theoretical_loss": 3.7695156855420784,
      "tokens_seen": 718436352
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039510531594784355,
      "loss": 2.9399,
      "theoretical_loss": 3.7694809351036214,
      "tokens_seen": 718501888
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039509528585757273,
      "loss": 2.8495,
      "theoretical_loss": 3.769446188722089,
      "tokens_seen": 718567424
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003950852557673019,
      "loss": 2.9849,
      "theoretical_loss": 3.7694114463966377,
      "tokens_seen": 718632960
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003950752256770311,
      "loss": 2.7779,
      "theoretical_loss": 3.7693767081264236,
      "tokens_seen": 718698496
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003950651955867603,
      "loss": 2.9114,
      "theoretical_loss": 3.7693419739106044,
      "tokens_seen": 718764032
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039505516549648946,
      "loss": 2.8417,
      "theoretical_loss": 3.7693072437483366,
      "tokens_seen": 718829568
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003950451354062187,
      "loss": 2.8272,
      "theoretical_loss": 3.7692725176387785,
      "tokens_seen": 718895104
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003950351053159478,
      "loss": 2.6431,
      "theoretical_loss": 3.7692377955810867,
      "tokens_seen": 718960640
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039502507522567705,
      "loss": 2.7929,
      "theoretical_loss": 3.76920307757442,
      "tokens_seen": 719026176
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003950150451354062,
      "loss": 2.8803,
      "theoretical_loss": 3.7691683636179367,
      "tokens_seen": 719091712
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003950050150451354,
      "loss": 2.8331,
      "theoretical_loss": 3.7691336537107945,
      "tokens_seen": 719157248
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003949949849548646,
      "loss": 2.8321,
      "theoretical_loss": 3.769098947852152,
      "tokens_seen": 719222784
    },
    {
      "epoch": 2.04,
      "objective/train/docs_used": 849012,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8266098499298096,
      "objective/train/theoretical_loss": 3.7690815964407562,
      "objective/train/tokens_used": 739715552,
      "theoretical_loss": 3.7690815964407562,
      "tokens_seen": 719255552
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003949849548645938,
      "loss": 2.8435,
      "theoretical_loss": 3.7690642460411694,
      "tokens_seen": 719288320
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039497492477432296,
      "loss": 2.9216,
      "theoretical_loss": 3.769029548277005,
      "tokens_seen": 719353856
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003949648946840522,
      "loss": 2.851,
      "theoretical_loss": 3.768994854558818,
      "tokens_seen": 719419392
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003949548645937813,
      "loss": 2.8108,
      "theoretical_loss": 3.768960164885769,
      "tokens_seen": 719484928
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039494483450351056,
      "loss": 2.9195,
      "theoretical_loss": 3.768925479257017,
      "tokens_seen": 719550464
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003949348044132397,
      "loss": 2.8193,
      "theoretical_loss": 3.768890797671723,
      "tokens_seen": 719616000
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003949247743229689,
      "loss": 2.7336,
      "theoretical_loss": 3.7688561201290467,
      "tokens_seen": 719681536
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003949147442326981,
      "loss": 2.9406,
      "theoretical_loss": 3.7688214466281496,
      "tokens_seen": 719747072
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003949047141424273,
      "loss": 2.7792,
      "theoretical_loss": 3.7687867771681924,
      "tokens_seen": 719812608
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003948946840521565,
      "loss": 2.9954,
      "theoretical_loss": 3.768752111748336,
      "tokens_seen": 719878144
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039488465396188564,
      "loss": 3.0322,
      "theoretical_loss": 3.768717450367742,
      "tokens_seen": 719943680
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003948746238716149,
      "loss": 2.8552,
      "theoretical_loss": 3.7686827930255724,
      "tokens_seen": 720009216
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039486459378134406,
      "loss": 2.9811,
      "theoretical_loss": 3.7686481397209883,
      "tokens_seen": 720074752
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039485456369107324,
      "loss": 2.9801,
      "theoretical_loss": 3.768613490453153,
      "tokens_seen": 720140288
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003948445336008024,
      "loss": 2.9666,
      "theoretical_loss": 3.7685788452212288,
      "tokens_seen": 720205824
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003948345035105316,
      "loss": 2.6315,
      "theoretical_loss": 3.768544204024378,
      "tokens_seen": 720271360
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003948244734202608,
      "loss": 2.6003,
      "theoretical_loss": 3.7685095668617636,
      "tokens_seen": 720336896
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039481444332999,
      "loss": 2.9072,
      "theoretical_loss": 3.7684749337325485,
      "tokens_seen": 720402432
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039480441323971915,
      "loss": 2.9633,
      "theoretical_loss": 3.768440304635897,
      "tokens_seen": 720467968
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003947943831494484,
      "loss": 2.8997,
      "theoretical_loss": 3.7684056795709724,
      "tokens_seen": 720533504
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039478435305917756,
      "loss": 2.8266,
      "theoretical_loss": 3.768371058536939,
      "tokens_seen": 720599040
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039477432296890674,
      "loss": 2.9913,
      "theoretical_loss": 3.76833644153296,
      "tokens_seen": 720664576
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003947642928786359,
      "loss": 2.8822,
      "theoretical_loss": 3.768301828558201,
      "tokens_seen": 720730112
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003947542627883651,
      "loss": 2.872,
      "theoretical_loss": 3.7682672196118254,
      "tokens_seen": 720795648
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003947442326980943,
      "loss": 2.8944,
      "theoretical_loss": 3.768232614693,
      "tokens_seen": 720861184
    },
    {
      "epoch": 2.04,
      "objective/train/docs_used": 850658,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.295666217803955,
      "objective/train/theoretical_loss": 3.768215313743657,
      "objective/train/tokens_used": 741353952,
      "theoretical_loss": 3.768215313743657,
      "tokens_seen": 720893952
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003947342026078235,
      "loss": 2.4079,
      "theoretical_loss": 3.7681980138008884,
      "tokens_seen": 720926720
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039472417251755265,
      "loss": 3.0534,
      "theoretical_loss": 3.768163416934657,
      "tokens_seen": 720992256
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003947141424272819,
      "loss": 2.7413,
      "theoretical_loss": 3.7681288240934707,
      "tokens_seen": 721057792
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.000394704112337011,
      "loss": 2.9531,
      "theoretical_loss": 3.768094235276496,
      "tokens_seen": 721123328
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039469408224674025,
      "loss": 2.8847,
      "theoretical_loss": 3.7680596504828987,
      "tokens_seen": 721188864
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039468405215646943,
      "loss": 2.7946,
      "theoretical_loss": 3.7680250697118467,
      "tokens_seen": 721254400
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003946740220661986,
      "loss": 2.9812,
      "theoretical_loss": 3.7679904929625043,
      "tokens_seen": 721319936
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003946639919759278,
      "loss": 2.6334,
      "theoretical_loss": 3.7679559202340407,
      "tokens_seen": 721385472
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039465396188565697,
      "loss": 3.0705,
      "theoretical_loss": 3.7679213515256214,
      "tokens_seen": 721451008
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039464393179538615,
      "loss": 2.9367,
      "theoretical_loss": 3.7678867868364145,
      "tokens_seen": 721516544
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003946339017051154,
      "loss": 2.859,
      "theoretical_loss": 3.767852226165588,
      "tokens_seen": 721582080
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003946238716148445,
      "loss": 2.6695,
      "theoretical_loss": 3.7678176695123096,
      "tokens_seen": 721647616
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039461384152457375,
      "loss": 2.6397,
      "theoretical_loss": 3.7677831168757474,
      "tokens_seen": 721713152
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039460381143430293,
      "loss": 2.8705,
      "theoretical_loss": 3.7677485682550698,
      "tokens_seen": 721778688
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003945937813440321,
      "loss": 2.848,
      "theoretical_loss": 3.7677140236494457,
      "tokens_seen": 721844224
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003945837512537613,
      "loss": 2.8654,
      "theoretical_loss": 3.767679483058044,
      "tokens_seen": 721909760
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003945737211634905,
      "loss": 2.7036,
      "theoretical_loss": 3.7676449464800337,
      "tokens_seen": 721975296
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039456369107321966,
      "loss": 2.7538,
      "theoretical_loss": 3.7676104139145843,
      "tokens_seen": 722040832
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003945536609829489,
      "loss": 2.9453,
      "theoretical_loss": 3.7675758853608654,
      "tokens_seen": 722106368
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.000394543630892678,
      "loss": 2.9468,
      "theoretical_loss": 3.767541360818047,
      "tokens_seen": 722171904
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039453360080240725,
      "loss": 2.9775,
      "theoretical_loss": 3.7675068402852996,
      "tokens_seen": 722237440
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003945235707121364,
      "loss": 3.0682,
      "theoretical_loss": 3.7674723237617926,
      "tokens_seen": 722302976
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003945135406218656,
      "loss": 2.7909,
      "theoretical_loss": 3.767437811246698,
      "tokens_seen": 722368512
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003945035105315948,
      "loss": 2.8995,
      "theoretical_loss": 3.7674033027391856,
      "tokens_seen": 722434048
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.000394493480441324,
      "loss": 2.6872,
      "theoretical_loss": 3.7673687982384276,
      "tokens_seen": 722499584
    },
    {
      "epoch": 2.04,
      "objective/train/docs_used": 851365,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.842928171157837,
      "objective/train/theoretical_loss": 3.767351547490322,
      "objective/train/tokens_used": 742992352,
      "theoretical_loss": 3.767351547490322,
      "tokens_seen": 722532352
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039448345035105316,
      "loss": 2.8114,
      "theoretical_loss": 3.767334297743594,
      "tokens_seen": 722565120
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003944734202607824,
      "loss": 2.9769,
      "theoretical_loss": 3.767299801253858,
      "tokens_seen": 722630656
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003944633901705115,
      "loss": 2.7042,
      "theoretical_loss": 3.7672653087683905,
      "tokens_seen": 722696192
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039445336008024076,
      "loss": 2.8248,
      "theoretical_loss": 3.767230820286364,
      "tokens_seen": 722761728
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003944433299899699,
      "loss": 3.0078,
      "theoretical_loss": 3.76719633580695,
      "tokens_seen": 722827264
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003944332998996991,
      "loss": 2.6326,
      "theoretical_loss": 3.7671618553293227,
      "tokens_seen": 722892800
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003944232698094283,
      "loss": 2.8485,
      "theoretical_loss": 3.7671273788526536,
      "tokens_seen": 722958336
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003944132397191575,
      "loss": 2.6747,
      "theoretical_loss": 3.767092906376117,
      "tokens_seen": 723023872
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039440320962888666,
      "loss": 2.8055,
      "theoretical_loss": 3.7670584378988856,
      "tokens_seen": 723089408
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039439317953861584,
      "loss": 2.9246,
      "theoretical_loss": 3.767023973420133,
      "tokens_seen": 723154944
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.000394383149448345,
      "loss": 2.8103,
      "theoretical_loss": 3.766989512939033,
      "tokens_seen": 723220480
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039437311935807426,
      "loss": 2.8357,
      "theoretical_loss": 3.7669550564547594,
      "tokens_seen": 723286016
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003943630892678034,
      "loss": 2.8102,
      "theoretical_loss": 3.7669206039664878,
      "tokens_seen": 723351552
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003943530591775326,
      "loss": 2.5936,
      "theoretical_loss": 3.766886155473392,
      "tokens_seen": 723417088
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039434302908726175,
      "loss": 3.0184,
      "theoretical_loss": 3.766851710974646,
      "tokens_seen": 723482624
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.000394332998996991,
      "loss": 2.7742,
      "theoretical_loss": 3.766817270469427,
      "tokens_seen": 723548160
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039432296890672017,
      "loss": 2.7539,
      "theoretical_loss": 3.766782833956908,
      "tokens_seen": 723613696
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039431293881644935,
      "loss": 2.7714,
      "theoretical_loss": 3.766748401436266,
      "tokens_seen": 723679232
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039430290872617853,
      "loss": 2.9656,
      "theoretical_loss": 3.7667139729066768,
      "tokens_seen": 723744768
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039429287863590776,
      "loss": 2.8626,
      "theoretical_loss": 3.766679548367316,
      "tokens_seen": 723810304
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003942828485456369,
      "loss": 2.9391,
      "theoretical_loss": 3.76664512781736,
      "tokens_seen": 723875840
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003942728184553661,
      "loss": 2.8405,
      "theoretical_loss": 3.766610711255985,
      "tokens_seen": 723941376
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039426278836509525,
      "loss": 2.9499,
      "theoretical_loss": 3.7665762986823683,
      "tokens_seen": 724006912
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003942527582748245,
      "loss": 2.621,
      "theoretical_loss": 3.766541890095687,
      "tokens_seen": 724072448
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039424272818455367,
      "loss": 2.8388,
      "theoretical_loss": 3.7665074854951186,
      "tokens_seen": 724137984
    },
    {
      "epoch": 2.04,
      "objective/train/docs_used": 852859,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8854544162750244,
      "objective/train/theoretical_loss": 3.766490284689369,
      "objective/train/tokens_used": 744630752,
      "theoretical_loss": 3.766490284689369,
      "tokens_seen": 724170752
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039423269809428285,
      "loss": 2.9438,
      "theoretical_loss": 3.76647308487984,
      "tokens_seen": 724203520
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039422266800401203,
      "loss": 2.8946,
      "theoretical_loss": 3.7664386882490293,
      "tokens_seen": 724269056
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003942126379137412,
      "loss": 3.0176,
      "theoretical_loss": 3.7664042956018644,
      "tokens_seen": 724334592
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003942026078234704,
      "loss": 2.879,
      "theoretical_loss": 3.7663699069375234,
      "tokens_seen": 724400128
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039419257773319963,
      "loss": 2.8402,
      "theoretical_loss": 3.7663355222551855,
      "tokens_seen": 724465664
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039418254764292876,
      "loss": 2.8686,
      "theoretical_loss": 3.766301141554029,
      "tokens_seen": 724531200
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.000394172517552658,
      "loss": 2.8314,
      "theoretical_loss": 3.7662667648332326,
      "tokens_seen": 724596736
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003941624874623871,
      "loss": 2.7861,
      "theoretical_loss": 3.7662323920919762,
      "tokens_seen": 724662272
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039415245737211635,
      "loss": 2.9663,
      "theoretical_loss": 3.7661980233294385,
      "tokens_seen": 724727808
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003941424272818456,
      "loss": 2.8261,
      "theoretical_loss": 3.7661636585447997,
      "tokens_seen": 724793344
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003941323971915747,
      "loss": 2.8967,
      "theoretical_loss": 3.7661292977372405,
      "tokens_seen": 724858880
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039412236710130395,
      "loss": 2.8327,
      "theoretical_loss": 3.7660949409059397,
      "tokens_seen": 724924416
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039411233701103313,
      "loss": 2.9669,
      "theoretical_loss": 3.7660605880500784,
      "tokens_seen": 724989952
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003941023069207623,
      "loss": 2.6587,
      "theoretical_loss": 3.7660262391688373,
      "tokens_seen": 725055488
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003940922768304915,
      "loss": 2.7309,
      "theoretical_loss": 3.7659918942613975,
      "tokens_seen": 725121024
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003940822467402207,
      "loss": 2.7392,
      "theoretical_loss": 3.76595755332694,
      "tokens_seen": 725186560
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039407221664994986,
      "loss": 2.8362,
      "theoretical_loss": 3.765923216364646,
      "tokens_seen": 725252096
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003940621865596791,
      "loss": 2.746,
      "theoretical_loss": 3.765888883373698,
      "tokens_seen": 725317632
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003940521564694082,
      "loss": 2.7861,
      "theoretical_loss": 3.765854554353277,
      "tokens_seen": 725383168
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039404212637913745,
      "loss": 2.9553,
      "theoretical_loss": 3.765820229302565,
      "tokens_seen": 725448704
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003940320962888666,
      "loss": 2.954,
      "theoretical_loss": 3.765785908220746,
      "tokens_seen": 725514240
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003940220661985958,
      "loss": 2.7014,
      "theoretical_loss": 3.7657515911070005,
      "tokens_seen": 725579776
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.000394012036108325,
      "loss": 2.7269,
      "theoretical_loss": 3.765717277960513,
      "tokens_seen": 725645312
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003940020060180542,
      "loss": 2.8501,
      "theoretical_loss": 3.7656829687804656,
      "tokens_seen": 725710848
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039399197592778336,
      "loss": 2.6812,
      "theoretical_loss": 3.7656486635660427,
      "tokens_seen": 725776384
    },
    {
      "epoch": 2.04,
      "objective/train/docs_used": 853613,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.0343728065490723,
      "objective/train/theoretical_loss": 3.765631512445685,
      "objective/train/tokens_used": 746269152,
      "theoretical_loss": 3.765631512445685,
      "tokens_seen": 725809152
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003939819458375126,
      "loss": 2.9185,
      "theoretical_loss": 3.765614362316427,
      "tokens_seen": 725841920
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003939719157472417,
      "loss": 2.9212,
      "theoretical_loss": 3.765580065030803,
      "tokens_seen": 725907456
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039396188565697096,
      "loss": 2.9167,
      "theoretical_loss": 3.7655457717083545,
      "tokens_seen": 725972992
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003939518555667001,
      "loss": 2.8155,
      "theoretical_loss": 3.7655114823482654,
      "tokens_seen": 726038528
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003939418254764293,
      "loss": 2.7203,
      "theoretical_loss": 3.7654771969497203,
      "tokens_seen": 726104064
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003939317953861585,
      "loss": 2.7505,
      "theoretical_loss": 3.765442915511905,
      "tokens_seen": 726169600
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003939217652958877,
      "loss": 2.9916,
      "theoretical_loss": 3.765408638034004,
      "tokens_seen": 726235136
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039391173520561686,
      "loss": 2.7676,
      "theoretical_loss": 3.7653743645152025,
      "tokens_seen": 726300672
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039390170511534604,
      "loss": 2.635,
      "theoretical_loss": 3.765340094954686,
      "tokens_seen": 726366208
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003938916750250752,
      "loss": 2.8757,
      "theoretical_loss": 3.7653058293516404,
      "tokens_seen": 726431744
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039388164493480446,
      "loss": 2.8451,
      "theoretical_loss": 3.7652715677052515,
      "tokens_seen": 726497280
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003938716148445336,
      "loss": 2.8985,
      "theoretical_loss": 3.765237310014706,
      "tokens_seen": 726562816
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003938615847542628,
      "loss": 2.8408,
      "theoretical_loss": 3.76520305627919,
      "tokens_seen": 726628352
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039385155466399195,
      "loss": 2.9088,
      "theoretical_loss": 3.76516880649789,
      "tokens_seen": 726693888
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003938415245737212,
      "loss": 3.0388,
      "theoretical_loss": 3.765134560669994,
      "tokens_seen": 726759424
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039383149448345037,
      "loss": 2.9202,
      "theoretical_loss": 3.7651003187946888,
      "tokens_seen": 726824960
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039382146439317955,
      "loss": 2.8556,
      "theoretical_loss": 3.7650660808711613,
      "tokens_seen": 726890496
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039381143430290873,
      "loss": 2.7686,
      "theoretical_loss": 3.7650318468985993,
      "tokens_seen": 726956032
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039380140421263796,
      "loss": 2.7862,
      "theoretical_loss": 3.764997616876191,
      "tokens_seen": 727021568
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003937913741223671,
      "loss": 3.0449,
      "theoretical_loss": 3.7649633908031257,
      "tokens_seen": 727087104
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003937813440320963,
      "loss": 2.6458,
      "theoretical_loss": 3.7649291686785897,
      "tokens_seen": 727152640
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039377131394182545,
      "loss": 2.7193,
      "theoretical_loss": 3.7648949505017733,
      "tokens_seen": 727218176
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003937612838515547,
      "loss": 2.8087,
      "theoretical_loss": 3.7648607362718645,
      "tokens_seen": 727283712
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039375125376128387,
      "loss": 2.97,
      "theoretical_loss": 3.7648265259880525,
      "tokens_seen": 727349248
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039374122367101305,
      "loss": 3.2369,
      "theoretical_loss": 3.764792319649527,
      "tokens_seen": 727414784
    },
    {
      "epoch": 2.04,
      "objective/train/docs_used": 854233,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.945094108581543,
      "objective/train/theoretical_loss": 3.764775217959494,
      "objective/train/tokens_used": 747907552,
      "theoretical_loss": 3.764775217959494,
      "tokens_seen": 727447552
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039373119358074223,
      "loss": 2.8022,
      "theoretical_loss": 3.7647581172554783,
      "tokens_seen": 727480320
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003937211634904714,
      "loss": 2.9182,
      "theoretical_loss": 3.764723918805095,
      "tokens_seen": 727545856
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003937111334002006,
      "loss": 2.775,
      "theoretical_loss": 3.7646897242975674,
      "tokens_seen": 727611392
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039370110330992983,
      "loss": 3.0035,
      "theoretical_loss": 3.7646555337320873,
      "tokens_seen": 727676928
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039369107321965896,
      "loss": 2.8621,
      "theoretical_loss": 3.764621347107843,
      "tokens_seen": 727742464
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003936810431293882,
      "loss": 3.0658,
      "theoretical_loss": 3.764587164424027,
      "tokens_seen": 727808000
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003936710130391173,
      "loss": 2.7817,
      "theoretical_loss": 3.76455298567983,
      "tokens_seen": 727873536
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039366098294884655,
      "loss": 2.9779,
      "theoretical_loss": 3.764518810874443,
      "tokens_seen": 727939072
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039365095285857573,
      "loss": 2.8654,
      "theoretical_loss": 3.7644846400070575,
      "tokens_seen": 728004608
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003936409227683049,
      "loss": 2.7282,
      "theoretical_loss": 3.7644504730768653,
      "tokens_seen": 728070144
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003936308926780341,
      "loss": 2.7755,
      "theoretical_loss": 3.764416310083059,
      "tokens_seen": 728135680
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039362086258776333,
      "loss": 2.9102,
      "theoretical_loss": 3.7643821510248303,
      "tokens_seen": 728201216
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039361083249749246,
      "loss": 2.9127,
      "theoretical_loss": 3.764347995901372,
      "tokens_seen": 728266752
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003936008024072217,
      "loss": 2.9826,
      "theoretical_loss": 3.764313844711876,
      "tokens_seen": 728332288
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003935907723169508,
      "loss": 2.7889,
      "theoretical_loss": 3.7642796974555366,
      "tokens_seen": 728397824
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039358074222668006,
      "loss": 2.7005,
      "theoretical_loss": 3.764245554131546,
      "tokens_seen": 728463360
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039357071213640924,
      "loss": 3.0278,
      "theoretical_loss": 3.7642114147390977,
      "tokens_seen": 728528896
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003935606820461384,
      "loss": 2.8777,
      "theoretical_loss": 3.764177279277386,
      "tokens_seen": 728594432
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003935506519558676,
      "loss": 3.0676,
      "theoretical_loss": 3.764143147745605,
      "tokens_seen": 728659968
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003935406218655968,
      "loss": 2.8788,
      "theoretical_loss": 3.7641090201429477,
      "tokens_seen": 728725504
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039353059177532596,
      "loss": 2.7074,
      "theoretical_loss": 3.764074896468609,
      "tokens_seen": 728791040
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003935205616850552,
      "loss": 2.8258,
      "theoretical_loss": 3.7640407767217843,
      "tokens_seen": 728856576
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003935105315947843,
      "loss": 2.9152,
      "theoretical_loss": 3.764006660901667,
      "tokens_seen": 728922112
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039350050150451356,
      "loss": 2.9142,
      "theoretical_loss": 3.7639725490074536,
      "tokens_seen": 728987648
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003934904714142427,
      "loss": 2.8464,
      "theoretical_loss": 3.7639384410383387,
      "tokens_seen": 729053184
    },
    {
      "epoch": 2.04,
      "objective/train/docs_used": 855726,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.673104763031006,
      "objective/train/theoretical_loss": 3.763921388525442,
      "objective/train/tokens_used": 749545952,
      "theoretical_loss": 3.763921388525442,
      "tokens_seen": 729085952
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003934804413239719,
      "loss": 2.9341,
      "theoretical_loss": 3.763904336993518,
      "tokens_seen": 729118720
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003934704112337011,
      "loss": 2.7971,
      "theoretical_loss": 3.7638702368721875,
      "tokens_seen": 729184256
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003934603811434303,
      "loss": 2.9572,
      "theoretical_loss": 3.7638361406735434,
      "tokens_seen": 729249792
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039345035105315946,
      "loss": 2.9707,
      "theoretical_loss": 3.7638020483967813,
      "tokens_seen": 729315328
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003934403209628887,
      "loss": 2.6327,
      "theoretical_loss": 3.7637679600410987,
      "tokens_seen": 729380864
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039343029087261783,
      "loss": 2.9909,
      "theoretical_loss": 3.7637338756056913,
      "tokens_seen": 729446400
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039342026078234706,
      "loss": 2.7681,
      "theoretical_loss": 3.7636997950897566,
      "tokens_seen": 729511936
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003934102306920762,
      "loss": 2.8921,
      "theoretical_loss": 3.763665718492492,
      "tokens_seen": 729577472
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003934002006018054,
      "loss": 2.948,
      "theoretical_loss": 3.7636316458130956,
      "tokens_seen": 729643008
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039339017051153466,
      "loss": 2.8759,
      "theoretical_loss": 3.7635975770507635,
      "tokens_seen": 729708544
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003933801404212638,
      "loss": 2.7037,
      "theoretical_loss": 3.763563512204695,
      "tokens_seen": 729774080
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.000393370110330993,
      "loss": 2.8733,
      "theoretical_loss": 3.7635294512740876,
      "tokens_seen": 729839616
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039336008024072215,
      "loss": 2.7322,
      "theoretical_loss": 3.7634953942581397,
      "tokens_seen": 729905152
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003933500501504514,
      "loss": 2.8427,
      "theoretical_loss": 3.7634613411560505,
      "tokens_seen": 729970688
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039334002006018057,
      "loss": 2.8633,
      "theoretical_loss": 3.7634272919670186,
      "tokens_seen": 730036224
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039332998996990975,
      "loss": 2.7698,
      "theoretical_loss": 3.763393246690243,
      "tokens_seen": 730101760
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039331995987963893,
      "loss": 2.883,
      "theoretical_loss": 3.7633592053249236,
      "tokens_seen": 730167296
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039330992978936816,
      "loss": 2.8142,
      "theoretical_loss": 3.763325167870259,
      "tokens_seen": 730232832
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003932998996990973,
      "loss": 3.0139,
      "theoretical_loss": 3.76329113432545,
      "tokens_seen": 730298368
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003932898696088265,
      "loss": 2.8159,
      "theoretical_loss": 3.7632571046896963,
      "tokens_seen": 730363904
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039327983951855565,
      "loss": 2.8133,
      "theoretical_loss": 3.763223078962198,
      "tokens_seen": 730429440
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003932698094282849,
      "loss": 2.9151,
      "theoretical_loss": 3.763189057142156,
      "tokens_seen": 730494976
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039325977933801407,
      "loss": 2.6749,
      "theoretical_loss": 3.7631550392287707,
      "tokens_seen": 730560512
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039324974924774325,
      "loss": 2.9721,
      "theoretical_loss": 3.7631210252212437,
      "tokens_seen": 730626048
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039323971915747243,
      "loss": 2.8593,
      "theoretical_loss": 3.763087015118776,
      "tokens_seen": 730691584
    },
    {
      "epoch": 2.04,
      "objective/train/docs_used": 856181,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.754777431488037,
      "objective/train/theoretical_loss": 3.7630700115316893,
      "objective/train/tokens_used": 751184352,
      "theoretical_loss": 3.7630700115316893,
      "tokens_seen": 730724352
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003932296890672016,
      "loss": 2.8694,
      "theoretical_loss": 3.7630530089205685,
      "tokens_seen": 730757120
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003932196589769308,
      "loss": 2.8381,
      "theoretical_loss": 3.7630190066258233,
      "tokens_seen": 730822656
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039320962888666003,
      "loss": 2.8851,
      "theoretical_loss": 3.762985008233743,
      "tokens_seen": 730888192
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039319959879638916,
      "loss": 3.0075,
      "theoretical_loss": 3.762951013743529,
      "tokens_seen": 730953728
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003931895687061184,
      "loss": 3.1236,
      "theoretical_loss": 3.762917023154384,
      "tokens_seen": 731019264
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003931795386158475,
      "loss": 2.8348,
      "theoretical_loss": 3.762883036465511,
      "tokens_seen": 731084800
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039316950852557675,
      "loss": 2.8395,
      "theoretical_loss": 3.7628490536761117,
      "tokens_seen": 731150336
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039315947843530593,
      "loss": 2.8552,
      "theoretical_loss": 3.7628150747853906,
      "tokens_seen": 731215872
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003931494483450351,
      "loss": 2.9167,
      "theoretical_loss": 3.7627810997925506,
      "tokens_seen": 731281408
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003931394182547643,
      "loss": 3.2027,
      "theoretical_loss": 3.7627471286967946,
      "tokens_seen": 731346944
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039312938816449353,
      "loss": 2.9124,
      "theoretical_loss": 3.7627131614973277,
      "tokens_seen": 731412480
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039311935807422266,
      "loss": 2.8646,
      "theoretical_loss": 3.762679198193353,
      "tokens_seen": 731478016
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003931093279839519,
      "loss": 2.9585,
      "theoretical_loss": 3.762645238784075,
      "tokens_seen": 731543552
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.000393099297893681,
      "loss": 2.7172,
      "theoretical_loss": 3.7626112832686984,
      "tokens_seen": 731609088
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039308926780341026,
      "loss": 2.8292,
      "theoretical_loss": 3.762577331646428,
      "tokens_seen": 731674624
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039307923771313944,
      "loss": 2.8594,
      "theoretical_loss": 3.762543383916469,
      "tokens_seen": 731740160
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003930692076228686,
      "loss": 2.9352,
      "theoretical_loss": 3.7625094400780257,
      "tokens_seen": 731805696
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003930591775325978,
      "loss": 2.8496,
      "theoretical_loss": 3.7624755001303045,
      "tokens_seen": 731871232
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.000393049147442327,
      "loss": 2.9329,
      "theoretical_loss": 3.7624415640725104,
      "tokens_seen": 731936768
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039303911735205616,
      "loss": 2.7525,
      "theoretical_loss": 3.7624076319038506,
      "tokens_seen": 732002304
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003930290872617854,
      "loss": 2.702,
      "theoretical_loss": 3.7623737036235294,
      "tokens_seen": 732067840
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003930190571715145,
      "loss": 2.7606,
      "theoretical_loss": 3.7623397792307545,
      "tokens_seen": 732133376
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039300902708124376,
      "loss": 2.7868,
      "theoretical_loss": 3.762305858724732,
      "tokens_seen": 732198912
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003929989969909729,
      "loss": 3.0959,
      "theoretical_loss": 3.7622719421046695,
      "tokens_seen": 732264448
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003929889669007021,
      "loss": 3.0838,
      "theoretical_loss": 3.7622380293697733,
      "tokens_seen": 732329984
    },
    {
      "epoch": 2.04,
      "objective/train/docs_used": 857755,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.856901168823242,
      "objective/train/theoretical_loss": 3.762221074459015,
      "objective/train/tokens_used": 752822752,
      "theoretical_loss": 3.762221074459015,
      "tokens_seen": 732362752
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003929789368104313,
      "loss": 2.7062,
      "theoretical_loss": 3.762204120519251,
      "tokens_seen": 732395520
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003929689067201605,
      "loss": 2.7705,
      "theoretical_loss": 3.76217021555231,
      "tokens_seen": 732461056
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039295887662988967,
      "loss": 2.9732,
      "theoretical_loss": 3.7621363144681585,
      "tokens_seen": 732526592
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003929488465396189,
      "loss": 2.9889,
      "theoretical_loss": 3.7621024172660045,
      "tokens_seen": 732592128
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039293881644934803,
      "loss": 2.7959,
      "theoretical_loss": 3.7620685239450555,
      "tokens_seen": 732657664
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039292878635907726,
      "loss": 2.9595,
      "theoretical_loss": 3.762034634504521,
      "tokens_seen": 732723200
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003929187562688064,
      "loss": 2.746,
      "theoretical_loss": 3.762000748943609,
      "tokens_seen": 732788736
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003929087261785356,
      "loss": 2.9863,
      "theoretical_loss": 3.7619668672615285,
      "tokens_seen": 732854272
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003928986960882648,
      "loss": 2.6103,
      "theoretical_loss": 3.761932989457489,
      "tokens_seen": 732919808
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.000392888665997994,
      "loss": 2.9042,
      "theoretical_loss": 3.7618991155307,
      "tokens_seen": 732985344
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039287863590772317,
      "loss": 2.9509,
      "theoretical_loss": 3.761865245480371,
      "tokens_seen": 733050880
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039286860581745235,
      "loss": 2.9416,
      "theoretical_loss": 3.7618313793057117,
      "tokens_seen": 733116416
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039285857572718153,
      "loss": 2.9895,
      "theoretical_loss": 3.761797517005933,
      "tokens_seen": 733181952
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039284854563691077,
      "loss": 2.6667,
      "theoretical_loss": 3.7617636585802434,
      "tokens_seen": 733247488
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003928385155466399,
      "loss": 2.8332,
      "theoretical_loss": 3.7617298040278557,
      "tokens_seen": 733313024
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039282848545636913,
      "loss": 2.8072,
      "theoretical_loss": 3.761695953347979,
      "tokens_seen": 733378560
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039281845536609825,
      "loss": 2.8533,
      "theoretical_loss": 3.7616621065398257,
      "tokens_seen": 733444096
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003928084252758275,
      "loss": 2.8354,
      "theoretical_loss": 3.761628263602606,
      "tokens_seen": 733509632
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039279839518555667,
      "loss": 2.695,
      "theoretical_loss": 3.7615944245355317,
      "tokens_seen": 733575168
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039278836509528585,
      "loss": 2.7853,
      "theoretical_loss": 3.7615605893378152,
      "tokens_seen": 733640704
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039277833500501503,
      "loss": 2.8961,
      "theoretical_loss": 3.7615267580086673,
      "tokens_seen": 733706240
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039276830491474427,
      "loss": 3.0298,
      "theoretical_loss": 3.761492930547301,
      "tokens_seen": 733771776
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003927582748244734,
      "loss": 2.914,
      "theoretical_loss": 3.7614591069529286,
      "tokens_seen": 733837312
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039274824473420263,
      "loss": 2.8187,
      "theoretical_loss": 3.761425287224763,
      "tokens_seen": 733902848
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039273821464393176,
      "loss": 2.7914,
      "theoretical_loss": 3.7613914713620167,
      "tokens_seen": 733968384
    },
    {
      "epoch": 2.04,
      "objective/train/docs_used": 858482,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.96612548828125,
      "objective/train/theoretical_loss": 3.7613745648799295,
      "objective/train/tokens_used": 754461152,
      "theoretical_loss": 3.7613745648799295,
      "tokens_seen": 734001152
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.000392728184553661,
      "loss": 2.6728,
      "theoretical_loss": 3.7613576593639024,
      "tokens_seen": 734033920
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003927181544633902,
      "loss": 2.9194,
      "theoretical_loss": 3.7613238512296348,
      "tokens_seen": 734099456
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039270812437311936,
      "loss": 2.7902,
      "theoretical_loss": 3.7612900469584263,
      "tokens_seen": 734164992
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039269809428284854,
      "loss": 3.0611,
      "theoretical_loss": 3.7612562465494905,
      "tokens_seen": 734230528
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003926880641925777,
      "loss": 2.6482,
      "theoretical_loss": 3.7612224500020424,
      "tokens_seen": 734296064
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003926780341023069,
      "loss": 2.7174,
      "theoretical_loss": 3.7611886573152957,
      "tokens_seen": 734361600
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039266800401203613,
      "loss": 2.6784,
      "theoretical_loss": 3.7611548684884655,
      "tokens_seen": 734427136
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039265797392176526,
      "loss": 2.886,
      "theoretical_loss": 3.7611210835207656,
      "tokens_seen": 734492672
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003926479438314945,
      "loss": 2.9471,
      "theoretical_loss": 3.7610873024114113,
      "tokens_seen": 734558208
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039263791374122373,
      "loss": 2.8858,
      "theoretical_loss": 3.7610535251596184,
      "tokens_seen": 734623744
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039262788365095286,
      "loss": 2.685,
      "theoretical_loss": 3.7610197517646022,
      "tokens_seen": 734689280
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003926178535606821,
      "loss": 2.6936,
      "theoretical_loss": 3.760985982225577,
      "tokens_seen": 734754816
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003926078234704112,
      "loss": 2.6473,
      "theoretical_loss": 3.76095221654176,
      "tokens_seen": 734820352
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039259779338014046,
      "loss": 2.7335,
      "theoretical_loss": 3.7609184547123675,
      "tokens_seen": 734885888
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039258776328986964,
      "loss": 2.9301,
      "theoretical_loss": 3.760884696736615,
      "tokens_seen": 734951424
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003925777331995988,
      "loss": 2.7825,
      "theoretical_loss": 3.7608509426137195,
      "tokens_seen": 735016960
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.000392567703109328,
      "loss": 2.8205,
      "theoretical_loss": 3.7608171923428975,
      "tokens_seen": 735082496
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003925576730190572,
      "loss": 2.8997,
      "theoretical_loss": 3.7607834459233667,
      "tokens_seen": 735148032
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039254764292878636,
      "loss": 2.6886,
      "theoretical_loss": 3.7607497033543433,
      "tokens_seen": 735213568
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003925376128385156,
      "loss": 2.9114,
      "theoretical_loss": 3.7607159646350454,
      "tokens_seen": 735279104
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003925275827482447,
      "loss": 2.9142,
      "theoretical_loss": 3.760682229764691,
      "tokens_seen": 735344640
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039251755265797396,
      "loss": 3.1446,
      "theoretical_loss": 3.7606484987424973,
      "tokens_seen": 735410176
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003925075225677031,
      "loss": 2.8202,
      "theoretical_loss": 3.760614771567683,
      "tokens_seen": 735475712
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003924974924774323,
      "loss": 2.784,
      "theoretical_loss": 3.760581048239466,
      "tokens_seen": 735541248
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003924874623871615,
      "loss": 2.8307,
      "theoretical_loss": 3.760547328757066,
      "tokens_seen": 735606784
    },
    {
      "epoch": 2.04,
      "objective/train/docs_used": 859784,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6183910369873047,
      "objective/train/theoretical_loss": 3.7605304704578026,
      "objective/train/tokens_used": 756099552,
      "theoretical_loss": 3.7605304704578026,
      "tokens_seen": 735639552
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003924774322968907,
      "loss": 2.844,
      "theoretical_loss": 3.7605136131197003,
      "tokens_seen": 735672320
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039246740220661987,
      "loss": 3.1915,
      "theoretical_loss": 3.7604799013265895,
      "tokens_seen": 735737856
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003924573721163491,
      "loss": 3.0552,
      "theoretical_loss": 3.760446193376952,
      "tokens_seen": 735803392
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039244734202607823,
      "loss": 2.871,
      "theoretical_loss": 3.7604124892700073,
      "tokens_seen": 735868928
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039243731193580746,
      "loss": 2.6401,
      "theoretical_loss": 3.7603787890049762,
      "tokens_seen": 735934464
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003924272818455366,
      "loss": 2.9156,
      "theoretical_loss": 3.760345092581077,
      "tokens_seen": 736000000
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003924172517552658,
      "loss": 2.7434,
      "theoretical_loss": 3.7603113999975317,
      "tokens_seen": 736065536
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.000392407221664995,
      "loss": 2.6808,
      "theoretical_loss": 3.7602777112535595,
      "tokens_seen": 736131072
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003923971915747242,
      "loss": 2.8149,
      "theoretical_loss": 3.7602440263483814,
      "tokens_seen": 736196608
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039238716148445337,
      "loss": 2.8053,
      "theoretical_loss": 3.7602103452812186,
      "tokens_seen": 736262144
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039237713139418255,
      "loss": 2.7973,
      "theoretical_loss": 3.760176668051291,
      "tokens_seen": 736327680
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039236710130391173,
      "loss": 2.9008,
      "theoretical_loss": 3.760142994657822,
      "tokens_seen": 736393216
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039235707121364097,
      "loss": 3.034,
      "theoretical_loss": 3.760109325100032,
      "tokens_seen": 736458752
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003923470411233701,
      "loss": 2.8678,
      "theoretical_loss": 3.7600756593771427,
      "tokens_seen": 736524288
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039233701103309933,
      "loss": 2.8121,
      "theoretical_loss": 3.7600419974883765,
      "tokens_seen": 736589824
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039232698094282846,
      "loss": 2.7452,
      "theoretical_loss": 3.7600083394329555,
      "tokens_seen": 736655360
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003923169508525577,
      "loss": 2.7862,
      "theoretical_loss": 3.7599746852101026,
      "tokens_seen": 736720896
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039230692076228687,
      "loss": 2.8344,
      "theoretical_loss": 3.7599410348190396,
      "tokens_seen": 736786432
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039229689067201605,
      "loss": 2.8182,
      "theoretical_loss": 3.7599073882589904,
      "tokens_seen": 736851968
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039228686058174523,
      "loss": 2.7019,
      "theoretical_loss": 3.759873745529178,
      "tokens_seen": 736917504
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039227683049147447,
      "loss": 2.7842,
      "theoretical_loss": 3.7598401066288254,
      "tokens_seen": 736983040
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003922668004012036,
      "loss": 2.854,
      "theoretical_loss": 3.759806471557156,
      "tokens_seen": 737048576
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039225677031093283,
      "loss": 2.7951,
      "theoretical_loss": 3.7597728403133948,
      "tokens_seen": 737114112
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039224674022066196,
      "loss": 3.0258,
      "theoretical_loss": 3.759739212896765,
      "tokens_seen": 737179648
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003922367101303912,
      "loss": 2.6118,
      "theoretical_loss": 3.759705589306491,
      "tokens_seen": 737245184
    },
    {
      "epoch": 2.04,
      "objective/train/docs_used": 860487,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.353097438812256,
      "objective/train/theoretical_loss": 3.759688778945995,
      "objective/train/tokens_used": 757737952,
      "theoretical_loss": 3.759688778945995,
      "tokens_seen": 737277952
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003922266800401204,
      "loss": 2.6294,
      "theoretical_loss": 3.759671969541797,
      "tokens_seen": 737310720
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039221664994984956,
      "loss": 2.9801,
      "theoretical_loss": 3.759638353601909,
      "tokens_seen": 737376256
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039220661985957874,
      "loss": 2.6816,
      "theoretical_loss": 3.7596047414860507,
      "tokens_seen": 737441792
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003921965897693079,
      "loss": 2.841,
      "theoretical_loss": 3.7595711331934476,
      "tokens_seen": 737507328
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003921865596790371,
      "loss": 2.8772,
      "theoretical_loss": 3.7595375287233255,
      "tokens_seen": 737572864
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039217652958876633,
      "loss": 2.7096,
      "theoretical_loss": 3.75950392807491,
      "tokens_seen": 737638400
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039216649949849546,
      "loss": 2.7568,
      "theoretical_loss": 3.7594703312474262,
      "tokens_seen": 737703936
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003921564694082247,
      "loss": 2.7315,
      "theoretical_loss": 3.7594367382401015,
      "tokens_seen": 737769472
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003921464393179539,
      "loss": 3.0837,
      "theoretical_loss": 3.759403149052161,
      "tokens_seen": 737835008
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039213640922768306,
      "loss": 2.7529,
      "theoretical_loss": 3.7593695636828324,
      "tokens_seen": 737900544
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039212637913741224,
      "loss": 2.9824,
      "theoretical_loss": 3.7593359821313417,
      "tokens_seen": 737966080
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003921163490471414,
      "loss": 2.8387,
      "theoretical_loss": 3.7593024043969163,
      "tokens_seen": 738031616
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003921063189568706,
      "loss": 2.9527,
      "theoretical_loss": 3.759268830478783,
      "tokens_seen": 738097152
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039209628886659984,
      "loss": 2.7311,
      "theoretical_loss": 3.7592352603761703,
      "tokens_seen": 738162688
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039208625877632896,
      "loss": 2.7515,
      "theoretical_loss": 3.7592016940883046,
      "tokens_seen": 738228224
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003920762286860582,
      "loss": 3.0343,
      "theoretical_loss": 3.7591681316144143,
      "tokens_seen": 738293760
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003920661985957873,
      "loss": 2.7702,
      "theoretical_loss": 3.759134572953728,
      "tokens_seen": 738359296
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039205616850551656,
      "loss": 2.8235,
      "theoretical_loss": 3.7591010181054734,
      "tokens_seen": 738424832
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039204613841524574,
      "loss": 2.6255,
      "theoretical_loss": 3.7590674670688795,
      "tokens_seen": 738490368
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003920361083249749,
      "loss": 2.8931,
      "theoretical_loss": 3.759033919843175,
      "tokens_seen": 738555904
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003920260782347041,
      "loss": 2.9703,
      "theoretical_loss": 3.759000376427589,
      "tokens_seen": 738621440
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003920160481444333,
      "loss": 2.9791,
      "theoretical_loss": 3.75896683682135,
      "tokens_seen": 738686976
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039200601805416247,
      "loss": 2.8601,
      "theoretical_loss": 3.758933301023689,
      "tokens_seen": 738752512
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003919959879638917,
      "loss": 2.8717,
      "theoretical_loss": 3.758899769033835,
      "tokens_seen": 738818048
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039198595787362083,
      "loss": 3.0159,
      "theoretical_loss": 3.7588662408510167,
      "tokens_seen": 738883584
    },
    {
      "epoch": 2.04,
      "objective/train/docs_used": 861921,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8210134506225586,
      "objective/train/theoretical_loss": 3.758849478187006,
      "objective/train/tokens_used": 759376352,
      "theoretical_loss": 3.758849478187006,
      "tokens_seen": 738916352
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039197592778335007,
      "loss": 2.9699,
      "theoretical_loss": 3.7588327164744664,
      "tokens_seen": 738949120
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039196589769307925,
      "loss": 2.7979,
      "theoretical_loss": 3.7587991959034133,
      "tokens_seen": 739014656
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039195586760280843,
      "loss": 2.8731,
      "theoretical_loss": 3.758765679137088,
      "tokens_seen": 739080192
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003919458375125376,
      "loss": 2.8623,
      "theoretical_loss": 3.7587321661747213,
      "tokens_seen": 739145728
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003919358074222668,
      "loss": 2.7755,
      "theoretical_loss": 3.758698657015545,
      "tokens_seen": 739211264
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039192577733199597,
      "loss": 2.9633,
      "theoretical_loss": 3.7586651516587892,
      "tokens_seen": 739276800
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003919157472417252,
      "loss": 2.9176,
      "theoretical_loss": 3.758631650103686,
      "tokens_seen": 739342336
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039190571715145433,
      "loss": 2.8127,
      "theoretical_loss": 3.7585981523494674,
      "tokens_seen": 739407872
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039189568706118357,
      "loss": 2.8019,
      "theoretical_loss": 3.7585646583953656,
      "tokens_seen": 739473408
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039188565697091275,
      "loss": 2.7282,
      "theoretical_loss": 3.7585311682406113,
      "tokens_seen": 739538944
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039187562688064193,
      "loss": 2.8132,
      "theoretical_loss": 3.758497681884439,
      "tokens_seen": 739604480
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039186559679037117,
      "loss": 3.1263,
      "theoretical_loss": 3.7584641993260792,
      "tokens_seen": 739670016
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003918555667001003,
      "loss": 2.7582,
      "theoretical_loss": 3.758430720564766,
      "tokens_seen": 739735552
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039184553660982953,
      "loss": 2.845,
      "theoretical_loss": 3.758397245599732,
      "tokens_seen": 739801088
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039183550651955866,
      "loss": 2.7409,
      "theoretical_loss": 3.7583637744302107,
      "tokens_seen": 739866624
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003918254764292879,
      "loss": 2.9599,
      "theoretical_loss": 3.758330307055436,
      "tokens_seen": 739932160
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039181544633901707,
      "loss": 2.9472,
      "theoretical_loss": 3.758296843474641,
      "tokens_seen": 739997696
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039180541624874625,
      "loss": 2.7435,
      "theoretical_loss": 3.7582633836870603,
      "tokens_seen": 740063232
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039179538615847543,
      "loss": 2.6432,
      "theoretical_loss": 3.758229927691927,
      "tokens_seen": 740128768
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039178535606820467,
      "loss": 3.0198,
      "theoretical_loss": 3.758196475488476,
      "tokens_seen": 740194304
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003917753259779338,
      "loss": 2.8823,
      "theoretical_loss": 3.758163027075943,
      "tokens_seen": 740259840
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039176529588766303,
      "loss": 2.7262,
      "theoretical_loss": 3.758129582453561,
      "tokens_seen": 740325376
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039175526579739216,
      "loss": 2.9131,
      "theoretical_loss": 3.758096141620566,
      "tokens_seen": 740390912
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003917452357071214,
      "loss": 2.8228,
      "theoretical_loss": 3.7580627045761936,
      "tokens_seen": 740456448
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003917352056168506,
      "loss": 2.7725,
      "theoretical_loss": 3.7580292713196792,
      "tokens_seen": 740521984
    },
    {
      "epoch": 2.04,
      "objective/train/docs_used": 862761,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8765251636505127,
      "objective/train/theoretical_loss": 3.7580125561116295,
      "objective/train/tokens_used": 761014752,
      "theoretical_loss": 3.7580125561116295,
      "tokens_seen": 740554752
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039172517552657976,
      "loss": 2.8848,
      "theoretical_loss": 3.757995841850258,
      "tokens_seen": 740587520
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039171514543630894,
      "loss": 2.6698,
      "theoretical_loss": 3.7579624161671665,
      "tokens_seen": 740653056
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003917051153460381,
      "loss": 2.8602,
      "theoretical_loss": 3.75792899426964,
      "tokens_seen": 740718592
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003916950852557673,
      "loss": 2.8744,
      "theoretical_loss": 3.757895576156916,
      "tokens_seen": 740784128
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039168505516549653,
      "loss": 2.9496,
      "theoretical_loss": 3.75786216182823,
      "tokens_seen": 740849664
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039167502507522566,
      "loss": 2.9308,
      "theoretical_loss": 3.757828751282821,
      "tokens_seen": 740915200
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003916649949849549,
      "loss": 3.0178,
      "theoretical_loss": 3.757795344519923,
      "tokens_seen": 740980736
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003916549648946841,
      "loss": 2.8622,
      "theoretical_loss": 3.7577619415387753,
      "tokens_seen": 741046272
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039164493480441326,
      "loss": 2.8771,
      "theoretical_loss": 3.757728542338615,
      "tokens_seen": 741111808
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039163490471414244,
      "loss": 2.652,
      "theoretical_loss": 3.75769514691868,
      "tokens_seen": 741177344
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003916248746238716,
      "loss": 2.9017,
      "theoretical_loss": 3.7576617552782077,
      "tokens_seen": 741242880
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003916148445336008,
      "loss": 2.9632,
      "theoretical_loss": 3.757628367416437,
      "tokens_seen": 741308416
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039160481444333004,
      "loss": 2.6733,
      "theoretical_loss": 3.757594983332605,
      "tokens_seen": 741373952
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039159478435305916,
      "loss": 2.6975,
      "theoretical_loss": 3.7575616030259518,
      "tokens_seen": 741439488
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003915847542627884,
      "loss": 2.9127,
      "theoretical_loss": 3.7575282264957153,
      "tokens_seen": 741505024
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003915747241725175,
      "loss": 2.8389,
      "theoretical_loss": 3.7574948537411355,
      "tokens_seen": 741570560
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039156469408224676,
      "loss": 3.0937,
      "theoretical_loss": 3.7574614847614503,
      "tokens_seen": 741636096
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039155466399197594,
      "loss": 2.979,
      "theoretical_loss": 3.7574281195559003,
      "tokens_seen": 741701632
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003915446339017051,
      "loss": 2.5656,
      "theoretical_loss": 3.7573947581237244,
      "tokens_seen": 741767168
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003915346038114343,
      "loss": 2.6176,
      "theoretical_loss": 3.757361400464163,
      "tokens_seen": 741832704
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0003915245737211635,
      "loss": 2.7122,
      "theoretical_loss": 3.7573280465764567,
      "tokens_seen": 741898240
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00039151454363089267,
      "loss": 3.0246,
      "theoretical_loss": 3.7572946964598453,
      "tokens_seen": 741963776
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003915045135406219,
      "loss": 2.7568,
      "theoretical_loss": 3.757261350113569,
      "tokens_seen": 742029312
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00039149448345035103,
      "loss": 2.8737,
      "theoretical_loss": 3.7572280075368694,
      "tokens_seen": 742094848
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00039148445336008027,
      "loss": 2.8704,
      "theoretical_loss": 3.757194668728987,
      "tokens_seen": 742160384
    },
    {
      "epoch": 2.05,
      "objective/train/docs_used": 864185,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.501842498779297,
      "objective/train/theoretical_loss": 3.757178000738115,
      "objective/train/tokens_used": 762653152,
      "theoretical_loss": 3.757178000738115,
      "tokens_seen": 742193152
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00039147442326980945,
      "loss": 3.1039,
      "theoretical_loss": 3.7571613336891634,
      "tokens_seen": 742225920
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00039146439317953863,
      "loss": 2.7548,
      "theoretical_loss": 3.75712800241664,
      "tokens_seen": 742291456
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003914543630892678,
      "loss": 2.8036,
      "theoretical_loss": 3.7570946749106584,
      "tokens_seen": 742356992
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.000391444332998997,
      "loss": 2.584,
      "theoretical_loss": 3.7570613511704605,
      "tokens_seen": 742422528
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00039143430290872617,
      "loss": 2.9916,
      "theoretical_loss": 3.7570280311952886,
      "tokens_seen": 742488064
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003914242728184554,
      "loss": 2.9727,
      "theoretical_loss": 3.7569947149843843,
      "tokens_seen": 742553600
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00039141424272818453,
      "loss": 2.9331,
      "theoretical_loss": 3.7569614025369917,
      "tokens_seen": 742619136
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00039140421263791377,
      "loss": 2.5328,
      "theoretical_loss": 3.756928093852352,
      "tokens_seen": 742684672
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003913941825476429,
      "loss": 2.8034,
      "theoretical_loss": 3.756894788929709,
      "tokens_seen": 742750208
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00039138415245737213,
      "loss": 2.7063,
      "theoretical_loss": 3.7568614877683064,
      "tokens_seen": 742815744
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003913741223671013,
      "loss": 2.8524,
      "theoretical_loss": 3.7568281903673864,
      "tokens_seen": 742881280
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003913640922768305,
      "loss": 2.6954,
      "theoretical_loss": 3.756794896726194,
      "tokens_seen": 742946816
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003913540621865597,
      "loss": 2.489,
      "theoretical_loss": 3.7567616068439715,
      "tokens_seen": 743012352
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00039134403209628886,
      "loss": 2.9102,
      "theoretical_loss": 3.756728320719964,
      "tokens_seen": 743077888
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00039133400200601804,
      "loss": 2.6978,
      "theoretical_loss": 3.756695038353416,
      "tokens_seen": 743143424
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00039132397191574727,
      "loss": 2.5701,
      "theoretical_loss": 3.7566617597435723,
      "tokens_seen": 743208960
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003913139418254764,
      "loss": 2.5741,
      "theoretical_loss": 3.7566284848896765,
      "tokens_seen": 743274496
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00039130391173520563,
      "loss": 2.8461,
      "theoretical_loss": 3.756595213790974,
      "tokens_seen": 743340032
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003912938816449348,
      "loss": 2.6907,
      "theoretical_loss": 3.756561946446711,
      "tokens_seen": 743405568
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.000391283851554664,
      "loss": 2.9518,
      "theoretical_loss": 3.756528682856131,
      "tokens_seen": 743471104
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003912738214643932,
      "loss": 2.6059,
      "theoretical_loss": 3.7564954230184817,
      "tokens_seen": 743536640
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00039126379137412236,
      "loss": 2.8419,
      "theoretical_loss": 3.756462166933008,
      "tokens_seen": 743602176
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00039125376128385154,
      "loss": 2.7538,
      "theoretical_loss": 3.7564289145989553,
      "tokens_seen": 743667712
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003912437311935808,
      "loss": 2.7312,
      "theoretical_loss": 3.756395666015571,
      "tokens_seen": 743733248
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003912337011033099,
      "loss": 3.0143,
      "theoretical_loss": 3.756362421182101,
      "tokens_seen": 743798784
    },
    {
      "epoch": 2.05,
      "objective/train/docs_used": 864916,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.2605478763580322,
      "objective/train/theoretical_loss": 3.7563458001713483,
      "objective/train/tokens_used": 764291552,
      "theoretical_loss": 3.7563458001713483,
      "tokens_seen": 743831552
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00039122367101303914,
      "loss": 2.9468,
      "theoretical_loss": 3.756329180097792,
      "tokens_seen": 743864320
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00039121364092276826,
      "loss": 2.9104,
      "theoretical_loss": 3.756295942761891,
      "tokens_seen": 743929856
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003912036108324975,
      "loss": 2.8218,
      "theoretical_loss": 3.7562627091736456,
      "tokens_seen": 743995392
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003911935807422267,
      "loss": 2.7912,
      "theoretical_loss": 3.7562294793323026,
      "tokens_seen": 744060928
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00039118355065195586,
      "loss": 2.9189,
      "theoretical_loss": 3.75619625323711,
      "tokens_seen": 744126464
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00039117352056168504,
      "loss": 2.9026,
      "theoretical_loss": 3.7561630308873153,
      "tokens_seen": 744192000
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003911634904714143,
      "loss": 2.9202,
      "theoretical_loss": 3.756129812282167,
      "tokens_seen": 744257536
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003911534603811434,
      "loss": 2.7454,
      "theoretical_loss": 3.7560965974209126,
      "tokens_seen": 744323072
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00039114343029087264,
      "loss": 3.0462,
      "theoretical_loss": 3.756063386302801,
      "tokens_seen": 744388608
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003911334002006018,
      "loss": 2.9229,
      "theoretical_loss": 3.7560301789270807,
      "tokens_seen": 744454144
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.000391123370110331,
      "loss": 2.6486,
      "theoretical_loss": 3.755996975293001,
      "tokens_seen": 744519680
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00039111334002006024,
      "loss": 2.7592,
      "theoretical_loss": 3.7559637753998105,
      "tokens_seen": 744585216
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00039110330992978936,
      "loss": 2.579,
      "theoretical_loss": 3.755930579246759,
      "tokens_seen": 744650752
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003910932798395186,
      "loss": 2.6565,
      "theoretical_loss": 3.7558973868330954,
      "tokens_seen": 744716288
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00039108324974924773,
      "loss": 2.6632,
      "theoretical_loss": 3.7558641981580703,
      "tokens_seen": 744781824
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00039107321965897696,
      "loss": 2.5718,
      "theoretical_loss": 3.7558310132209334,
      "tokens_seen": 744847360
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00039106318956870614,
      "loss": 2.6674,
      "theoretical_loss": 3.755797832020934,
      "tokens_seen": 744912896
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003910531594784353,
      "loss": 2.8298,
      "theoretical_loss": 3.755764654557324,
      "tokens_seen": 744978432
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003910431293881645,
      "loss": 3.0806,
      "theoretical_loss": 3.755731480829353,
      "tokens_seen": 745043968
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003910330992978937,
      "loss": 2.6907,
      "theoretical_loss": 3.755698310836272,
      "tokens_seen": 745109504
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00039102306920762287,
      "loss": 2.843,
      "theoretical_loss": 3.7556651445773324,
      "tokens_seen": 745175040
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003910130391173521,
      "loss": 2.8829,
      "theoretical_loss": 3.755631982051785,
      "tokens_seen": 745240576
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00039100300902708123,
      "loss": 2.7557,
      "theoretical_loss": 3.7555988232588815,
      "tokens_seen": 745306112
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00039099297893681047,
      "loss": 2.8817,
      "theoretical_loss": 3.7555656681978737,
      "tokens_seen": 745371648
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00039098294884653965,
      "loss": 2.8561,
      "theoretical_loss": 3.755532516868014,
      "tokens_seen": 745437184
    },
    {
      "epoch": 2.05,
      "objective/train/docs_used": 865470,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9077301025390625,
      "objective/train/theoretical_loss": 3.75551594260203,
      "objective/train/tokens_used": 765929952,
      "theoretical_loss": 3.75551594260203,
      "tokens_seen": 745469952
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00039097291875626883,
      "loss": 2.9101,
      "theoretical_loss": 3.755499369268553,
      "tokens_seen": 745502720
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.000390962888665998,
      "loss": 2.9018,
      "theoretical_loss": 3.7554662253987443,
      "tokens_seen": 745568256
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003909528585757272,
      "loss": 2.7832,
      "theoretical_loss": 3.755433085257841,
      "tokens_seen": 745633792
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00039094282848545637,
      "loss": 2.9756,
      "theoretical_loss": 3.755399948845094,
      "tokens_seen": 745699328
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003909327983951856,
      "loss": 2.7584,
      "theoretical_loss": 3.7553668161597584,
      "tokens_seen": 745764864
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00039092276830491473,
      "loss": 2.886,
      "theoretical_loss": 3.7553336872010856,
      "tokens_seen": 745830400
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00039091273821464397,
      "loss": 3.0352,
      "theoretical_loss": 3.7553005619683306,
      "tokens_seen": 745895936
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003909027081243731,
      "loss": 2.733,
      "theoretical_loss": 3.755267440460746,
      "tokens_seen": 745961472
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00039089267803410233,
      "loss": 2.7691,
      "theoretical_loss": 3.7552343226775857,
      "tokens_seen": 746027008
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003908826479438315,
      "loss": 2.886,
      "theoretical_loss": 3.7552012086181046,
      "tokens_seen": 746092544
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003908726178535607,
      "loss": 2.7298,
      "theoretical_loss": 3.755168098281556,
      "tokens_seen": 746158080
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003908625877632899,
      "loss": 2.7824,
      "theoretical_loss": 3.755134991667195,
      "tokens_seen": 746223616
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00039085255767301906,
      "loss": 2.8606,
      "theoretical_loss": 3.755101888774276,
      "tokens_seen": 746289152
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00039084252758274824,
      "loss": 2.481,
      "theoretical_loss": 3.755068789602054,
      "tokens_seen": 746354688
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00039083249749247747,
      "loss": 2.5479,
      "theoretical_loss": 3.755035694149784,
      "tokens_seen": 746420224
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003908224674022066,
      "loss": 2.9372,
      "theoretical_loss": 3.7550026024167225,
      "tokens_seen": 746485760
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00039081243731193583,
      "loss": 2.9928,
      "theoretical_loss": 3.7549695144021236,
      "tokens_seen": 746551296
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.000390802407221665,
      "loss": 2.8547,
      "theoretical_loss": 3.7549364301052437,
      "tokens_seen": 746616832
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003907923771313942,
      "loss": 2.6554,
      "theoretical_loss": 3.7549033495253386,
      "tokens_seen": 746682368
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003907823470411234,
      "loss": 2.6698,
      "theoretical_loss": 3.754870272661665,
      "tokens_seen": 746747904
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00039077231695085256,
      "loss": 2.8375,
      "theoretical_loss": 3.754837199513479,
      "tokens_seen": 746813440
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00039076228686058174,
      "loss": 2.6853,
      "theoretical_loss": 3.7548041300800366,
      "tokens_seen": 746878976
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.000390752256770311,
      "loss": 2.8241,
      "theoretical_loss": 3.7547710643605963,
      "tokens_seen": 746944512
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003907422266800401,
      "loss": 2.5937,
      "theoretical_loss": 3.7547380023544132,
      "tokens_seen": 747010048
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00039073219658976934,
      "loss": 2.6594,
      "theoretical_loss": 3.7547049440607467,
      "tokens_seen": 747075584
    },
    {
      "epoch": 2.05,
      "objective/train/docs_used": 865470,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7921359539031982,
      "objective/train/theoretical_loss": 3.7546884163058745,
      "objective/train/tokens_used": 767568352,
      "theoretical_loss": 3.7546884163058745,
      "tokens_seen": 747108352
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00039072216649949846,
      "loss": 2.9316,
      "theoretical_loss": 3.7546718894788524,
      "tokens_seen": 747141120
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003907121364092277,
      "loss": 2.9271,
      "theoretical_loss": 3.754638838607989,
      "tokens_seen": 747206656
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003907021063189569,
      "loss": 2.662,
      "theoretical_loss": 3.754605791447415,
      "tokens_seen": 747272192
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00039069207622868606,
      "loss": 2.846,
      "theoretical_loss": 3.754572747996387,
      "tokens_seen": 747337728
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00039068204613841524,
      "loss": 2.6688,
      "theoretical_loss": 3.754539708254164,
      "tokens_seen": 747403264
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003906720160481445,
      "loss": 3.0716,
      "theoretical_loss": 3.754506672220005,
      "tokens_seen": 747468800
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003906619859578736,
      "loss": 2.5733,
      "theoretical_loss": 3.7544736398931686,
      "tokens_seen": 747534336
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00039065195586760284,
      "loss": 2.8922,
      "theoretical_loss": 3.754440611272914,
      "tokens_seen": 747599872
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00039064192577733197,
      "loss": 2.7972,
      "theoretical_loss": 3.7544075863585,
      "tokens_seen": 747665408
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003906318956870612,
      "loss": 2.6071,
      "theoretical_loss": 3.7543745651491855,
      "tokens_seen": 747730944
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003906218655967904,
      "loss": 2.8335,
      "theoretical_loss": 3.7543415476442314,
      "tokens_seen": 747796480
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00039061183550651957,
      "loss": 3.1237,
      "theoretical_loss": 3.754308533842897,
      "tokens_seen": 747862016
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00039060180541624875,
      "loss": 2.6559,
      "theoretical_loss": 3.7542755237444423,
      "tokens_seen": 747927552
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00039059177532597793,
      "loss": 2.877,
      "theoretical_loss": 3.754242517348127,
      "tokens_seen": 747993088
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003905817452357071,
      "loss": 2.6894,
      "theoretical_loss": 3.754209514653213,
      "tokens_seen": 748058624
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00039057171514543634,
      "loss": 3.0517,
      "theoretical_loss": 3.7541765156589593,
      "tokens_seen": 748124160
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00039056168505516547,
      "loss": 2.8098,
      "theoretical_loss": 3.754143520364628,
      "tokens_seen": 748189696
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003905516549648947,
      "loss": 2.7217,
      "theoretical_loss": 3.7541105287694796,
      "tokens_seen": 748255232
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00039054162487462383,
      "loss": 2.8167,
      "theoretical_loss": 3.7540775408727765,
      "tokens_seen": 748320768
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00039053159478435307,
      "loss": 2.766,
      "theoretical_loss": 3.7540445566737786,
      "tokens_seen": 748386304
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00039052156469408225,
      "loss": 2.7154,
      "theoretical_loss": 3.754011576171749,
      "tokens_seen": 748451840
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00039051153460381143,
      "loss": 2.7892,
      "theoretical_loss": 3.7539785993659494,
      "tokens_seen": 748517376
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003905015045135406,
      "loss": 2.6884,
      "theoretical_loss": 3.7539456262556414,
      "tokens_seen": 748582912
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00039049147442326985,
      "loss": 3.0133,
      "theoretical_loss": 3.753912656840088,
      "tokens_seen": 748648448
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.000390481444332999,
      "loss": 2.7864,
      "theoretical_loss": 3.7538796911185512,
      "tokens_seen": 748713984
    },
    {
      "epoch": 2.05,
      "objective/train/docs_used": 866234,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.680988311767578,
      "objective/train/theoretical_loss": 3.7538632096428093,
      "objective/train/tokens_used": 769206752,
      "theoretical_loss": 3.7538632096428093,
      "tokens_seen": 748746752
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003904714142427282,
      "loss": 2.649,
      "theoretical_loss": 3.753846729090295,
      "tokens_seen": 748779520
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00039046138415245734,
      "loss": 2.9356,
      "theoretical_loss": 3.753813770754581,
      "tokens_seen": 748845056
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00039045135406218657,
      "loss": 2.8491,
      "theoretical_loss": 3.753780816110673,
      "tokens_seen": 748910592
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00039044132397191575,
      "loss": 2.9241,
      "theoretical_loss": 3.7537478651578358,
      "tokens_seen": 748976128
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00039043129388164493,
      "loss": 2.7688,
      "theoretical_loss": 3.7537149178953304,
      "tokens_seen": 749041664
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003904212637913741,
      "loss": 2.8314,
      "theoretical_loss": 3.7536819743224226,
      "tokens_seen": 749107200
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003904112337011033,
      "loss": 2.9138,
      "theoretical_loss": 3.753649034438376,
      "tokens_seen": 749172736
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003904012036108325,
      "loss": 2.713,
      "theoretical_loss": 3.753616098242455,
      "tokens_seen": 749238272
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003903911735205617,
      "loss": 2.8089,
      "theoretical_loss": 3.753583165733924,
      "tokens_seen": 749303808
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003903811434302909,
      "loss": 2.9077,
      "theoretical_loss": 3.7535502369120475,
      "tokens_seen": 749369344
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003903711133400201,
      "loss": 2.924,
      "theoretical_loss": 3.7535173117760907,
      "tokens_seen": 749434880
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00039036108324974926,
      "loss": 2.7313,
      "theoretical_loss": 3.753484390325319,
      "tokens_seen": 749500416
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00039035105315947844,
      "loss": 2.7019,
      "theoretical_loss": 3.753451472558997,
      "tokens_seen": 749565952
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00039034102306920767,
      "loss": 2.9285,
      "theoretical_loss": 3.7534185584763904,
      "tokens_seen": 749631488
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003903309929789368,
      "loss": 2.5482,
      "theoretical_loss": 3.7533856480767662,
      "tokens_seen": 749697024
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00039032096288866603,
      "loss": 2.7619,
      "theoretical_loss": 3.753352741359389,
      "tokens_seen": 749762560
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003903109327983952,
      "loss": 2.9135,
      "theoretical_loss": 3.7533198383235256,
      "tokens_seen": 749828096
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003903009027081244,
      "loss": 2.7497,
      "theoretical_loss": 3.753286938968442,
      "tokens_seen": 749893632
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003902908726178536,
      "loss": 2.8409,
      "theoretical_loss": 3.7532540432934054,
      "tokens_seen": 749959168
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00039028084252758276,
      "loss": 2.9118,
      "theoretical_loss": 3.7532211512976823,
      "tokens_seen": 750024704
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00039027081243731194,
      "loss": 2.8197,
      "theoretical_loss": 3.7531882629805393,
      "tokens_seen": 750090240
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003902607823470412,
      "loss": 2.7701,
      "theoretical_loss": 3.753155378341245,
      "tokens_seen": 750155776
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003902507522567703,
      "loss": 2.9495,
      "theoretical_loss": 3.7531224973790653,
      "tokens_seen": 750221312
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00039024072216649954,
      "loss": 2.7622,
      "theoretical_loss": 3.7530896200932684,
      "tokens_seen": 750286848
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00039023069207622866,
      "loss": 2.7241,
      "theoretical_loss": 3.7530567464831233,
      "tokens_seen": 750352384
    },
    {
      "epoch": 2.05,
      "objective/train/docs_used": 867529,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8023226261138916,
      "objective/train/theoretical_loss": 3.75304031105619,
      "objective/train/tokens_used": 770845152,
      "theoretical_loss": 3.75304031105619,
      "tokens_seen": 750385152
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003902206619859579,
      "loss": 2.8509,
      "theoretical_loss": 3.7530238765478963,
      "tokens_seen": 750417920
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003902106318956871,
      "loss": 2.8771,
      "theoretical_loss": 3.7529910102868564,
      "tokens_seen": 750483456
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00039020060180541626,
      "loss": 2.7444,
      "theoretical_loss": 3.752958147699272,
      "tokens_seen": 750548992
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00039019057171514544,
      "loss": 2.7908,
      "theoretical_loss": 3.7529252887844127,
      "tokens_seen": 750614528
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003901805416248747,
      "loss": 2.9501,
      "theoretical_loss": 3.7528924335415463,
      "tokens_seen": 750680064
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003901705115346038,
      "loss": 2.8367,
      "theoretical_loss": 3.7528595819699424,
      "tokens_seen": 750745600
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00039016048144433304,
      "loss": 2.6838,
      "theoretical_loss": 3.752826734068871,
      "tokens_seen": 750811136
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00039015045135406217,
      "loss": 2.7132,
      "theoretical_loss": 3.7527938898376,
      "tokens_seen": 750876672
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003901404212637914,
      "loss": 2.7584,
      "theoretical_loss": 3.752761049275401,
      "tokens_seen": 750942208
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003901303911735206,
      "loss": 2.8268,
      "theoretical_loss": 3.7527282123815424,
      "tokens_seen": 751007744
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00039012036108324977,
      "loss": 3.2001,
      "theoretical_loss": 3.7526953791552953,
      "tokens_seen": 751073280
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00039011033099297895,
      "loss": 2.7938,
      "theoretical_loss": 3.7526625495959296,
      "tokens_seen": 751138816
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00039010030090270813,
      "loss": 2.8428,
      "theoretical_loss": 3.752629723702717,
      "tokens_seen": 751204352
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003900902708124373,
      "loss": 2.7589,
      "theoretical_loss": 3.752596901474927,
      "tokens_seen": 751269888
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00039008024072216654,
      "loss": 2.8637,
      "theoretical_loss": 3.7525640829118307,
      "tokens_seen": 751335424
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00039007021063189567,
      "loss": 2.7897,
      "theoretical_loss": 3.7525312680127003,
      "tokens_seen": 751400960
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003900601805416249,
      "loss": 2.9186,
      "theoretical_loss": 3.7524984567768063,
      "tokens_seen": 751466496
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00039005015045135403,
      "loss": 2.819,
      "theoretical_loss": 3.7524656492034207,
      "tokens_seen": 751532032
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00039004012036108327,
      "loss": 2.9226,
      "theoretical_loss": 3.752432845291816,
      "tokens_seen": 751597568
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00039003009027081245,
      "loss": 2.8002,
      "theoretical_loss": 3.752400045041263,
      "tokens_seen": 751663104
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00039002006018054163,
      "loss": 2.6275,
      "theoretical_loss": 3.7523672484510344,
      "tokens_seen": 751728640
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003900100300902708,
      "loss": 3.0589,
      "theoretical_loss": 3.752334455520403,
      "tokens_seen": 751794176
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00039000000000000005,
      "loss": 3.0157,
      "theoretical_loss": 3.7523016662486413,
      "tokens_seen": 751859712
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003899899699097292,
      "loss": 2.6684,
      "theoretical_loss": 3.752268880635022,
      "tokens_seen": 751925248
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003899799398194584,
      "loss": 3.0901,
      "theoretical_loss": 3.752236098678819,
      "tokens_seen": 751990784
    },
    {
      "epoch": 2.05,
      "objective/train/docs_used": 868214,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.602017402648926,
      "objective/train/theoretical_loss": 3.752219709072021,
      "objective/train/tokens_used": 772483552,
      "theoretical_loss": 3.752219709072021,
      "tokens_seen": 752023552
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038996990972918754,
      "loss": 2.9366,
      "theoretical_loss": 3.7522033203793046,
      "tokens_seen": 752056320
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038995987963891677,
      "loss": 2.9654,
      "theoretical_loss": 3.752170545735753,
      "tokens_seen": 752121856
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038994984954864595,
      "loss": 2.8639,
      "theoretical_loss": 3.752137774747437,
      "tokens_seen": 752187392
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038993981945837513,
      "loss": 2.8474,
      "theoretical_loss": 3.7521050074136317,
      "tokens_seen": 752252928
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003899297893681043,
      "loss": 2.9317,
      "theoretical_loss": 3.7520722437336107,
      "tokens_seen": 752318464
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003899197592778335,
      "loss": 2.9675,
      "theoretical_loss": 3.752039483706648,
      "tokens_seen": 752384000
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003899097291875627,
      "loss": 3.0981,
      "theoretical_loss": 3.752006727332019,
      "tokens_seen": 752449536
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003898996990972919,
      "loss": 2.8965,
      "theoretical_loss": 3.7519739746089975,
      "tokens_seen": 752515072
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038988966900702104,
      "loss": 2.7717,
      "theoretical_loss": 3.751941225536859,
      "tokens_seen": 752580608
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003898796389167503,
      "loss": 2.8899,
      "theoretical_loss": 3.7519084801148788,
      "tokens_seen": 752646144
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003898696088264794,
      "loss": 2.9638,
      "theoretical_loss": 3.751875738342332,
      "tokens_seen": 752711680
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038985957873620864,
      "loss": 2.8868,
      "theoretical_loss": 3.7518430002184937,
      "tokens_seen": 752777216
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003898495486459378,
      "loss": 2.9566,
      "theoretical_loss": 3.751810265742641,
      "tokens_seen": 752842752
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.000389839518555667,
      "loss": 2.8835,
      "theoretical_loss": 3.7517775349140488,
      "tokens_seen": 752908288
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003898294884653962,
      "loss": 2.7896,
      "theoretical_loss": 3.7517448077319937,
      "tokens_seen": 752973824
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003898194583751254,
      "loss": 2.6797,
      "theoretical_loss": 3.751712084195752,
      "tokens_seen": 753039360
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038980942828485454,
      "loss": 2.8957,
      "theoretical_loss": 3.7516793643046005,
      "tokens_seen": 753104896
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003897993981945838,
      "loss": 2.8789,
      "theoretical_loss": 3.7516466480578154,
      "tokens_seen": 753170432
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003897893681043129,
      "loss": 2.9751,
      "theoretical_loss": 3.751613935454675,
      "tokens_seen": 753235968
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038977933801404214,
      "loss": 3.1071,
      "theoretical_loss": 3.7515812264944555,
      "tokens_seen": 753301504
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003897693079237713,
      "loss": 2.8043,
      "theoretical_loss": 3.751548521176434,
      "tokens_seen": 753367040
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003897592778335005,
      "loss": 2.8944,
      "theoretical_loss": 3.7515158194998897,
      "tokens_seen": 753432576
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003897492477432297,
      "loss": 2.8687,
      "theoretical_loss": 3.7514831214640987,
      "tokens_seen": 753498112
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038973921765295886,
      "loss": 2.8928,
      "theoretical_loss": 3.751450427068341,
      "tokens_seen": 753563648
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038972918756268805,
      "loss": 3.0386,
      "theoretical_loss": 3.7514177363118923,
      "tokens_seen": 753629184
    },
    {
      "epoch": 2.05,
      "objective/train/docs_used": 869556,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.754945755004883,
      "objective/train/theoretical_loss": 3.7514013922981846,
      "objective/train/tokens_used": 774121952,
      "theoretical_loss": 3.7514013922981846,
      "tokens_seen": 753661952
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003897191574724173,
      "loss": 2.9253,
      "theoretical_loss": 3.7513850491940337,
      "tokens_seen": 753694720
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003897091273821464,
      "loss": 2.8255,
      "theoretical_loss": 3.751352365714042,
      "tokens_seen": 753760256
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038969909729187564,
      "loss": 2.8426,
      "theoretical_loss": 3.7513196858711972,
      "tokens_seen": 753825792
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038968906720160477,
      "loss": 2.9487,
      "theoretical_loss": 3.751287009664778,
      "tokens_seen": 753891328
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.000389679037111334,
      "loss": 2.7612,
      "theoretical_loss": 3.751254337094063,
      "tokens_seen": 753956864
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003896690070210632,
      "loss": 3.0445,
      "theoretical_loss": 3.7512216681583332,
      "tokens_seen": 754022400
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038965897693079237,
      "loss": 2.9938,
      "theoretical_loss": 3.751189002856867,
      "tokens_seen": 754087936
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038964894684052155,
      "loss": 2.8911,
      "theoretical_loss": 3.751156341188945,
      "tokens_seen": 754153472
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003896389167502508,
      "loss": 2.8629,
      "theoretical_loss": 3.751123683153847,
      "tokens_seen": 754219008
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038962888665997997,
      "loss": 2.8401,
      "theoretical_loss": 3.7510910287508534,
      "tokens_seen": 754284544
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038961885656970915,
      "loss": 2.8568,
      "theoretical_loss": 3.751058377979245,
      "tokens_seen": 754350080
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038960882647943833,
      "loss": 2.8119,
      "theoretical_loss": 3.751025730838302,
      "tokens_seen": 754415616
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003895987963891675,
      "loss": 2.9211,
      "theoretical_loss": 3.7509930873273056,
      "tokens_seen": 754481152
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038958876629889674,
      "loss": 3.0088,
      "theoretical_loss": 3.7509604474455376,
      "tokens_seen": 754546688
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038957873620862587,
      "loss": 2.8507,
      "theoretical_loss": 3.750927811192278,
      "tokens_seen": 754612224
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003895687061183551,
      "loss": 2.8428,
      "theoretical_loss": 3.750895178566809,
      "tokens_seen": 754677760
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038955867602808423,
      "loss": 2.857,
      "theoretical_loss": 3.7508625495684136,
      "tokens_seen": 754743296
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038954864593781347,
      "loss": 2.9037,
      "theoretical_loss": 3.7508299241963714,
      "tokens_seen": 754808832
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038953861584754265,
      "loss": 2.688,
      "theoretical_loss": 3.7507973024499663,
      "tokens_seen": 754874368
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038952858575727183,
      "loss": 2.856,
      "theoretical_loss": 3.7507646843284803,
      "tokens_seen": 754939904
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.000389518555667001,
      "loss": 2.7452,
      "theoretical_loss": 3.7507320698311952,
      "tokens_seen": 755005440
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038950852557673025,
      "loss": 2.6562,
      "theoretical_loss": 3.7506994589573948,
      "tokens_seen": 755070976
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003894984954864594,
      "loss": 2.8558,
      "theoretical_loss": 3.7506668517063613,
      "tokens_seen": 755136512
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003894884653961886,
      "loss": 3.0562,
      "theoretical_loss": 3.750634248077379,
      "tokens_seen": 755202048
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038947843530591774,
      "loss": 2.8171,
      "theoretical_loss": 3.75060164806973,
      "tokens_seen": 755267584
    },
    {
      "epoch": 2.05,
      "objective/train/docs_used": 870017,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.080669403076172,
      "objective/train/theoretical_loss": 3.750585349423682,
      "objective/train/tokens_used": 775760352,
      "theoretical_loss": 3.750585349423682,
      "tokens_seen": 755300352
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038946840521564697,
      "loss": 2.8293,
      "theoretical_loss": 3.7505690516826986,
      "tokens_seen": 755333120
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038945837512537615,
      "loss": 2.8135,
      "theoretical_loss": 3.7505364589155685,
      "tokens_seen": 755398656
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038944834503510533,
      "loss": 2.8743,
      "theoretical_loss": 3.7505038697676234,
      "tokens_seen": 755464192
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003894383149448345,
      "loss": 2.8677,
      "theoretical_loss": 3.750471284238148,
      "tokens_seen": 755529728
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003894282848545637,
      "loss": 2.7691,
      "theoretical_loss": 3.7504387023264263,
      "tokens_seen": 755595264
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003894182547642929,
      "loss": 2.9833,
      "theoretical_loss": 3.750406124031743,
      "tokens_seen": 755660800
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003894082246740221,
      "loss": 2.9658,
      "theoretical_loss": 3.750373549353383,
      "tokens_seen": 755726336
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038939819458375124,
      "loss": 2.7576,
      "theoretical_loss": 3.750340978290632,
      "tokens_seen": 755791872
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003893881644934805,
      "loss": 2.7332,
      "theoretical_loss": 3.7503084108427736,
      "tokens_seen": 755857408
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003893781344032096,
      "loss": 2.7703,
      "theoretical_loss": 3.750275847009095,
      "tokens_seen": 755922944
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038936810431293884,
      "loss": 2.7189,
      "theoretical_loss": 3.7502432867888804,
      "tokens_seen": 755988480
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.000389358074222668,
      "loss": 2.7728,
      "theoretical_loss": 3.7502107301814167,
      "tokens_seen": 756054016
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003893480441323972,
      "loss": 2.6577,
      "theoretical_loss": 3.750178177185989,
      "tokens_seen": 756119552
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003893380140421264,
      "loss": 2.7496,
      "theoretical_loss": 3.750145627801884,
      "tokens_seen": 756185088
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003893279839518556,
      "loss": 2.9146,
      "theoretical_loss": 3.7501130820283883,
      "tokens_seen": 756250624
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038931795386158474,
      "loss": 2.7928,
      "theoretical_loss": 3.750080539864789,
      "tokens_seen": 756316160
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.000389307923771314,
      "loss": 2.8623,
      "theoretical_loss": 3.750048001310372,
      "tokens_seen": 756381696
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003892978936810431,
      "loss": 2.5961,
      "theoretical_loss": 3.750015466364424,
      "tokens_seen": 756447232
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038928786359077234,
      "loss": 2.6931,
      "theoretical_loss": 3.7499829350262335,
      "tokens_seen": 756512768
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003892778335005015,
      "loss": 2.8814,
      "theoretical_loss": 3.7499504072950876,
      "tokens_seen": 756578304
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003892678034102307,
      "loss": 2.7536,
      "theoretical_loss": 3.7499178831702737,
      "tokens_seen": 756643840
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003892577733199599,
      "loss": 2.8347,
      "theoretical_loss": 3.7498853626510797,
      "tokens_seen": 756709376
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038924774322968906,
      "loss": 2.7921,
      "theoretical_loss": 3.7498528457367932,
      "tokens_seen": 756774912
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038923771313941825,
      "loss": 2.9477,
      "theoretical_loss": 3.7498203324267037,
      "tokens_seen": 756840448
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003892276830491475,
      "loss": 3.0231,
      "theoretical_loss": 3.7497878227200983,
      "tokens_seen": 756905984
    },
    {
      "epoch": 2.05,
      "objective/train/docs_used": 871516,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.127725601196289,
      "objective/train/theoretical_loss": 3.74977156921788,
      "objective/train/tokens_used": 777398752,
      "theoretical_loss": 3.74977156921788,
      "tokens_seen": 756938752
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003892176529588766,
      "loss": 2.9308,
      "theoretical_loss": 3.749755316616267,
      "tokens_seen": 756971520
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038920762286860584,
      "loss": 2.8961,
      "theoretical_loss": 3.749722814114497,
      "tokens_seen": 757037056
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038919759277833497,
      "loss": 2.7532,
      "theoretical_loss": 3.749690315214079,
      "tokens_seen": 757102592
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003891875626880642,
      "loss": 3.0619,
      "theoretical_loss": 3.7496578199143014,
      "tokens_seen": 757168128
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003891775325977934,
      "loss": 2.6638,
      "theoretical_loss": 3.749625328214454,
      "tokens_seen": 757233664
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038916750250752257,
      "loss": 2.9055,
      "theoretical_loss": 3.7495928401138263,
      "tokens_seen": 757299200
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038915747241725175,
      "loss": 2.7871,
      "theoretical_loss": 3.749560355611708,
      "tokens_seen": 757364736
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.000389147442326981,
      "loss": 2.674,
      "theoretical_loss": 3.74952787470739,
      "tokens_seen": 757430272
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003891374122367101,
      "loss": 2.8836,
      "theoretical_loss": 3.7494953974001617,
      "tokens_seen": 757495808
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038912738214643935,
      "loss": 2.6771,
      "theoretical_loss": 3.7494629236893138,
      "tokens_seen": 757561344
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003891173520561685,
      "loss": 2.981,
      "theoretical_loss": 3.7494304535741376,
      "tokens_seen": 757626880
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003891073219658977,
      "loss": 3.0572,
      "theoretical_loss": 3.7493979870539227,
      "tokens_seen": 757692416
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003890972918756269,
      "loss": 3.0317,
      "theoretical_loss": 3.7493655241279615,
      "tokens_seen": 757757952
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038908726178535607,
      "loss": 2.9948,
      "theoretical_loss": 3.7493330647955445,
      "tokens_seen": 757823488
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038907723169508525,
      "loss": 2.7901,
      "theoretical_loss": 3.7493006090559637,
      "tokens_seen": 757889024
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038906720160481443,
      "loss": 2.6746,
      "theoretical_loss": 3.74926815690851,
      "tokens_seen": 757954560
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003890571715145436,
      "loss": 3.0293,
      "theoretical_loss": 3.7492357083524768,
      "tokens_seen": 758020096
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038904714142427285,
      "loss": 2.7772,
      "theoretical_loss": 3.7492032633871544,
      "tokens_seen": 758085632
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.000389037111334002,
      "loss": 2.8311,
      "theoretical_loss": 3.749170822011836,
      "tokens_seen": 758151168
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003890270812437312,
      "loss": 2.8405,
      "theoretical_loss": 3.749138384225814,
      "tokens_seen": 758216704
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003890170511534604,
      "loss": 2.9285,
      "theoretical_loss": 3.7491059500283814,
      "tokens_seen": 758282240
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003890070210631896,
      "loss": 2.8327,
      "theoretical_loss": 3.7490735194188307,
      "tokens_seen": 758347776
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038899699097291876,
      "loss": 2.7235,
      "theoretical_loss": 3.749041092396455,
      "tokens_seen": 758413312
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038898696088264794,
      "loss": 2.8355,
      "theoretical_loss": 3.7490086689605473,
      "tokens_seen": 758478848
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003889769307923771,
      "loss": 2.8129,
      "theoretical_loss": 3.748976249110402,
      "tokens_seen": 758544384
    },
    {
      "epoch": 2.05,
      "objective/train/docs_used": 872211,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.623302936553955,
      "objective/train/theoretical_loss": 3.748960040529769,
      "objective/train/tokens_used": 779037152,
      "theoretical_loss": 3.748960040529769,
      "tokens_seen": 758577152
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038896690070210635,
      "loss": 2.7462,
      "theoretical_loss": 3.7489438328453124,
      "tokens_seen": 758609920
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003889568706118355,
      "loss": 2.9954,
      "theoretical_loss": 3.748911420164572,
      "tokens_seen": 758675456
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003889468405215647,
      "loss": 2.8864,
      "theoretical_loss": 3.7488790110674755,
      "tokens_seen": 758740992
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038893681043129384,
      "loss": 3.003,
      "theoretical_loss": 3.7488466055533163,
      "tokens_seen": 758806528
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003889267803410231,
      "loss": 2.9352,
      "theoretical_loss": 3.74881420362139,
      "tokens_seen": 758872064
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038891675025075226,
      "loss": 2.7863,
      "theoretical_loss": 3.7487818052709905,
      "tokens_seen": 758937600
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038890672016048144,
      "loss": 2.9018,
      "theoretical_loss": 3.7487494105014125,
      "tokens_seen": 759003136
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003888966900702106,
      "loss": 2.5538,
      "theoretical_loss": 3.748717019311952,
      "tokens_seen": 759068672
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003888866599799398,
      "loss": 2.7822,
      "theoretical_loss": 3.748684631701904,
      "tokens_seen": 759134208
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038887662988966904,
      "loss": 2.7246,
      "theoretical_loss": 3.748652247670564,
      "tokens_seen": 759199744
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003888665997993982,
      "loss": 2.9017,
      "theoretical_loss": 3.748619867217227,
      "tokens_seen": 759265280
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003888565697091274,
      "loss": 2.9178,
      "theoretical_loss": 3.7485874903411895,
      "tokens_seen": 759330816
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003888465396188566,
      "loss": 2.8642,
      "theoretical_loss": 3.7485551170417484,
      "tokens_seen": 759396352
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003888365095285858,
      "loss": 2.8042,
      "theoretical_loss": 3.748522747318198,
      "tokens_seen": 759461888
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038882647943831494,
      "loss": 2.9502,
      "theoretical_loss": 3.7484903811698365,
      "tokens_seen": 759527424
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003888164493480442,
      "loss": 2.9731,
      "theoretical_loss": 3.7484580185959597,
      "tokens_seen": 759592960
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003888064192577733,
      "loss": 2.9822,
      "theoretical_loss": 3.748425659595865,
      "tokens_seen": 759658496
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038879638916750254,
      "loss": 3.0593,
      "theoretical_loss": 3.7483933041688493,
      "tokens_seen": 759724032
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003887863590772317,
      "loss": 2.6031,
      "theoretical_loss": 3.74836095231421,
      "tokens_seen": 759789568
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003887763289869609,
      "loss": 2.9151,
      "theoretical_loss": 3.7483286040312445,
      "tokens_seen": 759855104
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003887662988966901,
      "loss": 2.9807,
      "theoretical_loss": 3.7482962593192504,
      "tokens_seen": 759920640
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038875626880641927,
      "loss": 2.795,
      "theoretical_loss": 3.7482639181775252,
      "tokens_seen": 759986176
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038874623871614845,
      "loss": 2.9018,
      "theoretical_loss": 3.748231580605368,
      "tokens_seen": 760051712
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003887362086258777,
      "loss": 2.9098,
      "theoretical_loss": 3.748199246602076,
      "tokens_seen": 760117248
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003887261785356068,
      "loss": 2.8907,
      "theoretical_loss": 3.7481669161669484,
      "tokens_seen": 760182784
    },
    {
      "epoch": 2.05,
      "objective/train/docs_used": 873451,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.31072998046875,
      "objective/train/theoretical_loss": 3.748150752287227,
      "objective/train/tokens_used": 780675552,
      "theoretical_loss": 3.748150752287227,
      "tokens_seen": 760215552
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038871614844533604,
      "loss": 2.6467,
      "theoretical_loss": 3.7481345892992834,
      "tokens_seen": 760248320
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038870611835506517,
      "loss": 2.7994,
      "theoretical_loss": 3.7481022659983805,
      "tokens_seen": 760313856
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003886960882647944,
      "loss": 2.735,
      "theoretical_loss": 3.7480699462635383,
      "tokens_seen": 760379392
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003886860581745236,
      "loss": 2.9806,
      "theoretical_loss": 3.748037630094056,
      "tokens_seen": 760444928
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038867602808425277,
      "loss": 2.746,
      "theoretical_loss": 3.748005317489233,
      "tokens_seen": 760510464
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038866599799398195,
      "loss": 2.6938,
      "theoretical_loss": 3.74797300844837,
      "tokens_seen": 760576000
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003886559679037112,
      "loss": 2.9767,
      "theoretical_loss": 3.747940702970766,
      "tokens_seen": 760641536
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003886459378134403,
      "loss": 2.9013,
      "theoretical_loss": 3.7479084010557204,
      "tokens_seen": 760707072
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038863590772316955,
      "loss": 2.8877,
      "theoretical_loss": 3.7478761027025347,
      "tokens_seen": 760772608
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003886258776328987,
      "loss": 2.7881,
      "theoretical_loss": 3.7478438079105088,
      "tokens_seen": 760838144
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003886158475426279,
      "loss": 2.8087,
      "theoretical_loss": 3.747811516678943,
      "tokens_seen": 760903680
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003886058174523571,
      "loss": 3.0713,
      "theoretical_loss": 3.747779229007139,
      "tokens_seen": 760969216
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038859578736208627,
      "loss": 2.8158,
      "theoretical_loss": 3.7477469448943976,
      "tokens_seen": 761034752
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038858575727181545,
      "loss": 2.8475,
      "theoretical_loss": 3.7477146643400197,
      "tokens_seen": 761100288
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038857572718154463,
      "loss": 2.8359,
      "theoretical_loss": 3.747682387343307,
      "tokens_seen": 761165824
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003885656970912738,
      "loss": 2.5743,
      "theoretical_loss": 3.7476501139035605,
      "tokens_seen": 761231360
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038855566700100305,
      "loss": 2.9382,
      "theoretical_loss": 3.747617844020083,
      "tokens_seen": 761296896
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003885456369107322,
      "loss": 2.8919,
      "theoretical_loss": 3.7475855776921763,
      "tokens_seen": 761362432
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003885356068204614,
      "loss": 2.9157,
      "theoretical_loss": 3.7475533149191422,
      "tokens_seen": 761427968
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003885255767301906,
      "loss": 2.7488,
      "theoretical_loss": 3.7475210557002834,
      "tokens_seen": 761493504
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003885155466399198,
      "loss": 2.7937,
      "theoretical_loss": 3.747488800034903,
      "tokens_seen": 761559040
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038850551654964896,
      "loss": 2.6836,
      "theoretical_loss": 3.747456547922303,
      "tokens_seen": 761624576
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038849548645937814,
      "loss": 2.8512,
      "theoretical_loss": 3.747424299361787,
      "tokens_seen": 761690112
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003884854563691073,
      "loss": 2.7041,
      "theoretical_loss": 3.747392054352658,
      "tokens_seen": 761755648
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038847542627883655,
      "loss": 2.8625,
      "theoretical_loss": 3.7473598128942194,
      "tokens_seen": 761821184
    },
    {
      "epoch": 2.05,
      "objective/train/docs_used": 874119,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.929797649383545,
      "objective/train/theoretical_loss": 3.7473436934962914,
      "objective/train/tokens_used": 782313952,
      "theoretical_loss": 3.7473436934962914,
      "tokens_seen": 761853952
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003884653961885657,
      "loss": 2.8039,
      "theoretical_loss": 3.747327574985775,
      "tokens_seen": 761886720
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003884553660982949,
      "loss": 2.8516,
      "theoretical_loss": 3.7472953406266276,
      "tokens_seen": 761952256
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038844533600802404,
      "loss": 2.7222,
      "theoretical_loss": 3.747263109816083,
      "tokens_seen": 762017792
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003884353059177533,
      "loss": 2.9182,
      "theoretical_loss": 3.747230882553444,
      "tokens_seen": 762083328
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038842527582748246,
      "loss": 2.6432,
      "theoretical_loss": 3.747198658838016,
      "tokens_seen": 762148864
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038841524573721164,
      "loss": 2.661,
      "theoretical_loss": 3.747166438669103,
      "tokens_seen": 762214400
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003884052156469408,
      "loss": 2.8589,
      "theoretical_loss": 3.747134222046009,
      "tokens_seen": 762279936
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038839518555667,
      "loss": 2.9799,
      "theoretical_loss": 3.7471020089680405,
      "tokens_seen": 762345472
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003883851554663992,
      "loss": 2.8731,
      "theoretical_loss": 3.7470697994345024,
      "tokens_seen": 762411008
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003883751253761284,
      "loss": 2.8761,
      "theoretical_loss": 3.7470375934446984,
      "tokens_seen": 762476544
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038836509528585755,
      "loss": 3.1062,
      "theoretical_loss": 3.7470053909979364,
      "tokens_seen": 762542080
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003883550651955868,
      "loss": 2.8829,
      "theoretical_loss": 3.746973192093521,
      "tokens_seen": 762607616
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038834503510531596,
      "loss": 2.6719,
      "theoretical_loss": 3.746940996730758,
      "tokens_seen": 762673152
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038833500501504514,
      "loss": 2.8273,
      "theoretical_loss": 3.7469088049089545,
      "tokens_seen": 762738688
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003883249749247743,
      "loss": 2.8338,
      "theoretical_loss": 3.7468766166274157,
      "tokens_seen": 762804224
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003883149448345035,
      "loss": 2.8078,
      "theoretical_loss": 3.7468444318854486,
      "tokens_seen": 762869760
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003883049147442327,
      "loss": 2.7179,
      "theoretical_loss": 3.7468122506823605,
      "tokens_seen": 762935296
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003882948846539619,
      "loss": 2.871,
      "theoretical_loss": 3.746780073017457,
      "tokens_seen": 763000832
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038828485456369105,
      "loss": 2.8244,
      "theoretical_loss": 3.746747898890047,
      "tokens_seen": 763066368
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003882748244734203,
      "loss": 2.6443,
      "theoretical_loss": 3.7467157282994368,
      "tokens_seen": 763131904
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003882647943831494,
      "loss": 2.8023,
      "theoretical_loss": 3.7466835612449336,
      "tokens_seen": 763197440
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038825476429287865,
      "loss": 2.7588,
      "theoretical_loss": 3.746651397725846,
      "tokens_seen": 763262976
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038824473420260783,
      "loss": 2.9649,
      "theoretical_loss": 3.7466192377414815,
      "tokens_seen": 763328512
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.000388234704112337,
      "loss": 3.0581,
      "theoretical_loss": 3.746587081291148,
      "tokens_seen": 763394048
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003882246740220662,
      "loss": 2.9048,
      "theoretical_loss": 3.746554928374154,
      "tokens_seen": 763459584
    },
    {
      "epoch": 2.05,
      "objective/train/docs_used": 875620,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.904193878173828,
      "objective/train/theoretical_loss": 3.7465388532404433,
      "objective/train/tokens_used": 783952352,
      "theoretical_loss": 3.7465388532404433,
      "tokens_seen": 763492352
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038821464393179537,
      "loss": 2.9749,
      "theoretical_loss": 3.7465227789898083,
      "tokens_seen": 763525120
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038820461384152455,
      "loss": 2.8214,
      "theoretical_loss": 3.7464906331374195,
      "tokens_seen": 763590656
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003881945837512538,
      "loss": 2.7125,
      "theoretical_loss": 3.7464584908162957,
      "tokens_seen": 763656192
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003881845536609829,
      "loss": 2.9804,
      "theoretical_loss": 3.746426352025747,
      "tokens_seen": 763721728
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038817452357071215,
      "loss": 2.7881,
      "theoretical_loss": 3.7463942167650823,
      "tokens_seen": 763787264
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038816449348044133,
      "loss": 2.871,
      "theoretical_loss": 3.746362085033611,
      "tokens_seen": 763852800
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003881544633901705,
      "loss": 3.2335,
      "theoretical_loss": 3.746329956830643,
      "tokens_seen": 763918336
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003881444332998997,
      "loss": 2.8939,
      "theoretical_loss": 3.746297832155488,
      "tokens_seen": 763983872
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003881344032096289,
      "loss": 2.8657,
      "theoretical_loss": 3.7462657110074558,
      "tokens_seen": 764049408
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003881243731193581,
      "loss": 2.9157,
      "theoretical_loss": 3.746233593385857,
      "tokens_seen": 764114944
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003881143430290873,
      "loss": 2.9297,
      "theoretical_loss": 3.7462014792900016,
      "tokens_seen": 764180480
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038810431293881647,
      "loss": 2.9193,
      "theoretical_loss": 3.746169368719201,
      "tokens_seen": 764246016
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038809428284854565,
      "loss": 2.6569,
      "theoretical_loss": 3.7461372616727657,
      "tokens_seen": 764311552
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038808425275827483,
      "loss": 2.8671,
      "theoretical_loss": 3.746105158150007,
      "tokens_seen": 764377088
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.000388074222668004,
      "loss": 2.8244,
      "theoretical_loss": 3.746073058150235,
      "tokens_seen": 764442624
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038806419257773325,
      "loss": 2.7417,
      "theoretical_loss": 3.7460409616727626,
      "tokens_seen": 764508160
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003880541624874624,
      "loss": 2.8929,
      "theoretical_loss": 3.7460088687169,
      "tokens_seen": 764573696
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003880441323971916,
      "loss": 2.9315,
      "theoretical_loss": 3.7459767792819605,
      "tokens_seen": 764639232
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003880341023069208,
      "loss": 2.8129,
      "theoretical_loss": 3.745944693367255,
      "tokens_seen": 764704768
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038802407221665,
      "loss": 2.9542,
      "theoretical_loss": 3.7459126109720957,
      "tokens_seen": 764770304
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038801404212637916,
      "loss": 2.9054,
      "theoretical_loss": 3.7458805320957955,
      "tokens_seen": 764835840
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038800401203610834,
      "loss": 2.7932,
      "theoretical_loss": 3.745848456737667,
      "tokens_seen": 764901376
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003879939819458375,
      "loss": 2.8687,
      "theoretical_loss": 3.7458163848970223,
      "tokens_seen": 764966912
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038798395185556675,
      "loss": 2.8049,
      "theoretical_loss": 3.7457843165731752,
      "tokens_seen": 765032448
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003879739217652959,
      "loss": 2.9075,
      "theoretical_loss": 3.7457522517654382,
      "tokens_seen": 765097984
    },
    {
      "epoch": 2.05,
      "objective/train/docs_used": 876301,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8112714290618896,
      "objective/train/theoretical_loss": 3.7457362206798965,
      "objective/train/tokens_used": 785590752,
      "theoretical_loss": 3.7457362206798965,
      "tokens_seen": 765130752
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003879638916750251,
      "loss": 3.0486,
      "theoretical_loss": 3.7457201904731248,
      "tokens_seen": 765163520
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038795386158475424,
      "loss": 2.8881,
      "theoretical_loss": 3.745688132695549,
      "tokens_seen": 765229056
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003879438314944835,
      "loss": 3.0385,
      "theoretical_loss": 3.745656078432024,
      "tokens_seen": 765294592
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038793380140421266,
      "loss": 2.8354,
      "theoretical_loss": 3.7456240276818633,
      "tokens_seen": 765360128
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038792377131394184,
      "loss": 2.731,
      "theoretical_loss": 3.7455919804443827,
      "tokens_seen": 765425664
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.000387913741223671,
      "loss": 2.9255,
      "theoretical_loss": 3.7455599367188945,
      "tokens_seen": 765491200
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003879037111334002,
      "loss": 2.842,
      "theoretical_loss": 3.745527896504715,
      "tokens_seen": 765556736
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003878936810431294,
      "loss": 3.034,
      "theoretical_loss": 3.7454958598011574,
      "tokens_seen": 765622272
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003878836509528586,
      "loss": 2.7549,
      "theoretical_loss": 3.745463826607537,
      "tokens_seen": 765687808
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038787362086258775,
      "loss": 2.9639,
      "theoretical_loss": 3.7454317969231696,
      "tokens_seen": 765753344
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.000387863590772317,
      "loss": 2.8361,
      "theoretical_loss": 3.74539977074737,
      "tokens_seen": 765818880
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038785356068204616,
      "loss": 2.8789,
      "theoretical_loss": 3.7453677480794534,
      "tokens_seen": 765884416
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038784353059177534,
      "loss": 2.9602,
      "theoretical_loss": 3.745335728918736,
      "tokens_seen": 765949952
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003878335005015045,
      "loss": 2.8919,
      "theoretical_loss": 3.745303713264533,
      "tokens_seen": 766015488
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003878234704112337,
      "loss": 2.8526,
      "theoretical_loss": 3.745271701116161,
      "tokens_seen": 766081024
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003878134403209629,
      "loss": 3.0369,
      "theoretical_loss": 3.7452396924729356,
      "tokens_seen": 766146560
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003878034102306921,
      "loss": 2.8223,
      "theoretical_loss": 3.745207687334174,
      "tokens_seen": 766212096
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038779338014042125,
      "loss": 2.8681,
      "theoretical_loss": 3.745175685699193,
      "tokens_seen": 766277632
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003877833500501505,
      "loss": 2.7303,
      "theoretical_loss": 3.745143687567308,
      "tokens_seen": 766343168
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003877733199598796,
      "loss": 2.9516,
      "theoretical_loss": 3.7451116929378374,
      "tokens_seen": 766408704
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038776328986960885,
      "loss": 2.7378,
      "theoretical_loss": 3.7450797018100976,
      "tokens_seen": 766474240
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038775325977933803,
      "loss": 2.6424,
      "theoretical_loss": 3.745047714183406,
      "tokens_seen": 766539776
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003877432296890672,
      "loss": 2.5332,
      "theoretical_loss": 3.7450157300570814,
      "tokens_seen": 766605312
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003877331995987964,
      "loss": 2.8007,
      "theoretical_loss": 3.74498374943044,
      "tokens_seen": 766670848
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038772316950852557,
      "loss": 2.8851,
      "theoretical_loss": 3.7449517723028007,
      "tokens_seen": 766736384
    },
    {
      "epoch": 2.05,
      "objective/train/docs_used": 877352,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.128386974334717,
      "objective/train/theoretical_loss": 3.7449357850508935,
      "objective/train/tokens_used": 787229152,
      "theoretical_loss": 3.7449357850508935,
      "tokens_seen": 766769152
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038771313941825475,
      "loss": 2.7447,
      "theoretical_loss": 3.7449197986734806,
      "tokens_seen": 766801920
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.000387703109327984,
      "loss": 2.6875,
      "theoretical_loss": 3.744887828541799,
      "tokens_seen": 766867456
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003876930792377131,
      "loss": 2.916,
      "theoretical_loss": 3.744855861907075,
      "tokens_seen": 766932992
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038768304914744235,
      "loss": 2.8973,
      "theoretical_loss": 3.744823898768626,
      "tokens_seen": 766998528
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038767301905717153,
      "loss": 2.7641,
      "theoretical_loss": 3.744791939125771,
      "tokens_seen": 767064064
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003876629889669007,
      "loss": 2.8003,
      "theoretical_loss": 3.74475998297783,
      "tokens_seen": 767129600
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003876529588766299,
      "loss": 2.837,
      "theoretical_loss": 3.7447280303241213,
      "tokens_seen": 767195136
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003876429287863591,
      "loss": 2.6501,
      "theoretical_loss": 3.7446960811639656,
      "tokens_seen": 767260672
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038763289869608826,
      "loss": 3.0486,
      "theoretical_loss": 3.7446641354966816,
      "tokens_seen": 767326208
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003876228686058175,
      "loss": 2.7624,
      "theoretical_loss": 3.744632193321589,
      "tokens_seen": 767391744
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003876128385155466,
      "loss": 2.5624,
      "theoretical_loss": 3.7446002546380086,
      "tokens_seen": 767457280
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038760280842527585,
      "loss": 2.7834,
      "theoretical_loss": 3.7445683194452606,
      "tokens_seen": 767522816
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.000387592778335005,
      "loss": 2.562,
      "theoretical_loss": 3.744536387742665,
      "tokens_seen": 767588352
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003875827482447342,
      "loss": 2.966,
      "theoretical_loss": 3.7445044595295425,
      "tokens_seen": 767653888
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003875727181544634,
      "loss": 2.8074,
      "theoretical_loss": 3.7444725348052144,
      "tokens_seen": 767719424
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003875626880641926,
      "loss": 2.9211,
      "theoretical_loss": 3.744440613569001,
      "tokens_seen": 767784960
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038755265797392176,
      "loss": 2.9349,
      "theoretical_loss": 3.7444086958202245,
      "tokens_seen": 767850496
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.000387542627883651,
      "loss": 2.8785,
      "theoretical_loss": 3.7443767815582047,
      "tokens_seen": 767916032
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003875325977933801,
      "loss": 2.9355,
      "theoretical_loss": 3.744344870782265,
      "tokens_seen": 767981568
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038752256770310936,
      "loss": 3.0958,
      "theoretical_loss": 3.7443129634917254,
      "tokens_seen": 768047104
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003875125376128385,
      "loss": 3.0222,
      "theoretical_loss": 3.7442810596859095,
      "tokens_seen": 768112640
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003875025075225677,
      "loss": 2.7181,
      "theoretical_loss": 3.7442491593641387,
      "tokens_seen": 768178176
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003874924774322969,
      "loss": 2.8857,
      "theoretical_loss": 3.744217262525735,
      "tokens_seen": 768243712
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003874824473420261,
      "loss": 2.7694,
      "theoretical_loss": 3.744185369170021,
      "tokens_seen": 768309248
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038747241725175526,
      "loss": 2.9329,
      "theoretical_loss": 3.7441534792963203,
      "tokens_seen": 768374784
    },
    {
      "epoch": 2.05,
      "objective/train/docs_used": 878097,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6164374351501465,
      "objective/train/theoretical_loss": 3.744137535665013,
      "objective/train/tokens_used": 788867552,
      "theoretical_loss": 3.744137535665013,
      "tokens_seen": 768407552
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038746238716148444,
      "loss": 2.7017,
      "theoretical_loss": 3.744121592903955,
      "tokens_seen": 768440320
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003874523570712136,
      "loss": 2.9756,
      "theoretical_loss": 3.744089709992248,
      "tokens_seen": 768505856
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038744232698094286,
      "loss": 2.8532,
      "theoretical_loss": 3.744057830560523,
      "tokens_seen": 768571392
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.000387432296890672,
      "loss": 2.8827,
      "theoretical_loss": 3.7440259546081043,
      "tokens_seen": 768636928
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003874222668004012,
      "loss": 2.9375,
      "theoretical_loss": 3.7439940821343134,
      "tokens_seen": 768702464
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038741223671013035,
      "loss": 2.7816,
      "theoretical_loss": 3.7439622131384764,
      "tokens_seen": 768768000
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003874022066198596,
      "loss": 2.5517,
      "theoretical_loss": 3.7439303476199166,
      "tokens_seen": 768833536
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038739217652958876,
      "loss": 2.8064,
      "theoretical_loss": 3.743898485577957,
      "tokens_seen": 768899072
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038738214643931795,
      "loss": 2.8739,
      "theoretical_loss": 3.743866627011924,
      "tokens_seen": 768964608
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003873721163490472,
      "loss": 3.0709,
      "theoretical_loss": 3.743834771921141,
      "tokens_seen": 769030144
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038736208625877636,
      "loss": 2.7886,
      "theoretical_loss": 3.7438029203049332,
      "tokens_seen": 769095680
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038735205616850554,
      "loss": 2.9046,
      "theoretical_loss": 3.743771072162625,
      "tokens_seen": 769161216
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003873420260782347,
      "loss": 2.8764,
      "theoretical_loss": 3.7437392274935424,
      "tokens_seen": 769226752
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003873319959879639,
      "loss": 3.125,
      "theoretical_loss": 3.74370738629701,
      "tokens_seen": 769292288
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003873219658976931,
      "loss": 2.9003,
      "theoretical_loss": 3.7436755485723543,
      "tokens_seen": 769357824
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003873119358074223,
      "loss": 2.7719,
      "theoretical_loss": 3.7436437143189005,
      "tokens_seen": 769423360
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038730190571715145,
      "loss": 2.894,
      "theoretical_loss": 3.7436118835359737,
      "tokens_seen": 769488896
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003872918756268807,
      "loss": 2.7787,
      "theoretical_loss": 3.7435800562229016,
      "tokens_seen": 769554432
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003872818455366098,
      "loss": 2.7862,
      "theoretical_loss": 3.7435482323790095,
      "tokens_seen": 769619968
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038727181544633905,
      "loss": 2.8708,
      "theoretical_loss": 3.743516412003624,
      "tokens_seen": 769685504
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038726178535606823,
      "loss": 2.8517,
      "theoretical_loss": 3.7434845950960725,
      "tokens_seen": 769751040
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003872517552657974,
      "loss": 2.9194,
      "theoretical_loss": 3.74345278165568,
      "tokens_seen": 769816576
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003872417251755266,
      "loss": 2.7662,
      "theoretical_loss": 3.743420971681776,
      "tokens_seen": 769882112
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038723169508525577,
      "loss": 2.7826,
      "theoretical_loss": 3.743389165173686,
      "tokens_seen": 769947648
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038722166499498495,
      "loss": 3.1014,
      "theoretical_loss": 3.7433573621307383,
      "tokens_seen": 770013184
    },
    {
      "epoch": 2.05,
      "objective/train/docs_used": 879358,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.580174684524536,
      "objective/train/theoretical_loss": 3.7433414619084826,
      "objective/train/tokens_used": 790505952,
      "theoretical_loss": 3.7433414619084826,
      "tokens_seen": 770045952
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003872116349047142,
      "loss": 2.6705,
      "theoretical_loss": 3.7433255625522603,
      "tokens_seen": 770078720
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003872016048144433,
      "loss": 2.8358,
      "theoretical_loss": 3.7432937664375796,
      "tokens_seen": 770144256
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038719157472417255,
      "loss": 2.8843,
      "theoretical_loss": 3.7432619737860247,
      "tokens_seen": 770209792
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038718154463390173,
      "loss": 2.9577,
      "theoretical_loss": 3.7432301845969227,
      "tokens_seen": 770275328
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003871715145436309,
      "loss": 2.9062,
      "theoretical_loss": 3.743198398869603,
      "tokens_seen": 770340864
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003871614844533601,
      "loss": 2.9596,
      "theoretical_loss": 3.743166616603394,
      "tokens_seen": 770406400
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003871514543630893,
      "loss": 2.9192,
      "theoretical_loss": 3.7431348377976246,
      "tokens_seen": 770471936
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038714142427281846,
      "loss": 3.0561,
      "theoretical_loss": 3.743103062451623,
      "tokens_seen": 770537472
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003871313941825477,
      "loss": 2.8688,
      "theoretical_loss": 3.7430712905647185,
      "tokens_seen": 770603008
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003871213640922768,
      "loss": 2.7494,
      "theoretical_loss": 3.7430395221362414,
      "tokens_seen": 770668544
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038711133400200605,
      "loss": 2.8565,
      "theoretical_loss": 3.74300775716552,
      "tokens_seen": 770734080
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003871013039117352,
      "loss": 2.9329,
      "theoretical_loss": 3.7429759956518844,
      "tokens_seen": 770799616
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003870912738214644,
      "loss": 2.809,
      "theoretical_loss": 3.7429442375946644,
      "tokens_seen": 770865152
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003870812437311936,
      "loss": 2.8026,
      "theoretical_loss": 3.742912482993191,
      "tokens_seen": 770930688
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003870712136409228,
      "loss": 2.8708,
      "theoretical_loss": 3.7428807318467925,
      "tokens_seen": 770996224
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038706118355065196,
      "loss": 2.9398,
      "theoretical_loss": 3.742848984154801,
      "tokens_seen": 771061760
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003870511534603812,
      "loss": 2.8979,
      "theoretical_loss": 3.7428172399165467,
      "tokens_seen": 771127296
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003870411233701103,
      "loss": 2.8455,
      "theoretical_loss": 3.74278549913136,
      "tokens_seen": 771192832
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038703109327983956,
      "loss": 2.998,
      "theoretical_loss": 3.742753761798572,
      "tokens_seen": 771258368
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003870210631895687,
      "loss": 2.836,
      "theoretical_loss": 3.742722027917514,
      "tokens_seen": 771323904
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003870110330992979,
      "loss": 2.9174,
      "theoretical_loss": 3.7426902974875182,
      "tokens_seen": 771389440
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003870010030090271,
      "loss": 2.7942,
      "theoretical_loss": 3.7426585705079147,
      "tokens_seen": 771454976
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003869909729187563,
      "loss": 3.0534,
      "theoretical_loss": 3.7426268469780357,
      "tokens_seen": 771520512
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038698094282848546,
      "loss": 2.9767,
      "theoretical_loss": 3.7425951268972133,
      "tokens_seen": 771586048
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038697091273821464,
      "loss": 2.7632,
      "theoretical_loss": 3.74256341026478,
      "tokens_seen": 771651584
    },
    {
      "epoch": 2.05,
      "objective/train/docs_used": 880093,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.807681083679199,
      "objective/train/theoretical_loss": 3.7425475532415007,
      "objective/train/tokens_used": 792144352,
      "theoretical_loss": 3.7425475532415007,
      "tokens_seen": 771684352
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003869608826479438,
      "loss": 2.8143,
      "theoretical_loss": 3.742531697080068,
      "tokens_seen": 771717120
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038695085255767306,
      "loss": 2.8962,
      "theoretical_loss": 3.7424999873424087,
      "tokens_seen": 771782656
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003869408224674022,
      "loss": 2.7382,
      "theoretical_loss": 3.742468281051136,
      "tokens_seen": 771848192
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003869307923771314,
      "loss": 3.0543,
      "theoretical_loss": 3.7424365782055817,
      "tokens_seen": 771913728
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038692076228686055,
      "loss": 2.814,
      "theoretical_loss": 3.74240487880508,
      "tokens_seen": 771979264
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003869107321965898,
      "loss": 2.8605,
      "theoretical_loss": 3.7423731828489633,
      "tokens_seen": 772044800
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038690070210631896,
      "loss": 2.9045,
      "theoretical_loss": 3.7423414903365653,
      "tokens_seen": 772110336
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038689067201604815,
      "loss": 2.7878,
      "theoretical_loss": 3.7423098012672193,
      "tokens_seen": 772175872
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038688064192577733,
      "loss": 2.7309,
      "theoretical_loss": 3.74227811564026,
      "tokens_seen": 772241408
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038687061183550656,
      "loss": 2.947,
      "theoretical_loss": 3.7422464334550196,
      "tokens_seen": 772306944
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003868605817452357,
      "loss": 2.7292,
      "theoretical_loss": 3.742214754710834,
      "tokens_seen": 772372480
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003868505516549649,
      "loss": 2.8713,
      "theoretical_loss": 3.742183079407037,
      "tokens_seen": 772438016
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038684052156469405,
      "loss": 2.9634,
      "theoretical_loss": 3.7421514075429627,
      "tokens_seen": 772503552
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003868304914744233,
      "loss": 2.7913,
      "theoretical_loss": 3.742119739117946,
      "tokens_seen": 772569088
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038682046138415247,
      "loss": 3.1463,
      "theoretical_loss": 3.7420880741313223,
      "tokens_seen": 772634624
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038681043129388165,
      "loss": 2.6388,
      "theoretical_loss": 3.742056412582426,
      "tokens_seen": 772700160
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038680040120361083,
      "loss": 2.694,
      "theoretical_loss": 3.7420247544705925,
      "tokens_seen": 772765696
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038679037111334,
      "loss": 2.935,
      "theoretical_loss": 3.741993099795158,
      "tokens_seen": 772831232
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003867803410230692,
      "loss": 3.0614,
      "theoretical_loss": 3.7419614485554566,
      "tokens_seen": 772896768
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038677031093279843,
      "loss": 2.7906,
      "theoretical_loss": 3.7419298007508255,
      "tokens_seen": 772962304
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038676028084252755,
      "loss": 3.0528,
      "theoretical_loss": 3.7418981563806,
      "tokens_seen": 773027840
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003867502507522568,
      "loss": 2.9,
      "theoretical_loss": 3.7418665154441166,
      "tokens_seen": 773093376
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003867402206619859,
      "loss": 2.9703,
      "theoretical_loss": 3.741834877940711,
      "tokens_seen": 773158912
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038673019057171515,
      "loss": 3.0419,
      "theoretical_loss": 3.7418032438697213,
      "tokens_seen": 773224448
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038672016048144433,
      "loss": 2.969,
      "theoretical_loss": 3.7417716132304824,
      "tokens_seen": 773289984
    },
    {
      "epoch": 2.05,
      "objective/train/docs_used": 881313,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6865322589874268,
      "objective/train/theoretical_loss": 3.741755799197563,
      "objective/train/tokens_used": 793782752,
      "theoretical_loss": 3.741755799197563,
      "tokens_seen": 773322752
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003867101303911735,
      "loss": 2.8573,
      "theoretical_loss": 3.741739986022333,
      "tokens_seen": 773355520
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003867001003009027,
      "loss": 2.991,
      "theoretical_loss": 3.7417083622446086,
      "tokens_seen": 773421056
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038669007021063193,
      "loss": 2.9242,
      "theoretical_loss": 3.7416767418966472,
      "tokens_seen": 773486592
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038668004012036106,
      "loss": 2.8793,
      "theoretical_loss": 3.7416451249777865,
      "tokens_seen": 773552128
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003866700100300903,
      "loss": 3.0631,
      "theoretical_loss": 3.7416135114873637,
      "tokens_seen": 773617664
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003866599799398194,
      "loss": 2.7995,
      "theoretical_loss": 3.741581901424717,
      "tokens_seen": 773683200
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038664994984954866,
      "loss": 2.7504,
      "theoretical_loss": 3.7415502947891843,
      "tokens_seen": 773748736
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038663991975927784,
      "loss": 2.9901,
      "theoretical_loss": 3.7415186915801035,
      "tokens_seen": 773814272
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.000386629889669007,
      "loss": 2.9467,
      "theoretical_loss": 3.7414870917968135,
      "tokens_seen": 773879808
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038661985957873625,
      "loss": 2.7866,
      "theoretical_loss": 3.7414554954386525,
      "tokens_seen": 773945344
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003866098294884654,
      "loss": 2.6642,
      "theoretical_loss": 3.7414239025049594,
      "tokens_seen": 774010880
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003865997993981946,
      "loss": 2.7879,
      "theoretical_loss": 3.7413923129950737,
      "tokens_seen": 774076416
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003865897693079238,
      "loss": 3.0073,
      "theoretical_loss": 3.741360726908333,
      "tokens_seen": 774141952
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.000386579739217653,
      "loss": 2.7793,
      "theoretical_loss": 3.7413291442440784,
      "tokens_seen": 774207488
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038656970912738216,
      "loss": 3.0301,
      "theoretical_loss": 3.7412975650016485,
      "tokens_seen": 774273024
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003865596790371114,
      "loss": 2.8888,
      "theoretical_loss": 3.741265989180383,
      "tokens_seen": 774338560
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003865496489468405,
      "loss": 2.9951,
      "theoretical_loss": 3.7412344167796214,
      "tokens_seen": 774404096
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038653961885656976,
      "loss": 2.8143,
      "theoretical_loss": 3.7412028477987045,
      "tokens_seen": 774469632
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003865295887662989,
      "loss": 2.9018,
      "theoretical_loss": 3.7411712822369725,
      "tokens_seen": 774535168
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003865195586760281,
      "loss": 3.0448,
      "theoretical_loss": 3.741139720093765,
      "tokens_seen": 774600704
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003865095285857573,
      "loss": 2.9451,
      "theoretical_loss": 3.741108161368423,
      "tokens_seen": 774666240
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.0003864994984954865,
      "loss": 2.8807,
      "theoretical_loss": 3.7410766060602887,
      "tokens_seen": 774731776
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038648946840521566,
      "loss": 3.0032,
      "theoretical_loss": 3.7410450541687004,
      "tokens_seen": 774797312
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038647943831494484,
      "loss": 2.6414,
      "theoretical_loss": 3.741013505693001,
      "tokens_seen": 774862848
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.000386469408224674,
      "loss": 2.8847,
      "theoretical_loss": 3.740981960632532,
      "tokens_seen": 774928384
    },
    {
      "epoch": 2.05,
      "objective/train/docs_used": 882028,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.3022451400756836,
      "objective/train/theoretical_loss": 3.740966189382802,
      "objective/train/tokens_used": 795421152,
      "theoretical_loss": 3.740966189382802,
      "tokens_seen": 774961152
    },
    {
      "epoch": 2.05,
      "learning_rate": 0.00038645937813440326,
      "loss": 2.8122,
      "theoretical_loss": 3.7409504189866336,
      "tokens_seen": 774993920
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003864493480441324,
      "loss": 2.8759,
      "theoretical_loss": 3.7409188807546485,
      "tokens_seen": 775059456
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003864393179538616,
      "loss": 2.8984,
      "theoretical_loss": 3.740887345935918,
      "tokens_seen": 775124992
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038642928786359075,
      "loss": 2.902,
      "theoretical_loss": 3.7408558145297848,
      "tokens_seen": 775190528
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038641925777332,
      "loss": 2.8342,
      "theoretical_loss": 3.7408242865355907,
      "tokens_seen": 775256064
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038640922768304917,
      "loss": 2.9901,
      "theoretical_loss": 3.7407927619526786,
      "tokens_seen": 775321600
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038639919759277835,
      "loss": 2.8286,
      "theoretical_loss": 3.7407612407803903,
      "tokens_seen": 775387136
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038638916750250753,
      "loss": 2.8755,
      "theoretical_loss": 3.7407297230180694,
      "tokens_seen": 775452672
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038637913741223676,
      "loss": 2.7858,
      "theoretical_loss": 3.7406982086650578,
      "tokens_seen": 775518208
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003863691073219659,
      "loss": 2.7104,
      "theoretical_loss": 3.7406666977207,
      "tokens_seen": 775583744
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003863590772316951,
      "loss": 2.8863,
      "theoretical_loss": 3.740635190184338,
      "tokens_seen": 775649280
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038634904714142425,
      "loss": 2.9168,
      "theoretical_loss": 3.7406036860553167,
      "tokens_seen": 775714816
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003863390170511535,
      "loss": 2.9828,
      "theoretical_loss": 3.7405721853329785,
      "tokens_seen": 775780352
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038632898696088267,
      "loss": 2.6336,
      "theoretical_loss": 3.7405406880166687,
      "tokens_seen": 775845888
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038631895687061185,
      "loss": 2.7923,
      "theoretical_loss": 3.74050919410573,
      "tokens_seen": 775911424
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038630892678034103,
      "loss": 2.8397,
      "theoretical_loss": 3.740477703599507,
      "tokens_seen": 775976960
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003862988966900702,
      "loss": 2.8475,
      "theoretical_loss": 3.740446216497344,
      "tokens_seen": 776042496
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003862888665997994,
      "loss": 2.7932,
      "theoretical_loss": 3.7404147327985866,
      "tokens_seen": 776108032
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038627883650952863,
      "loss": 2.8546,
      "theoretical_loss": 3.7403832525025784,
      "tokens_seen": 776173568
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038626880641925775,
      "loss": 2.9357,
      "theoretical_loss": 3.740351775608665,
      "tokens_seen": 776239104
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.000386258776328987,
      "loss": 2.8414,
      "theoretical_loss": 3.7403203021161913,
      "tokens_seen": 776304640
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003862487462387161,
      "loss": 2.916,
      "theoretical_loss": 3.740288832024503,
      "tokens_seen": 776370176
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038623871614844535,
      "loss": 2.9987,
      "theoretical_loss": 3.740257365332945,
      "tokens_seen": 776435712
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038622868605817453,
      "loss": 2.8645,
      "theoretical_loss": 3.740225902040864,
      "tokens_seen": 776501248
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003862186559679037,
      "loss": 2.8856,
      "theoretical_loss": 3.740194442147604,
      "tokens_seen": 776566784
    },
    {
      "epoch": 2.06,
      "objective/train/docs_used": 883113,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.753779411315918,
      "objective/train/theoretical_loss": 3.740178713475329,
      "objective/train/tokens_used": 797059552,
      "theoretical_loss": 3.740178713475329,
      "tokens_seen": 776599552
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003862086258776329,
      "loss": 2.8486,
      "theoretical_loss": 3.740162985652513,
      "tokens_seen": 776632320
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038619859578736213,
      "loss": 2.783,
      "theoretical_loss": 3.740131532554937,
      "tokens_seen": 776697856
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038618856569709126,
      "loss": 2.7231,
      "theoretical_loss": 3.7401000828542212,
      "tokens_seen": 776763392
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003861785356068205,
      "loss": 2.8215,
      "theoretical_loss": 3.740068636549713,
      "tokens_seen": 776828928
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003861685055165496,
      "loss": 2.8173,
      "theoretical_loss": 3.740037193640759,
      "tokens_seen": 776894464
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038615847542627886,
      "loss": 2.7441,
      "theoretical_loss": 3.7400057541267064,
      "tokens_seen": 776960000
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038614844533600804,
      "loss": 2.7466,
      "theoretical_loss": 3.739974318006902,
      "tokens_seen": 777025536
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003861384152457372,
      "loss": 2.9235,
      "theoretical_loss": 3.7399428852806937,
      "tokens_seen": 777091072
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003861283851554664,
      "loss": 2.9365,
      "theoretical_loss": 3.7399114559474285,
      "tokens_seen": 777156608
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003861183550651956,
      "loss": 2.8048,
      "theoretical_loss": 3.739880030006454,
      "tokens_seen": 777222144
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038610832497492476,
      "loss": 2.7476,
      "theoretical_loss": 3.7398486074571182,
      "tokens_seen": 777287680
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.000386098294884654,
      "loss": 2.7782,
      "theoretical_loss": 3.73981718829877,
      "tokens_seen": 777353216
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003860882647943831,
      "loss": 2.8754,
      "theoretical_loss": 3.7397857725307553,
      "tokens_seen": 777418752
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038607823470411236,
      "loss": 2.8833,
      "theoretical_loss": 3.7397543601524252,
      "tokens_seen": 777484288
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003860682046138415,
      "loss": 2.569,
      "theoretical_loss": 3.739722951163127,
      "tokens_seen": 777549824
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003860581745235707,
      "loss": 2.8597,
      "theoretical_loss": 3.7396915455622093,
      "tokens_seen": 777615360
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003860481444332999,
      "loss": 2.9052,
      "theoretical_loss": 3.7396601433490213,
      "tokens_seen": 777680896
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003860381143430291,
      "loss": 2.8995,
      "theoretical_loss": 3.7396287445229124,
      "tokens_seen": 777746432
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038602808425275826,
      "loss": 2.6981,
      "theoretical_loss": 3.7395973490832315,
      "tokens_seen": 777811968
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003860180541624875,
      "loss": 2.683,
      "theoretical_loss": 3.739565957029328,
      "tokens_seen": 777877504
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003860080240722166,
      "loss": 2.7552,
      "theoretical_loss": 3.7395345683605523,
      "tokens_seen": 777943040
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038599799398194586,
      "loss": 2.8937,
      "theoretical_loss": 3.739503183076253,
      "tokens_seen": 778008576
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.000385987963891675,
      "loss": 2.646,
      "theoretical_loss": 3.7394718011757817,
      "tokens_seen": 778074112
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003859779338014042,
      "loss": 2.8996,
      "theoretical_loss": 3.739440422658487,
      "tokens_seen": 778139648
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003859679037111334,
      "loss": 2.947,
      "theoretical_loss": 3.7394090475237203,
      "tokens_seen": 778205184
    },
    {
      "epoch": 2.06,
      "objective/train/docs_used": 883877,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5676019191741943,
      "objective/train/theoretical_loss": 3.7393933612245824,
      "objective/train/tokens_used": 798697952,
      "theoretical_loss": 3.7393933612245824,
      "tokens_seen": 778237952
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003859578736208626,
      "loss": 2.6764,
      "theoretical_loss": 3.7393776757708324,
      "tokens_seen": 778270720
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038594784353059177,
      "loss": 2.8117,
      "theoretical_loss": 3.7393463073991726,
      "tokens_seen": 778336256
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038593781344032095,
      "loss": 2.9232,
      "theoretical_loss": 3.7393149424080936,
      "tokens_seen": 778401792
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038592778335005013,
      "loss": 2.7402,
      "theoretical_loss": 3.739283580796945,
      "tokens_seen": 778467328
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038591775325977937,
      "loss": 2.8319,
      "theoretical_loss": 3.7392522225650793,
      "tokens_seen": 778532864
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003859077231695085,
      "loss": 2.9142,
      "theoretical_loss": 3.739220867711847,
      "tokens_seen": 778598400
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038589769307923773,
      "loss": 2.896,
      "theoretical_loss": 3.7391895162366002,
      "tokens_seen": 778663936
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038588766298896685,
      "loss": 2.5929,
      "theoretical_loss": 3.7391581681386907,
      "tokens_seen": 778729472
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003858776328986961,
      "loss": 2.9375,
      "theoretical_loss": 3.7391268234174704,
      "tokens_seen": 778795008
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003858676028084253,
      "loss": 2.8272,
      "theoretical_loss": 3.7390954820722917,
      "tokens_seen": 778860544
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038585757271815445,
      "loss": 2.8612,
      "theoretical_loss": 3.7390641441025068,
      "tokens_seen": 778926080
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003858475426278837,
      "loss": 3.0562,
      "theoretical_loss": 3.7390328095074676,
      "tokens_seen": 778991616
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038583751253761287,
      "loss": 2.9644,
      "theoretical_loss": 3.7390014782865277,
      "tokens_seen": 779057152
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038582748244734205,
      "loss": 2.7878,
      "theoretical_loss": 3.73897015043904,
      "tokens_seen": 779122688
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038581745235707123,
      "loss": 2.7606,
      "theoretical_loss": 3.7389388259643566,
      "tokens_seen": 779188224
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003858074222668004,
      "loss": 2.8087,
      "theoretical_loss": 3.7389075048618317,
      "tokens_seen": 779253760
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003857973921765296,
      "loss": 2.8811,
      "theoretical_loss": 3.7388761871308187,
      "tokens_seen": 779319296
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038578736208625883,
      "loss": 3.0463,
      "theoretical_loss": 3.7388448727706707,
      "tokens_seen": 779384832
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038577733199598796,
      "loss": 2.8511,
      "theoretical_loss": 3.738813561780742,
      "tokens_seen": 779450368
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003857673019057172,
      "loss": 2.763,
      "theoretical_loss": 3.7387822541603857,
      "tokens_seen": 779515904
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003857572718154463,
      "loss": 2.9689,
      "theoretical_loss": 3.738750949908957,
      "tokens_seen": 779581440
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038574724172517555,
      "loss": 2.7182,
      "theoretical_loss": 3.738719649025809,
      "tokens_seen": 779646976
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038573721163490473,
      "loss": 2.8999,
      "theoretical_loss": 3.7386883515102975,
      "tokens_seen": 779712512
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003857271815446339,
      "loss": 2.9385,
      "theoretical_loss": 3.7386570573617766,
      "tokens_seen": 779778048
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003857171514543631,
      "loss": 2.7616,
      "theoretical_loss": 3.738625766579601,
      "tokens_seen": 779843584
    },
    {
      "epoch": 2.06,
      "objective/train/docs_used": 885041,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8538613319396973,
      "objective/train/theoretical_loss": 3.738610122450691,
      "objective/train/tokens_used": 800336352,
      "theoretical_loss": 3.738610122450691,
      "tokens_seen": 779876352
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038570712136409233,
      "loss": 2.8281,
      "theoretical_loss": 3.7385944791631256,
      "tokens_seen": 779909120
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038569709127382146,
      "loss": 2.8344,
      "theoretical_loss": 3.738563195111706,
      "tokens_seen": 779974656
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003856870611835507,
      "loss": 3.0381,
      "theoretical_loss": 3.7385319144246973,
      "tokens_seen": 780040192
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003856770310932798,
      "loss": 2.6559,
      "theoretical_loss": 3.7385006371014553,
      "tokens_seen": 780105728
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038566700100300906,
      "loss": 3.129,
      "theoretical_loss": 3.738469363141336,
      "tokens_seen": 780171264
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038565697091273824,
      "loss": 2.6653,
      "theoretical_loss": 3.738438092543694,
      "tokens_seen": 780236800
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003856469408224674,
      "loss": 2.7597,
      "theoretical_loss": 3.7384068253078873,
      "tokens_seen": 780302336
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003856369107321966,
      "loss": 2.9662,
      "theoretical_loss": 3.7383755614332705,
      "tokens_seen": 780367872
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003856268806419258,
      "loss": 3.0753,
      "theoretical_loss": 3.7383443009192012,
      "tokens_seen": 780433408
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038561685055165496,
      "loss": 2.7267,
      "theoretical_loss": 3.738313043765035,
      "tokens_seen": 780498944
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003856068204613842,
      "loss": 3.0858,
      "theoretical_loss": 3.73828178997013,
      "tokens_seen": 780564480
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003855967903711133,
      "loss": 2.8874,
      "theoretical_loss": 3.738250539533842,
      "tokens_seen": 780630016
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038558676028084256,
      "loss": 2.8781,
      "theoretical_loss": 3.738219292455528,
      "tokens_seen": 780695552
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003855767301905717,
      "loss": 2.8912,
      "theoretical_loss": 3.738188048734547,
      "tokens_seen": 780761088
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003855667001003009,
      "loss": 2.9454,
      "theoretical_loss": 3.738156808370255,
      "tokens_seen": 780826624
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003855566700100301,
      "loss": 2.9832,
      "theoretical_loss": 3.7381255713620103,
      "tokens_seen": 780892160
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003855466399197593,
      "loss": 2.8575,
      "theoretical_loss": 3.73809433770917,
      "tokens_seen": 780957696
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038553660982948846,
      "loss": 2.751,
      "theoretical_loss": 3.7380631074110933,
      "tokens_seen": 781023232
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003855265797392177,
      "loss": 2.7593,
      "theoretical_loss": 3.7380318804671377,
      "tokens_seen": 781088768
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003855165496489468,
      "loss": 2.6554,
      "theoretical_loss": 3.7380006568766615,
      "tokens_seen": 781154304
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038550651955867606,
      "loss": 2.8923,
      "theoretical_loss": 3.7379694366390237,
      "tokens_seen": 781219840
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003854964894684052,
      "loss": 2.8653,
      "theoretical_loss": 3.737938219753583,
      "tokens_seen": 781285376
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003854864593781344,
      "loss": 2.7881,
      "theoretical_loss": 3.7379070062196975,
      "tokens_seen": 781350912
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003854764292878636,
      "loss": 2.9967,
      "theoretical_loss": 3.7378757960367275,
      "tokens_seen": 781416448
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003854663991975928,
      "loss": 2.7699,
      "theoretical_loss": 3.7378445892040317,
      "tokens_seen": 781481984
    },
    {
      "epoch": 2.06,
      "objective/train/docs_used": 890004,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.93015456199646,
      "objective/train/theoretical_loss": 3.7378289870438364,
      "objective/train/tokens_used": 801974752,
      "theoretical_loss": 3.7378289870438364,
      "tokens_seen": 781514752
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038545636910732197,
      "loss": 2.823,
      "theoretical_loss": 3.737813385720969,
      "tokens_seen": 781547520
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038544633901705115,
      "loss": 2.6779,
      "theoretical_loss": 3.7377821855869007,
      "tokens_seen": 781613056
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038543630892678033,
      "loss": 2.9013,
      "theoretical_loss": 3.7377509888011846,
      "tokens_seen": 781678592
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038542627883650957,
      "loss": 2.8486,
      "theoretical_loss": 3.7377197953631818,
      "tokens_seen": 781744128
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003854162487462387,
      "loss": 2.8487,
      "theoretical_loss": 3.737688605272252,
      "tokens_seen": 781809664
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038540621865596793,
      "loss": 2.7669,
      "theoretical_loss": 3.7376574185277565,
      "tokens_seen": 781875200
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038539618856569705,
      "loss": 2.6736,
      "theoretical_loss": 3.7376262351290546,
      "tokens_seen": 781940736
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003853861584754263,
      "loss": 3.0875,
      "theoretical_loss": 3.7375950550755075,
      "tokens_seen": 782006272
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038537612838515547,
      "loss": 2.9593,
      "theoretical_loss": 3.737563878366476,
      "tokens_seen": 782071808
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038536609829488465,
      "loss": 2.6679,
      "theoretical_loss": 3.737532705001321,
      "tokens_seen": 782137344
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038535606820461383,
      "loss": 3.0975,
      "theoretical_loss": 3.7375015349794047,
      "tokens_seen": 782202880
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038534603811434307,
      "loss": 2.8719,
      "theoretical_loss": 3.7374703683000865,
      "tokens_seen": 782268416
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003853360080240722,
      "loss": 3.0239,
      "theoretical_loss": 3.7374392049627296,
      "tokens_seen": 782333952
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038532597793380143,
      "loss": 2.6991,
      "theoretical_loss": 3.7374080449666955,
      "tokens_seen": 782399488
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038531594784353056,
      "loss": 3.0373,
      "theoretical_loss": 3.7373768883113456,
      "tokens_seen": 782465024
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003853059177532598,
      "loss": 2.7967,
      "theoretical_loss": 3.737345734996042,
      "tokens_seen": 782530560
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.000385295887662989,
      "loss": 2.7177,
      "theoretical_loss": 3.7373145850201475,
      "tokens_seen": 782596096
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038528585757271816,
      "loss": 2.8801,
      "theoretical_loss": 3.7372834383830247,
      "tokens_seen": 782661632
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038527582748244734,
      "loss": 2.6214,
      "theoretical_loss": 3.7372522950840352,
      "tokens_seen": 782727168
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003852657973921765,
      "loss": 2.8109,
      "theoretical_loss": 3.7372211551225423,
      "tokens_seen": 782792704
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003852557673019057,
      "loss": 3.0358,
      "theoretical_loss": 3.737190018497909,
      "tokens_seen": 782858240
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038524573721163493,
      "loss": 2.8844,
      "theoretical_loss": 3.7371588852094986,
      "tokens_seen": 782923776
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038523570712136406,
      "loss": 2.9968,
      "theoretical_loss": 3.7371277552566737,
      "tokens_seen": 782989312
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003852256770310933,
      "loss": 2.758,
      "theoretical_loss": 3.737096628638799,
      "tokens_seen": 783054848
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003852156469408225,
      "loss": 2.8207,
      "theoretical_loss": 3.737065505355237,
      "tokens_seen": 783120384
    },
    {
      "epoch": 2.06,
      "objective/train/docs_used": 897763,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7294697761535645,
      "objective/train/theoretical_loss": 3.7370499449636245,
      "objective/train/tokens_used": 803613152,
      "theoretical_loss": 3.7370499449636245,
      "tokens_seen": 783153152
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038520561685055166,
      "loss": 2.6323,
      "theoretical_loss": 3.7370343854053525,
      "tokens_seen": 783185920
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038519558676028084,
      "loss": 2.9722,
      "theoretical_loss": 3.7370032687885084,
      "tokens_seen": 783251456
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038518555667001,
      "loss": 2.8461,
      "theoretical_loss": 3.7369721555040702,
      "tokens_seen": 783316992
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003851755265797392,
      "loss": 2.7199,
      "theoretical_loss": 3.736941045551401,
      "tokens_seen": 783382528
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038516549648946844,
      "loss": 2.8761,
      "theoretical_loss": 3.736909938929866,
      "tokens_seen": 783448064
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038515546639919756,
      "loss": 3.1976,
      "theoretical_loss": 3.7368788356388305,
      "tokens_seen": 783513600
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003851454363089268,
      "loss": 2.904,
      "theoretical_loss": 3.7368477356776584,
      "tokens_seen": 783579136
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003851354062186559,
      "loss": 2.8346,
      "theoretical_loss": 3.7368166390457147,
      "tokens_seen": 783644672
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038512537612838516,
      "loss": 2.7427,
      "theoretical_loss": 3.7367855457423653,
      "tokens_seen": 783710208
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003851153460381144,
      "loss": 2.7978,
      "theoretical_loss": 3.736754455766975,
      "tokens_seen": 783775744
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003851053159478435,
      "loss": 2.9312,
      "theoretical_loss": 3.73672336911891,
      "tokens_seen": 783841280
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038509528585757276,
      "loss": 2.7718,
      "theoretical_loss": 3.7366922857975355,
      "tokens_seen": 783906816
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003850852557673019,
      "loss": 2.7926,
      "theoretical_loss": 3.7366612058022177,
      "tokens_seen": 783972352
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003850752256770311,
      "loss": 2.9423,
      "theoretical_loss": 3.736630129132323,
      "tokens_seen": 784037888
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003850651955867603,
      "loss": 2.71,
      "theoretical_loss": 3.7365990557872166,
      "tokens_seen": 784103424
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003850551654964895,
      "loss": 2.9139,
      "theoretical_loss": 3.7365679857662664,
      "tokens_seen": 784168960
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038504513540621866,
      "loss": 2.8589,
      "theoretical_loss": 3.736536919068838,
      "tokens_seen": 784234496
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003850351053159479,
      "loss": 2.8182,
      "theoretical_loss": 3.7365058556942987,
      "tokens_seen": 784300032
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.000385025075225677,
      "loss": 2.9608,
      "theoretical_loss": 3.736474795642015,
      "tokens_seen": 784365568
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038501504513540626,
      "loss": 2.8076,
      "theoretical_loss": 3.736443738911354,
      "tokens_seen": 784431104
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003850050150451354,
      "loss": 2.8234,
      "theoretical_loss": 3.7364126855016835,
      "tokens_seen": 784496640
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003849949849548646,
      "loss": 2.6401,
      "theoretical_loss": 3.7363816354123713,
      "tokens_seen": 784562176
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003849849548645938,
      "loss": 2.8178,
      "theoretical_loss": 3.736350588642784,
      "tokens_seen": 784627712
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.000384974924774323,
      "loss": 2.9572,
      "theoretical_loss": 3.7363195451922904,
      "tokens_seen": 784693248
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038496489468405217,
      "loss": 2.9677,
      "theoretical_loss": 3.7362885050602577,
      "tokens_seen": 784758784
    },
    {
      "epoch": 2.06,
      "objective/train/docs_used": 900073,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.1294591426849365,
      "objective/train/theoretical_loss": 3.7362729862384674,
      "objective/train/tokens_used": 805251552,
      "theoretical_loss": 3.7362729862384674,
      "tokens_seen": 784791552
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038495486459378135,
      "loss": 2.8987,
      "theoretical_loss": 3.736257468246055,
      "tokens_seen": 784824320
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038494483450351053,
      "loss": 2.8781,
      "theoretical_loss": 3.7362264347490504,
      "tokens_seen": 784889856
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038493480441323977,
      "loss": 2.9147,
      "theoretical_loss": 3.7361954045686114,
      "tokens_seen": 784955392
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003849247743229689,
      "loss": 2.7068,
      "theoretical_loss": 3.7361643777041076,
      "tokens_seen": 785020928
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038491474423269813,
      "loss": 2.716,
      "theoretical_loss": 3.7361333541549078,
      "tokens_seen": 785086464
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038490471414242725,
      "loss": 2.923,
      "theoretical_loss": 3.7361023339203814,
      "tokens_seen": 785152000
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003848946840521565,
      "loss": 2.7817,
      "theoretical_loss": 3.736071316999897,
      "tokens_seen": 785217536
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038488465396188567,
      "loss": 2.8024,
      "theoretical_loss": 3.7360403033928242,
      "tokens_seen": 785283072
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038487462387161485,
      "loss": 2.9971,
      "theoretical_loss": 3.7360092930985322,
      "tokens_seen": 785348608
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038486459378134403,
      "loss": 3.0687,
      "theoretical_loss": 3.7359782861163917,
      "tokens_seen": 785414144
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038485456369107327,
      "loss": 2.8731,
      "theoretical_loss": 3.735947282445772,
      "tokens_seen": 785479680
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003848445336008024,
      "loss": 2.8119,
      "theoretical_loss": 3.735916282086043,
      "tokens_seen": 785545216
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038483450351053163,
      "loss": 2.7403,
      "theoretical_loss": 3.735885285036575,
      "tokens_seen": 785610752
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038482447342026076,
      "loss": 2.7575,
      "theoretical_loss": 3.7358542912967385,
      "tokens_seen": 785676288
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038481444332999,
      "loss": 2.912,
      "theoretical_loss": 3.7358233008659045,
      "tokens_seen": 785741824
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003848044132397192,
      "loss": 2.9174,
      "theoretical_loss": 3.7357923137434437,
      "tokens_seen": 785807360
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038479438314944836,
      "loss": 2.7221,
      "theoretical_loss": 3.7357613299287262,
      "tokens_seen": 785872896
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038478435305917754,
      "loss": 2.7952,
      "theoretical_loss": 3.7357303494211243,
      "tokens_seen": 785938432
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003847743229689067,
      "loss": 2.7357,
      "theoretical_loss": 3.735699372220008,
      "tokens_seen": 786003968
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003847642928786359,
      "loss": 2.6275,
      "theoretical_loss": 3.7356683983247496,
      "tokens_seen": 786069504
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038475426278836513,
      "loss": 2.6422,
      "theoretical_loss": 3.735637427734721,
      "tokens_seen": 786135040
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038474423269809426,
      "loss": 2.8061,
      "theoretical_loss": 3.7356064604492936,
      "tokens_seen": 786200576
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003847342026078235,
      "loss": 2.8188,
      "theoretical_loss": 3.735575496467839,
      "tokens_seen": 786266112
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003847241725175527,
      "loss": 3.1006,
      "theoretical_loss": 3.73554453578973,
      "tokens_seen": 786331648
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038471414242728186,
      "loss": 2.8199,
      "theoretical_loss": 3.7355135784143387,
      "tokens_seen": 786397184
    },
    {
      "epoch": 2.06,
      "objective/train/docs_used": 905150,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.815542459487915,
      "objective/train/theoretical_loss": 3.735498100964966,
      "objective/train/tokens_used": 806889952,
      "theoretical_loss": 3.735498100964966,
      "tokens_seen": 786429952
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038470411233701104,
      "loss": 2.7891,
      "theoretical_loss": 3.735482624341037,
      "tokens_seen": 786462720
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003846940822467402,
      "loss": 2.8774,
      "theoretical_loss": 3.735451673569199,
      "tokens_seen": 786528256
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003846840521564694,
      "loss": 2.9415,
      "theoretical_loss": 3.7354207260981953,
      "tokens_seen": 786593792
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038467402206619864,
      "loss": 2.9055,
      "theoretical_loss": 3.7353897819274016,
      "tokens_seen": 786659328
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038466399197592776,
      "loss": 2.7626,
      "theoretical_loss": 3.735358841056189,
      "tokens_seen": 786724864
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.000384653961885657,
      "loss": 2.9134,
      "theoretical_loss": 3.735327903483931,
      "tokens_seen": 786790400
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003846439317953861,
      "loss": 2.9937,
      "theoretical_loss": 3.7352969692100024,
      "tokens_seen": 786855936
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038463390170511536,
      "loss": 2.9249,
      "theoretical_loss": 3.7352660382337763,
      "tokens_seen": 786921472
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038462387161484454,
      "loss": 2.8463,
      "theoretical_loss": 3.735235110554626,
      "tokens_seen": 786987008
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003846138415245737,
      "loss": 2.9076,
      "theoretical_loss": 3.735204186171926,
      "tokens_seen": 787052544
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003846038114343029,
      "loss": 2.7049,
      "theoretical_loss": 3.7351732650850504,
      "tokens_seen": 787118080
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003845937813440321,
      "loss": 2.7692,
      "theoretical_loss": 3.7351423472933734,
      "tokens_seen": 787183616
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038458375125376127,
      "loss": 2.8375,
      "theoretical_loss": 3.7351114327962702,
      "tokens_seen": 787249152
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003845737211634905,
      "loss": 2.9005,
      "theoretical_loss": 3.735080521593115,
      "tokens_seen": 787314688
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038456369107321963,
      "loss": 2.7006,
      "theoretical_loss": 3.7350496136832825,
      "tokens_seen": 787380224
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038455366098294886,
      "loss": 2.8368,
      "theoretical_loss": 3.735018709066148,
      "tokens_seen": 787445760
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038454363089267805,
      "loss": 2.8671,
      "theoretical_loss": 3.7349878077410867,
      "tokens_seen": 787511296
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038453360080240723,
      "loss": 2.8951,
      "theoretical_loss": 3.734956909707474,
      "tokens_seen": 787576832
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003845235707121364,
      "loss": 2.8302,
      "theoretical_loss": 3.734926014964686,
      "tokens_seen": 787642368
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003845135406218656,
      "loss": 2.7686,
      "theoretical_loss": 3.7348951235120977,
      "tokens_seen": 787707904
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038450351053159477,
      "loss": 2.7914,
      "theoretical_loss": 3.734864235349085,
      "tokens_seen": 787773440
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.000384493480441324,
      "loss": 2.7126,
      "theoretical_loss": 3.7348333504750246,
      "tokens_seen": 787838976
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038448345035105313,
      "loss": 3.0019,
      "theoretical_loss": 3.734802468889292,
      "tokens_seen": 787904512
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038447342026078237,
      "loss": 3.1167,
      "theoretical_loss": 3.734771590591264,
      "tokens_seen": 787970048
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003844633901705115,
      "loss": 2.921,
      "theoretical_loss": 3.7347407155803176,
      "tokens_seen": 788035584
    },
    {
      "epoch": 2.06,
      "objective/train/docs_used": 910242,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.993212938308716,
      "objective/train/theoretical_loss": 3.734725279307305,
      "objective/train/tokens_used": 808528352,
      "theoretical_loss": 3.734725279307305,
      "tokens_seen": 788068352
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038445336008024073,
      "loss": 3.0711,
      "theoretical_loss": 3.734709843855829,
      "tokens_seen": 788101120
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003844433299899699,
      "loss": 2.9976,
      "theoretical_loss": 3.734678975417175,
      "tokens_seen": 788166656
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003844332998996991,
      "loss": 2.8315,
      "theoretical_loss": 3.7346481102637332,
      "tokens_seen": 788232192
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003844232698094283,
      "loss": 2.8954,
      "theoretical_loss": 3.7346172483948803,
      "tokens_seen": 788297728
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038441323971915745,
      "loss": 2.7735,
      "theoretical_loss": 3.7345863898099942,
      "tokens_seen": 788363264
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038440320962888664,
      "loss": 2.8864,
      "theoretical_loss": 3.7345555345084525,
      "tokens_seen": 788428800
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038439317953861587,
      "loss": 3.0191,
      "theoretical_loss": 3.7345246824896328,
      "tokens_seen": 788494336
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038438314944834505,
      "loss": 2.9265,
      "theoretical_loss": 3.7344938337529134,
      "tokens_seen": 788559872
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038437311935807423,
      "loss": 2.7427,
      "theoretical_loss": 3.7344629882976714,
      "tokens_seen": 788625408
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038436308926780347,
      "loss": 2.8258,
      "theoretical_loss": 3.7344321461232863,
      "tokens_seen": 788690944
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003843530591775326,
      "loss": 3.0298,
      "theoretical_loss": 3.734401307229136,
      "tokens_seen": 788756480
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038434302908726183,
      "loss": 2.9562,
      "theoretical_loss": 3.734370471614599,
      "tokens_seen": 788822016
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038433299899699096,
      "loss": 2.8713,
      "theoretical_loss": 3.734339639279055,
      "tokens_seen": 788887552
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003843229689067202,
      "loss": 2.6562,
      "theoretical_loss": 3.734308810221881,
      "tokens_seen": 788953088
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003843129388164494,
      "loss": 3.0193,
      "theoretical_loss": 3.7342779844424583,
      "tokens_seen": 789018624
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038430290872617856,
      "loss": 2.7786,
      "theoretical_loss": 3.734247161940165,
      "tokens_seen": 789084160
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038429287863590774,
      "loss": 2.7898,
      "theoretical_loss": 3.734216342714381,
      "tokens_seen": 789149696
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003842828485456369,
      "loss": 2.8075,
      "theoretical_loss": 3.7341855267644855,
      "tokens_seen": 789215232
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003842728184553661,
      "loss": 2.9642,
      "theoretical_loss": 3.7341547140898586,
      "tokens_seen": 789280768
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038426278836509533,
      "loss": 2.824,
      "theoretical_loss": 3.7341239046898806,
      "tokens_seen": 789346304
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038425275827482446,
      "loss": 2.7216,
      "theoretical_loss": 3.734093098563931,
      "tokens_seen": 789411840
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003842427281845537,
      "loss": 2.6687,
      "theoretical_loss": 3.7340622957113907,
      "tokens_seen": 789477376
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003842326980942829,
      "loss": 2.8243,
      "theoretical_loss": 3.7340314961316396,
      "tokens_seen": 789542912
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038422266800401206,
      "loss": 2.8471,
      "theoretical_loss": 3.734000699824059,
      "tokens_seen": 789608448
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038421263791374124,
      "loss": 2.8609,
      "theoretical_loss": 3.7339699067880296,
      "tokens_seen": 789673984
    },
    {
      "epoch": 2.06,
      "objective/train/docs_used": 911523,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8604464530944824,
      "objective/train/theoretical_loss": 3.733954511496653,
      "objective/train/tokens_used": 810166752,
      "theoretical_loss": 3.733954511496653,
      "tokens_seen": 789706752
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003842026078234704,
      "loss": 2.9005,
      "theoretical_loss": 3.7339391170229317,
      "tokens_seen": 789739520
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003841925777331996,
      "loss": 2.9414,
      "theoretical_loss": 3.733908330528147,
      "tokens_seen": 789805056
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038418254764292884,
      "loss": 2.8559,
      "theoretical_loss": 3.7338775473030568,
      "tokens_seen": 789870592
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038417251755265796,
      "loss": 3.1541,
      "theoretical_loss": 3.733846767347043,
      "tokens_seen": 789936128
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003841624874623872,
      "loss": 2.7023,
      "theoretical_loss": 3.7338159906594863,
      "tokens_seen": 790001664
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003841524573721163,
      "loss": 2.8034,
      "theoretical_loss": 3.7337852172397694,
      "tokens_seen": 790067200
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038414242728184556,
      "loss": 2.7727,
      "theoretical_loss": 3.7337544470872737,
      "tokens_seen": 790132736
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038413239719157474,
      "loss": 2.7565,
      "theoretical_loss": 3.733723680201382,
      "tokens_seen": 790198272
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003841223671013039,
      "loss": 2.8344,
      "theoretical_loss": 3.733692916581476,
      "tokens_seen": 790263808
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003841123370110331,
      "loss": 2.8147,
      "theoretical_loss": 3.7336621562269383,
      "tokens_seen": 790329344
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003841023069207623,
      "loss": 2.8566,
      "theoretical_loss": 3.7336313991371517,
      "tokens_seen": 790394880
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038409227683049147,
      "loss": 2.9304,
      "theoretical_loss": 3.7336006453114994,
      "tokens_seen": 790460416
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003840822467402207,
      "loss": 2.9737,
      "theoretical_loss": 3.7335698947493645,
      "tokens_seen": 790525952
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038407221664994983,
      "loss": 2.7628,
      "theoretical_loss": 3.733539147450129,
      "tokens_seen": 790591488
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038406218655967907,
      "loss": 2.6498,
      "theoretical_loss": 3.733508403413177,
      "tokens_seen": 790657024
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038405215646940825,
      "loss": 2.9293,
      "theoretical_loss": 3.7334776626378927,
      "tokens_seen": 790722560
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038404212637913743,
      "loss": 2.8448,
      "theoretical_loss": 3.7334469251236584,
      "tokens_seen": 790788096
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003840320962888666,
      "loss": 2.8674,
      "theoretical_loss": 3.733416190869859,
      "tokens_seen": 790853632
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003840220661985958,
      "loss": 2.9442,
      "theoretical_loss": 3.733385459875878,
      "tokens_seen": 790919168
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038401203610832497,
      "loss": 2.7676,
      "theoretical_loss": 3.7333547321411,
      "tokens_seen": 790984704
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003840020060180542,
      "loss": 2.5382,
      "theoretical_loss": 3.7333240076649092,
      "tokens_seen": 791050240
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038399197592778333,
      "loss": 2.9655,
      "theoretical_loss": 3.7332932864466897,
      "tokens_seen": 791115776
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038398194583751257,
      "loss": 2.9938,
      "theoretical_loss": 3.7332625684858263,
      "tokens_seen": 791181312
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003839719157472417,
      "loss": 2.8769,
      "theoretical_loss": 3.733231853781705,
      "tokens_seen": 791246848
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038396188565697093,
      "loss": 2.8183,
      "theoretical_loss": 3.7332011423337086,
      "tokens_seen": 791312384
    },
    {
      "epoch": 2.06,
      "objective/train/docs_used": 912176,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.054990291595459,
      "objective/train/theoretical_loss": 3.733185787830566,
      "objective/train/tokens_used": 811805152,
      "theoretical_loss": 3.733185787830566,
      "tokens_seen": 791345152
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003839518555667001,
      "loss": 2.8538,
      "theoretical_loss": 3.733170434141224,
      "tokens_seen": 791377920
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003839418254764293,
      "loss": 2.9927,
      "theoretical_loss": 3.7331397292036366,
      "tokens_seen": 791443456
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003839317953861585,
      "loss": 2.9339,
      "theoretical_loss": 3.7331090275203307,
      "tokens_seen": 791508992
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038392176529588765,
      "loss": 2.7836,
      "theoretical_loss": 3.733078329090693,
      "tokens_seen": 791574528
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038391173520561684,
      "loss": 2.5344,
      "theoretical_loss": 3.733047633914109,
      "tokens_seen": 791640064
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038390170511534607,
      "loss": 3.0145,
      "theoretical_loss": 3.7330169419899653,
      "tokens_seen": 791705600
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003838916750250752,
      "loss": 2.9672,
      "theoretical_loss": 3.7329862533176463,
      "tokens_seen": 791771136
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038388164493480443,
      "loss": 2.8305,
      "theoretical_loss": 3.732955567896541,
      "tokens_seen": 791836672
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003838716148445336,
      "loss": 2.9355,
      "theoretical_loss": 3.7329248857260335,
      "tokens_seen": 791902208
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003838615847542628,
      "loss": 2.7786,
      "theoretical_loss": 3.7328942068055118,
      "tokens_seen": 791967744
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.000383851554663992,
      "loss": 2.8358,
      "theoretical_loss": 3.7328635311343623,
      "tokens_seen": 792033280
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038384152457372116,
      "loss": 2.6601,
      "theoretical_loss": 3.732832858711973,
      "tokens_seen": 792098816
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038383149448345034,
      "loss": 2.618,
      "theoretical_loss": 3.7328021895377295,
      "tokens_seen": 792164352
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003838214643931796,
      "loss": 2.9023,
      "theoretical_loss": 3.73277152361102,
      "tokens_seen": 792229888
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003838114343029087,
      "loss": 2.7241,
      "theoretical_loss": 3.732740860931232,
      "tokens_seen": 792295424
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038380140421263794,
      "loss": 2.7667,
      "theoretical_loss": 3.7327102014977527,
      "tokens_seen": 792360960
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038379137412236706,
      "loss": 2.6392,
      "theoretical_loss": 3.732679545309971,
      "tokens_seen": 792426496
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003837813440320963,
      "loss": 2.877,
      "theoretical_loss": 3.732648892367274,
      "tokens_seen": 792492032
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003837713139418255,
      "loss": 2.8724,
      "theoretical_loss": 3.7326182426690506,
      "tokens_seen": 792557568
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038376128385155466,
      "loss": 2.7809,
      "theoretical_loss": 3.732587596214688,
      "tokens_seen": 792623104
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038375125376128384,
      "loss": 2.858,
      "theoretical_loss": 3.732556953003576,
      "tokens_seen": 792688640
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003837412236710131,
      "loss": 2.7547,
      "theoretical_loss": 3.732526313035102,
      "tokens_seen": 792754176
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003837311935807422,
      "loss": 2.8265,
      "theoretical_loss": 3.7324956763086563,
      "tokens_seen": 792819712
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038372116349047144,
      "loss": 3.1426,
      "theoretical_loss": 3.732465042823627,
      "tokens_seen": 792885248
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038371113340020057,
      "loss": 2.7527,
      "theoretical_loss": 3.732434412579403,
      "tokens_seen": 792950784
    },
    {
      "epoch": 2.06,
      "objective/train/docs_used": 913766,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.490590810775757,
      "objective/train/theoretical_loss": 3.732419098672402,
      "objective/train/tokens_used": 813443552,
      "theoretical_loss": 3.732419098672402,
      "tokens_seen": 792983552
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003837011033099298,
      "loss": 2.7067,
      "theoretical_loss": 3.7324037855753742,
      "tokens_seen": 793016320
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.000383691073219659,
      "loss": 2.7344,
      "theoretical_loss": 3.73237316181093,
      "tokens_seen": 793081856
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038368104312938816,
      "loss": 2.8914,
      "theoretical_loss": 3.73234254128546,
      "tokens_seen": 793147392
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038367101303911735,
      "loss": 2.8415,
      "theoretical_loss": 3.7323119239983544,
      "tokens_seen": 793212928
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003836609829488465,
      "loss": 2.8043,
      "theoretical_loss": 3.732281309949002,
      "tokens_seen": 793278464
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003836509528585757,
      "loss": 2.803,
      "theoretical_loss": 3.732250699136795,
      "tokens_seen": 793344000
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038364092276830494,
      "loss": 2.7812,
      "theoretical_loss": 3.732220091561122,
      "tokens_seen": 793409536
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003836308926780341,
      "loss": 2.6567,
      "theoretical_loss": 3.7321894872213743,
      "tokens_seen": 793475072
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003836208625877633,
      "loss": 2.6744,
      "theoretical_loss": 3.7321588861169417,
      "tokens_seen": 793540608
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003836108324974925,
      "loss": 2.8148,
      "theoretical_loss": 3.7321282882472167,
      "tokens_seen": 793606144
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038360080240722167,
      "loss": 2.8026,
      "theoretical_loss": 3.7320976936115886,
      "tokens_seen": 793671680
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003835907723169509,
      "loss": 2.8677,
      "theoretical_loss": 3.7320671022094496,
      "tokens_seen": 793737216
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038358074222668003,
      "loss": 2.9286,
      "theoretical_loss": 3.7320365140401903,
      "tokens_seen": 793802752
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038357071213640927,
      "loss": 2.9299,
      "theoretical_loss": 3.7320059291032024,
      "tokens_seen": 793868288
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038356068204613845,
      "loss": 2.8521,
      "theoretical_loss": 3.7319753473978783,
      "tokens_seen": 793933824
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038355065195586763,
      "loss": 2.7029,
      "theoretical_loss": 3.7319447689236087,
      "tokens_seen": 793999360
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003835406218655968,
      "loss": 2.8717,
      "theoretical_loss": 3.7319141936797866,
      "tokens_seen": 794064896
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.000383530591775326,
      "loss": 2.5418,
      "theoretical_loss": 3.731883621665803,
      "tokens_seen": 794130432
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038352056168505517,
      "loss": 2.831,
      "theoretical_loss": 3.731853052881051,
      "tokens_seen": 794195968
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003835105315947844,
      "loss": 2.8505,
      "theoretical_loss": 3.7318224873249237,
      "tokens_seen": 794261504
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038350050150451353,
      "loss": 2.7467,
      "theoretical_loss": 3.731791924996812,
      "tokens_seen": 794327040
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038349047141424277,
      "loss": 2.6475,
      "theoretical_loss": 3.7317613658961104,
      "tokens_seen": 794392576
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003834804413239719,
      "loss": 2.829,
      "theoretical_loss": 3.731730810022211,
      "tokens_seen": 794458112
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038347041123370113,
      "loss": 2.832,
      "theoretical_loss": 3.7317002573745066,
      "tokens_seen": 794523648
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003834603811434303,
      "loss": 2.8858,
      "theoretical_loss": 3.731669707952391,
      "tokens_seen": 794589184
    },
    {
      "epoch": 2.06,
      "objective/train/docs_used": 914513,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7244105339050293,
      "objective/train/theoretical_loss": 3.7316544344507396,
      "objective/train/tokens_used": 815081952,
      "theoretical_loss": 3.7316544344507396,
      "tokens_seen": 794621952
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003834503510531595,
      "loss": 2.8052,
      "theoretical_loss": 3.731639161755258,
      "tokens_seen": 794654720
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003834403209628887,
      "loss": 2.9226,
      "theoretical_loss": 3.731608618782501,
      "tokens_seen": 794720256
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038343029087261786,
      "loss": 2.8461,
      "theoretical_loss": 3.731578079033513,
      "tokens_seen": 794785792
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038342026078234704,
      "loss": 2.7635,
      "theoretical_loss": 3.7315475425076894,
      "tokens_seen": 794851328
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038341023069207627,
      "loss": 2.7834,
      "theoretical_loss": 3.731517009204423,
      "tokens_seen": 794916864
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003834002006018054,
      "loss": 2.6976,
      "theoretical_loss": 3.7314864791231086,
      "tokens_seen": 794982400
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038339017051153463,
      "loss": 2.7814,
      "theoretical_loss": 3.7314559522631408,
      "tokens_seen": 795047936
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003833801404212638,
      "loss": 2.6001,
      "theoretical_loss": 3.7314254286239144,
      "tokens_seen": 795113472
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.000383370110330993,
      "loss": 2.9477,
      "theoretical_loss": 3.7313949082048232,
      "tokens_seen": 795179008
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003833600802407222,
      "loss": 2.9413,
      "theoretical_loss": 3.731364391005263,
      "tokens_seen": 795244544
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038335005015045136,
      "loss": 2.7242,
      "theoretical_loss": 3.731333877024629,
      "tokens_seen": 795310080
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038334002006018054,
      "loss": 2.7499,
      "theoretical_loss": 3.7313033662623156,
      "tokens_seen": 795375616
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003833299899699098,
      "loss": 2.7898,
      "theoretical_loss": 3.7312728587177193,
      "tokens_seen": 795441152
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003833199598796389,
      "loss": 2.7396,
      "theoretical_loss": 3.7312423543902344,
      "tokens_seen": 795506688
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038330992978936814,
      "loss": 2.7025,
      "theoretical_loss": 3.731211853279258,
      "tokens_seen": 795572224
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038329989969909726,
      "loss": 2.7248,
      "theoretical_loss": 3.7311813553841855,
      "tokens_seen": 795637760
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003832898696088265,
      "loss": 2.7345,
      "theoretical_loss": 3.731150860704413,
      "tokens_seen": 795703296
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003832798395185557,
      "loss": 2.8407,
      "theoretical_loss": 3.7311203692393358,
      "tokens_seen": 795768832
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038326980942828486,
      "loss": 2.9334,
      "theoretical_loss": 3.731089880988352,
      "tokens_seen": 795834368
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038325977933801404,
      "loss": 2.7804,
      "theoretical_loss": 3.731059395950857,
      "tokens_seen": 795899904
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003832497492477433,
      "loss": 2.7048,
      "theoretical_loss": 3.731028914126248,
      "tokens_seen": 795965440
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003832397191574724,
      "loss": 2.8948,
      "theoretical_loss": 3.7309984355139214,
      "tokens_seen": 796030976
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038322968906720164,
      "loss": 2.953,
      "theoretical_loss": 3.730967960113275,
      "tokens_seen": 796096512
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038321965897693077,
      "loss": 2.8011,
      "theoretical_loss": 3.730937487923705,
      "tokens_seen": 796162048
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038320962888666,
      "loss": 2.8264,
      "theoretical_loss": 3.7309070189446096,
      "tokens_seen": 796227584
    },
    {
      "epoch": 2.06,
      "objective/train/docs_used": 915896,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.109905481338501,
      "objective/train/theoretical_loss": 3.7308917856588018,
      "objective/train/tokens_used": 816720352,
      "theoretical_loss": 3.7308917856588018,
      "tokens_seen": 796260352
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003831995987963892,
      "loss": 2.8979,
      "theoretical_loss": 3.7308765531753867,
      "tokens_seen": 796293120
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038318956870611836,
      "loss": 2.908,
      "theoretical_loss": 3.730846090615433,
      "tokens_seen": 796358656
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038317953861584755,
      "loss": 2.7226,
      "theoretical_loss": 3.730815631264147,
      "tokens_seen": 796424192
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003831695085255767,
      "loss": 2.5663,
      "theoretical_loss": 3.7307851751209267,
      "tokens_seen": 796489728
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003831594784353059,
      "loss": 2.7183,
      "theoretical_loss": 3.7307547221851696,
      "tokens_seen": 796555264
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038314944834503514,
      "loss": 2.8574,
      "theoretical_loss": 3.730724272456275,
      "tokens_seen": 796620800
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038313941825476427,
      "loss": 2.8418,
      "theoretical_loss": 3.730693825933641,
      "tokens_seen": 796686336
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003831293881644935,
      "loss": 2.5915,
      "theoretical_loss": 3.7306633826166666,
      "tokens_seen": 796751872
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038311935807422263,
      "loss": 2.9564,
      "theoretical_loss": 3.73063294250475,
      "tokens_seen": 796817408
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038310932798395187,
      "loss": 2.8469,
      "theoretical_loss": 3.730602505597291,
      "tokens_seen": 796882944
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038309929789368105,
      "loss": 2.887,
      "theoretical_loss": 3.730572071893688,
      "tokens_seen": 796948480
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038308926780341023,
      "loss": 2.8746,
      "theoretical_loss": 3.7305416413933408,
      "tokens_seen": 797014016
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003830792377131394,
      "loss": 2.9179,
      "theoretical_loss": 3.7305112140956487,
      "tokens_seen": 797079552
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038306920762286865,
      "loss": 2.9636,
      "theoretical_loss": 3.730480790000011,
      "tokens_seen": 797145088
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003830591775325978,
      "loss": 2.7864,
      "theoretical_loss": 3.730450369105829,
      "tokens_seen": 797210624
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.000383049147442327,
      "loss": 3.0204,
      "theoretical_loss": 3.7304199514125007,
      "tokens_seen": 797276160
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038303911735205614,
      "loss": 2.8526,
      "theoretical_loss": 3.7303895369194273,
      "tokens_seen": 797341696
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038302908726178537,
      "loss": 2.7846,
      "theoretical_loss": 3.730359125626009,
      "tokens_seen": 797407232
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038301905717151455,
      "loss": 2.941,
      "theoretical_loss": 3.7303287175316466,
      "tokens_seen": 797472768
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038300902708124373,
      "loss": 2.7052,
      "theoretical_loss": 3.73029831263574,
      "tokens_seen": 797538304
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003829989969909729,
      "loss": 2.7305,
      "theoretical_loss": 3.73026791093769,
      "tokens_seen": 797603840
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003829889669007021,
      "loss": 2.7913,
      "theoretical_loss": 3.730237512436899,
      "tokens_seen": 797669376
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003829789368104313,
      "loss": 2.7654,
      "theoretical_loss": 3.730207117132766,
      "tokens_seen": 797734912
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003829689067201605,
      "loss": 2.7138,
      "theoretical_loss": 3.7301767250246938,
      "tokens_seen": 797800448
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038295887662988964,
      "loss": 2.9547,
      "theoretical_loss": 3.7301463361120835,
      "tokens_seen": 797865984
    },
    {
      "epoch": 2.06,
      "objective/train/docs_used": 916418,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8996529579162598,
      "objective/train/theoretical_loss": 3.730131142853889,
      "objective/train/tokens_used": 818358752,
      "theoretical_loss": 3.730131142853889,
      "tokens_seen": 797898752
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003829488465396189,
      "loss": 3.0152,
      "theoretical_loss": 3.7301159503943357,
      "tokens_seen": 797931520
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.000382938816449348,
      "loss": 3.0106,
      "theoretical_loss": 3.7300855678708533,
      "tokens_seen": 797997056
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038292878635907724,
      "loss": 2.8793,
      "theoretical_loss": 3.7300551885410385,
      "tokens_seen": 798062592
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003829187562688064,
      "loss": 2.7218,
      "theoretical_loss": 3.730024812404292,
      "tokens_seen": 798128128
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003829087261785356,
      "loss": 2.8625,
      "theoretical_loss": 3.7299944394600173,
      "tokens_seen": 798193664
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003828986960882648,
      "loss": 2.5771,
      "theoretical_loss": 3.729964069707616,
      "tokens_seen": 798259200
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.000382888665997994,
      "loss": 2.8149,
      "theoretical_loss": 3.729933703146491,
      "tokens_seen": 798324736
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003828786359077232,
      "loss": 2.8857,
      "theoretical_loss": 3.729903339776045,
      "tokens_seen": 798390272
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003828686058174524,
      "loss": 2.9546,
      "theoretical_loss": 3.7298729795956804,
      "tokens_seen": 798455808
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038285857572718156,
      "loss": 2.9196,
      "theoretical_loss": 3.729842622604801,
      "tokens_seen": 798521344
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038284854563691074,
      "loss": 2.921,
      "theoretical_loss": 3.72981226880281,
      "tokens_seen": 798586880
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038283851554664,
      "loss": 3.0057,
      "theoretical_loss": 3.7297819181891105,
      "tokens_seen": 798652416
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003828284854563691,
      "loss": 3.0165,
      "theoretical_loss": 3.7297515707631055,
      "tokens_seen": 798717952
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038281845536609834,
      "loss": 3.0633,
      "theoretical_loss": 3.7297212265242,
      "tokens_seen": 798783488
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038280842527582746,
      "loss": 2.8644,
      "theoretical_loss": 3.729690885471796,
      "tokens_seen": 798849024
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003827983951855567,
      "loss": 2.935,
      "theoretical_loss": 3.7296605476052993,
      "tokens_seen": 798914560
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003827883650952859,
      "loss": 2.9549,
      "theoretical_loss": 3.7296302129241132,
      "tokens_seen": 798980096
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038277833500501506,
      "loss": 2.798,
      "theoretical_loss": 3.729599881427642,
      "tokens_seen": 799045632
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038276830491474424,
      "loss": 2.9321,
      "theoretical_loss": 3.72956955311529,
      "tokens_seen": 799111168
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003827582748244735,
      "loss": 2.8404,
      "theoretical_loss": 3.7295392279864625,
      "tokens_seen": 799176704
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003827482447342026,
      "loss": 2.5635,
      "theoretical_loss": 3.729508906040564,
      "tokens_seen": 799242240
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038273821464393184,
      "loss": 2.7403,
      "theoretical_loss": 3.7294785872769993,
      "tokens_seen": 799307776
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038272818455366097,
      "loss": 2.9023,
      "theoretical_loss": 3.7294482716951736,
      "tokens_seen": 799373312
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003827181544633902,
      "loss": 2.5704,
      "theoretical_loss": 3.7294179592944925,
      "tokens_seen": 799438848
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003827081243731194,
      "loss": 2.8978,
      "theoretical_loss": 3.729387650074361,
      "tokens_seen": 799504384
    },
    {
      "epoch": 2.06,
      "objective/train/docs_used": 917678,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9319193363189697,
      "objective/train/theoretical_loss": 3.729372496656816,
      "objective/train/tokens_used": 819997152,
      "theoretical_loss": 3.729372496656816,
      "tokens_seen": 799537152
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038269809428284856,
      "loss": 2.9678,
      "theoretical_loss": 3.729357344034185,
      "tokens_seen": 799569920
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038268806419257775,
      "loss": 2.7522,
      "theoretical_loss": 3.72932704117337,
      "tokens_seen": 799635456
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003826780341023069,
      "loss": 2.9152,
      "theoretical_loss": 3.7292967414913223,
      "tokens_seen": 799700992
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003826680040120361,
      "loss": 2.5198,
      "theoretical_loss": 3.7292664449874477,
      "tokens_seen": 799766528
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038265797392176534,
      "loss": 2.8087,
      "theoretical_loss": 3.729236151661153,
      "tokens_seen": 799832064
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038264794383149447,
      "loss": 2.8923,
      "theoretical_loss": 3.7292058615118435,
      "tokens_seen": 799897600
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003826379137412237,
      "loss": 2.7683,
      "theoretical_loss": 3.729175574538927,
      "tokens_seen": 799963136
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038262788365095283,
      "loss": 2.7146,
      "theoretical_loss": 3.7291452907418092,
      "tokens_seen": 800028672
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038261785356068207,
      "loss": 2.9047,
      "theoretical_loss": 3.729115010119897,
      "tokens_seen": 800094208
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038260782347041125,
      "loss": 2.712,
      "theoretical_loss": 3.7290847326725984,
      "tokens_seen": 800159744
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038259779338014043,
      "loss": 2.7399,
      "theoretical_loss": 3.72905445839932,
      "tokens_seen": 800225280
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003825877632898696,
      "loss": 2.6716,
      "theoretical_loss": 3.729024187299469,
      "tokens_seen": 800290816
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038257773319959885,
      "loss": 2.878,
      "theoretical_loss": 3.7289939193724537,
      "tokens_seen": 800356352
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.000382567703109328,
      "loss": 2.7631,
      "theoretical_loss": 3.728963654617681,
      "tokens_seen": 800421888
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003825576730190572,
      "loss": 2.8419,
      "theoretical_loss": 3.7289333930345587,
      "tokens_seen": 800487424
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038254764292878634,
      "loss": 2.5932,
      "theoretical_loss": 3.7289031346224952,
      "tokens_seen": 800552960
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038253761283851557,
      "loss": 2.8904,
      "theoretical_loss": 3.7288728793808987,
      "tokens_seen": 800618496
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038252758274824475,
      "loss": 3.065,
      "theoretical_loss": 3.7288426273091773,
      "tokens_seen": 800684032
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038251755265797393,
      "loss": 2.7377,
      "theoretical_loss": 3.7288123784067393,
      "tokens_seen": 800749568
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003825075225677031,
      "loss": 2.8729,
      "theoretical_loss": 3.7287821326729933,
      "tokens_seen": 800815104
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003824974924774323,
      "loss": 2.7359,
      "theoretical_loss": 3.728751890107349,
      "tokens_seen": 800880640
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003824874623871615,
      "loss": 2.8713,
      "theoretical_loss": 3.7287216507092142,
      "tokens_seen": 800946176
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003824774322968907,
      "loss": 2.5688,
      "theoretical_loss": 3.728691414477999,
      "tokens_seen": 801011712
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038246740220661984,
      "loss": 2.7509,
      "theoretical_loss": 3.728661181413112,
      "tokens_seen": 801077248
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003824573721163491,
      "loss": 2.6961,
      "theoretical_loss": 3.7286309515139626,
      "tokens_seen": 801142784
    },
    {
      "epoch": 2.06,
      "objective/train/docs_used": 918390,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.938758134841919,
      "objective/train/theoretical_loss": 3.728615837751355,
      "objective/train/tokens_used": 821635552,
      "theoretical_loss": 3.728615837751355,
      "tokens_seen": 801175552
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003824473420260782,
      "loss": 3.0026,
      "theoretical_loss": 3.7286007247799606,
      "tokens_seen": 801208320
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038243731193580744,
      "loss": 2.6613,
      "theoretical_loss": 3.7285705012105157,
      "tokens_seen": 801273856
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003824272818455366,
      "loss": 2.668,
      "theoretical_loss": 3.7285402808050376,
      "tokens_seen": 801339392
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003824172517552658,
      "loss": 2.5081,
      "theoretical_loss": 3.728510063562937,
      "tokens_seen": 801404928
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.000382407221664995,
      "loss": 2.5455,
      "theoretical_loss": 3.7284798494836235,
      "tokens_seen": 801470464
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003823971915747242,
      "loss": 2.7998,
      "theoretical_loss": 3.7284496385665076,
      "tokens_seen": 801536000
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038238716148445334,
      "loss": 2.8151,
      "theoretical_loss": 3.728419430811,
      "tokens_seen": 801601536
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003823771313941826,
      "loss": 2.7684,
      "theoretical_loss": 3.7283892262165117,
      "tokens_seen": 801667072
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003823671013039117,
      "loss": 3.0378,
      "theoretical_loss": 3.7283590247824527,
      "tokens_seen": 801732608
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038235707121364094,
      "loss": 2.9177,
      "theoretical_loss": 3.7283288265082346,
      "tokens_seen": 801798144
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003823470411233701,
      "loss": 3.1085,
      "theoretical_loss": 3.7282986313932684,
      "tokens_seen": 801863680
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003823370110330993,
      "loss": 2.8762,
      "theoretical_loss": 3.728268439436966,
      "tokens_seen": 801929216
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003823269809428285,
      "loss": 2.7137,
      "theoretical_loss": 3.7282382506387375,
      "tokens_seen": 801994752
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038231695085255766,
      "loss": 2.524,
      "theoretical_loss": 3.728208064997996,
      "tokens_seen": 802060288
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038230692076228685,
      "loss": 2.8362,
      "theoretical_loss": 3.728177882514152,
      "tokens_seen": 802125824
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003822968906720161,
      "loss": 2.9109,
      "theoretical_loss": 3.728147703186619,
      "tokens_seen": 802191360
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003822868605817452,
      "loss": 2.7985,
      "theoretical_loss": 3.728117527014808,
      "tokens_seen": 802256896
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038227683049147444,
      "loss": 2.8394,
      "theoretical_loss": 3.7280873539981307,
      "tokens_seen": 802322432
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038226680040120357,
      "loss": 2.7245,
      "theoretical_loss": 3.7280571841360013,
      "tokens_seen": 802387968
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003822567703109328,
      "loss": 2.8553,
      "theoretical_loss": 3.7280270174278316,
      "tokens_seen": 802453504
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.000382246740220662,
      "loss": 2.6918,
      "theoretical_loss": 3.7279968538730333,
      "tokens_seen": 802519040
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038223671013039117,
      "loss": 2.6755,
      "theoretical_loss": 3.7279666934710205,
      "tokens_seen": 802584576
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038222668004012035,
      "loss": 2.7522,
      "theoretical_loss": 3.727936536221206,
      "tokens_seen": 802650112
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003822166499498496,
      "loss": 2.6767,
      "theoretical_loss": 3.7279063821230034,
      "tokens_seen": 802715648
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003822066198595787,
      "loss": 2.8416,
      "theoretical_loss": 3.727876231175825,
      "tokens_seen": 802781184
    },
    {
      "epoch": 2.06,
      "objective/train/docs_used": 919804,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.2665350437164307,
      "objective/train/theoretical_loss": 3.727861156883687,
      "objective/train/tokens_used": 823273952,
      "theoretical_loss": 3.727861156883687,
      "tokens_seen": 802813952
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038219658976930795,
      "loss": 2.9272,
      "theoretical_loss": 3.7278460833790854,
      "tokens_seen": 802846720
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003821865596790371,
      "loss": 2.8213,
      "theoretical_loss": 3.7278159387321974,
      "tokens_seen": 802912256
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003821765295887663,
      "loss": 2.7012,
      "theoretical_loss": 3.727785797234575,
      "tokens_seen": 802977792
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003821664994984955,
      "loss": 2.6509,
      "theoretical_loss": 3.7277556588856333,
      "tokens_seen": 803043328
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038215646940822467,
      "loss": 2.8218,
      "theoretical_loss": 3.7277255236847853,
      "tokens_seen": 803108864
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038214643931795385,
      "loss": 2.9506,
      "theoretical_loss": 3.7276953916314453,
      "tokens_seen": 803174400
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038213640922768303,
      "loss": 2.8434,
      "theoretical_loss": 3.727665262725029,
      "tokens_seen": 803239936
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038212637913741227,
      "loss": 2.9449,
      "theoretical_loss": 3.7276351369649494,
      "tokens_seen": 803305472
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038211634904714145,
      "loss": 2.6784,
      "theoretical_loss": 3.7276050143506225,
      "tokens_seen": 803371008
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038210631895687063,
      "loss": 2.9453,
      "theoretical_loss": 3.727574894881462,
      "tokens_seen": 803436544
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003820962888665998,
      "loss": 2.7106,
      "theoretical_loss": 3.7275447785568847,
      "tokens_seen": 803502080
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038208625877632905,
      "loss": 2.8271,
      "theoretical_loss": 3.7275146653763045,
      "tokens_seen": 803567616
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003820762286860582,
      "loss": 2.9294,
      "theoretical_loss": 3.727484555339137,
      "tokens_seen": 803633152
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003820661985957874,
      "loss": 2.5488,
      "theoretical_loss": 3.7274544484447985,
      "tokens_seen": 803698688
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038205616850551654,
      "loss": 2.9373,
      "theoretical_loss": 3.727424344692704,
      "tokens_seen": 803764224
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038204613841524577,
      "loss": 2.9778,
      "theoretical_loss": 3.7273942440822694,
      "tokens_seen": 803829760
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038203610832497495,
      "loss": 2.9941,
      "theoretical_loss": 3.727364146612911,
      "tokens_seen": 803895296
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038202607823470413,
      "loss": 2.6439,
      "theoretical_loss": 3.727334052284045,
      "tokens_seen": 803960832
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003820160481444333,
      "loss": 2.7207,
      "theoretical_loss": 3.7273039610950875,
      "tokens_seen": 804026368
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003820060180541625,
      "loss": 2.9206,
      "theoretical_loss": 3.7272738730454558,
      "tokens_seen": 804091904
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003819959879638917,
      "loss": 2.8644,
      "theoretical_loss": 3.7272437881345652,
      "tokens_seen": 804157440
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003819859578736209,
      "loss": 2.964,
      "theoretical_loss": 3.7272137063618334,
      "tokens_seen": 804222976
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038197592778335004,
      "loss": 2.8123,
      "theoretical_loss": 3.727183627726677,
      "tokens_seen": 804288512
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003819658976930793,
      "loss": 2.7005,
      "theoretical_loss": 3.727153552228514,
      "tokens_seen": 804354048
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003819558676028084,
      "loss": 3.1517,
      "theoretical_loss": 3.7271234798667603,
      "tokens_seen": 804419584
    },
    {
      "epoch": 2.06,
      "objective/train/docs_used": 920531,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.036098003387451,
      "objective/train/theoretical_loss": 3.727108444861855,
      "objective/train/tokens_used": 824912352,
      "theoretical_loss": 3.727108444861855,
      "tokens_seen": 804452352
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038194583751253764,
      "loss": 2.6692,
      "theoretical_loss": 3.727093410640835,
      "tokens_seen": 804485120
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003819358074222668,
      "loss": 2.885,
      "theoretical_loss": 3.727063344550154,
      "tokens_seen": 804550656
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.000381925777331996,
      "loss": 3.1506,
      "theoretical_loss": 3.727033281594136,
      "tokens_seen": 804616192
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003819157472417252,
      "loss": 2.9126,
      "theoretical_loss": 3.7270032217721987,
      "tokens_seen": 804681728
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003819057171514544,
      "loss": 2.7855,
      "theoretical_loss": 3.72697316508376,
      "tokens_seen": 804747264
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038189568706118354,
      "loss": 2.5874,
      "theoretical_loss": 3.7269431115282385,
      "tokens_seen": 804812800
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003818856569709128,
      "loss": 2.7009,
      "theoretical_loss": 3.7269130611050523,
      "tokens_seen": 804878336
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003818756268806419,
      "loss": 2.4843,
      "theoretical_loss": 3.72688301381362,
      "tokens_seen": 804943872
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038186559679037114,
      "loss": 2.7415,
      "theoretical_loss": 3.72685296965336,
      "tokens_seen": 805009408
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003818555667001003,
      "loss": 2.5227,
      "theoretical_loss": 3.7268229286236916,
      "tokens_seen": 805074944
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003818455366098295,
      "loss": 2.7426,
      "theoretical_loss": 3.7267928907240337,
      "tokens_seen": 805140480
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003818355065195587,
      "loss": 2.7953,
      "theoretical_loss": 3.7267628559538055,
      "tokens_seen": 805206016
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038182547642928786,
      "loss": 2.7316,
      "theoretical_loss": 3.7267328243124256,
      "tokens_seen": 805271552
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038181544633901705,
      "loss": 2.7126,
      "theoretical_loss": 3.726702795799314,
      "tokens_seen": 805337088
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003818054162487463,
      "loss": 2.8651,
      "theoretical_loss": 3.7266727704138907,
      "tokens_seen": 805402624
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003817953861584754,
      "loss": 2.7344,
      "theoretical_loss": 3.726642748155575,
      "tokens_seen": 805468160
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038178535606820464,
      "loss": 2.8418,
      "theoretical_loss": 3.7266127290237865,
      "tokens_seen": 805533696
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038177532597793377,
      "loss": 2.887,
      "theoretical_loss": 3.7265827130179465,
      "tokens_seen": 805599232
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.000381765295887663,
      "loss": 2.8043,
      "theoretical_loss": 3.7265527001374736,
      "tokens_seen": 805664768
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003817552657973922,
      "loss": 2.8358,
      "theoretical_loss": 3.7265226903817896,
      "tokens_seen": 805730304
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038174523570712137,
      "loss": 2.6254,
      "theoretical_loss": 3.726492683750314,
      "tokens_seen": 805795840
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038173520561685055,
      "loss": 2.6947,
      "theoretical_loss": 3.7264626802424683,
      "tokens_seen": 805861376
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003817251755265798,
      "loss": 2.6898,
      "theoretical_loss": 3.7264326798576723,
      "tokens_seen": 805926912
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003817151454363089,
      "loss": 2.7434,
      "theoretical_loss": 3.7264026825953485,
      "tokens_seen": 805992448
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038170511534603815,
      "loss": 2.557,
      "theoretical_loss": 3.7263726884549166,
      "tokens_seen": 806057984
    },
    {
      "epoch": 2.06,
      "objective/train/docs_used": 921184,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.061922788619995,
      "objective/train/theoretical_loss": 3.7263576925552298,
      "objective/train/tokens_used": 826550752,
      "theoretical_loss": 3.7263576925552298,
      "tokens_seen": 806090752
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003816950852557673,
      "loss": 2.9332,
      "theoretical_loss": 3.7263426974357987,
      "tokens_seen": 806123520
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003816850551654965,
      "loss": 2.8546,
      "theoretical_loss": 3.7263127095374164,
      "tokens_seen": 806189056
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003816750250752257,
      "loss": 2.9827,
      "theoretical_loss": 3.7262827247591908,
      "tokens_seen": 806254592
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038166499498495487,
      "loss": 2.7669,
      "theoretical_loss": 3.7262527431005443,
      "tokens_seen": 806320128
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038165496489468405,
      "loss": 2.9307,
      "theoretical_loss": 3.726222764560898,
      "tokens_seen": 806385664
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038164493480441323,
      "loss": 2.855,
      "theoretical_loss": 3.7261927891396747,
      "tokens_seen": 806451200
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003816349047141424,
      "loss": 2.6853,
      "theoretical_loss": 3.726162816836296,
      "tokens_seen": 806516736
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038162487462387165,
      "loss": 2.6587,
      "theoretical_loss": 3.7261328476501845,
      "tokens_seen": 806582272
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003816148445336008,
      "loss": 2.7978,
      "theoretical_loss": 3.7261028815807635,
      "tokens_seen": 806647808
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038160481444333,
      "loss": 2.8737,
      "theoretical_loss": 3.726072918627455,
      "tokens_seen": 806713344
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038159478435305914,
      "loss": 2.6942,
      "theoretical_loss": 3.7260429587896815,
      "tokens_seen": 806778880
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003815847542627884,
      "loss": 2.8407,
      "theoretical_loss": 3.7260130020668667,
      "tokens_seen": 806844416
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038157472417251756,
      "loss": 2.8455,
      "theoretical_loss": 3.725983048458434,
      "tokens_seen": 806909952
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038156469408224674,
      "loss": 2.7394,
      "theoretical_loss": 3.7259530979638056,
      "tokens_seen": 806975488
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003815546639919759,
      "loss": 2.8217,
      "theoretical_loss": 3.725923150582406,
      "tokens_seen": 807041024
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038154463390170515,
      "loss": 2.7258,
      "theoretical_loss": 3.7258932063136583,
      "tokens_seen": 807106560
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003815346038114343,
      "loss": 2.4661,
      "theoretical_loss": 3.7258632651569865,
      "tokens_seen": 807172096
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003815245737211635,
      "loss": 2.8683,
      "theoretical_loss": 3.725833327111814,
      "tokens_seen": 807237632
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038151454363089264,
      "loss": 2.3263,
      "theoretical_loss": 3.725803392177566,
      "tokens_seen": 807303168
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003815045135406219,
      "loss": 2.5954,
      "theoretical_loss": 3.7257734603536656,
      "tokens_seen": 807368704
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038149448345035106,
      "loss": 2.9207,
      "theoretical_loss": 3.7257435316395373,
      "tokens_seen": 807434240
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038148445336008024,
      "loss": 3.1429,
      "theoretical_loss": 3.7257136060346063,
      "tokens_seen": 807499776
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003814744232698094,
      "loss": 2.795,
      "theoretical_loss": 3.725683683538297,
      "tokens_seen": 807565312
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003814643931795386,
      "loss": 2.5561,
      "theoretical_loss": 3.7256537641500347,
      "tokens_seen": 807630848
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003814543630892678,
      "loss": 2.9045,
      "theoretical_loss": 3.725623847869244,
      "tokens_seen": 807696384
    },
    {
      "epoch": 2.06,
      "objective/train/docs_used": 922293,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.627511501312256,
      "objective/train/theoretical_loss": 3.72560889089397,
      "objective/train/tokens_used": 828189152,
      "theoretical_loss": 3.72560889089397,
      "tokens_seen": 807729152
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.000381444332998997,
      "loss": 2.818,
      "theoretical_loss": 3.725593934695349,
      "tokens_seen": 807761920
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038143430290872614,
      "loss": 2.9807,
      "theoretical_loss": 3.7255640246277766,
      "tokens_seen": 807827456
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.0003814242728184554,
      "loss": 2.9548,
      "theoretical_loss": 3.7255341176659513,
      "tokens_seen": 807892992
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038141424272818456,
      "loss": 2.9676,
      "theoretical_loss": 3.725504213809299,
      "tokens_seen": 807958528
    },
    {
      "epoch": 2.06,
      "learning_rate": 0.00038140421263791374,
      "loss": 2.7428,
      "theoretical_loss": 3.725474313057246,
      "tokens_seen": 808024064
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003813941825476429,
      "loss": 2.8616,
      "theoretical_loss": 3.7254444154092177,
      "tokens_seen": 808089600
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003813841524573721,
      "loss": 2.9295,
      "theoretical_loss": 3.72541452086464,
      "tokens_seen": 808155136
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00038137412236710134,
      "loss": 2.6852,
      "theoretical_loss": 3.7253846294229396,
      "tokens_seen": 808220672
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003813640922768305,
      "loss": 2.6936,
      "theoretical_loss": 3.7253547410835424,
      "tokens_seen": 808286208
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003813540621865597,
      "loss": 2.5662,
      "theoretical_loss": 3.725324855845875,
      "tokens_seen": 808351744
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003813440320962889,
      "loss": 2.8743,
      "theoretical_loss": 3.7252949737093646,
      "tokens_seen": 808417280
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00038133400200601806,
      "loss": 2.6674,
      "theoretical_loss": 3.7252650946734374,
      "tokens_seen": 808482816
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00038132397191574725,
      "loss": 2.8592,
      "theoretical_loss": 3.7252352187375206,
      "tokens_seen": 808548352
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003813139418254765,
      "loss": 2.9754,
      "theoretical_loss": 3.725205345901042,
      "tokens_seen": 808613888
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003813039117352056,
      "loss": 2.6084,
      "theoretical_loss": 3.725175476163427,
      "tokens_seen": 808679424
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00038129388164493484,
      "loss": 2.7187,
      "theoretical_loss": 3.7251456095241053,
      "tokens_seen": 808744960
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00038128385155466397,
      "loss": 2.8086,
      "theoretical_loss": 3.725115745982503,
      "tokens_seen": 808810496
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003812738214643932,
      "loss": 2.8193,
      "theoretical_loss": 3.725085885538049,
      "tokens_seen": 808876032
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003812637913741224,
      "loss": 2.8032,
      "theoretical_loss": 3.7250560281901706,
      "tokens_seen": 808941568
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00038125376128385157,
      "loss": 2.766,
      "theoretical_loss": 3.7250261739382955,
      "tokens_seen": 809007104
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00038124373119358075,
      "loss": 2.82,
      "theoretical_loss": 3.724996322781852,
      "tokens_seen": 809072640
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00038123370110331,
      "loss": 2.5854,
      "theoretical_loss": 3.7249664747202686,
      "tokens_seen": 809138176
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003812236710130391,
      "loss": 2.4659,
      "theoretical_loss": 3.7249366297529742,
      "tokens_seen": 809203712
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00038121364092276835,
      "loss": 2.7821,
      "theoretical_loss": 3.724906787879397,
      "tokens_seen": 809269248
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003812036108324975,
      "loss": 2.5499,
      "theoretical_loss": 3.7248769490989666,
      "tokens_seen": 809334784
    },
    {
      "epoch": 2.07,
      "objective/train/docs_used": 923007,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.85902738571167,
      "objective/train/theoretical_loss": 3.7248620308685023,
      "objective/train/tokens_used": 829827552,
      "theoretical_loss": 3.7248620308685023,
      "tokens_seen": 809367552
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003811935807422267,
      "loss": 2.8503,
      "theoretical_loss": 3.7248471134111103,
      "tokens_seen": 809400320
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003811835506519559,
      "loss": 2.7954,
      "theoretical_loss": 3.724817280815259,
      "tokens_seen": 809465856
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00038117352056168507,
      "loss": 2.7549,
      "theoretical_loss": 3.7247874513108403,
      "tokens_seen": 809531392
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00038116349047141425,
      "loss": 2.5741,
      "theoretical_loss": 3.7247576248972853,
      "tokens_seen": 809596928
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00038115346038114343,
      "loss": 2.7751,
      "theoretical_loss": 3.724727801574023,
      "tokens_seen": 809662464
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003811434302908726,
      "loss": 3.0212,
      "theoretical_loss": 3.724697981340482,
      "tokens_seen": 809728000
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00038113340020060185,
      "loss": 2.8603,
      "theoretical_loss": 3.7246681641960935,
      "tokens_seen": 809793536
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.000381123370110331,
      "loss": 2.8533,
      "theoretical_loss": 3.724638350140287,
      "tokens_seen": 809859072
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003811133400200602,
      "loss": 2.8063,
      "theoretical_loss": 3.7246085391724932,
      "tokens_seen": 809924608
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00038110330992978934,
      "loss": 2.7438,
      "theoretical_loss": 3.7245787312921417,
      "tokens_seen": 809990144
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003810932798395186,
      "loss": 2.7521,
      "theoretical_loss": 3.724548926498663,
      "tokens_seen": 810055680
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00038108324974924776,
      "loss": 2.7021,
      "theoretical_loss": 3.7245191247914886,
      "tokens_seen": 810121216
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00038107321965897694,
      "loss": 2.6705,
      "theoretical_loss": 3.7244893261700485,
      "tokens_seen": 810186752
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003810631895687061,
      "loss": 2.8608,
      "theoretical_loss": 3.7244595306337733,
      "tokens_seen": 810252288
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00038105315947843535,
      "loss": 2.8353,
      "theoretical_loss": 3.7244297381820948,
      "tokens_seen": 810317824
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003810431293881645,
      "loss": 2.9204,
      "theoretical_loss": 3.7243999488144444,
      "tokens_seen": 810383360
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003810330992978937,
      "loss": 3.0504,
      "theoretical_loss": 3.7243701625302528,
      "tokens_seen": 810448896
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00038102306920762284,
      "loss": 2.8244,
      "theoretical_loss": 3.724340379328952,
      "tokens_seen": 810514432
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003810130391173521,
      "loss": 2.8698,
      "theoretical_loss": 3.724310599209973,
      "tokens_seen": 810579968
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00038100300902708126,
      "loss": 2.9134,
      "theoretical_loss": 3.724280822172749,
      "tokens_seen": 810645504
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00038099297893681044,
      "loss": 2.5862,
      "theoretical_loss": 3.724251048216711,
      "tokens_seen": 810711040
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003809829488465396,
      "loss": 2.6886,
      "theoretical_loss": 3.7242212773412913,
      "tokens_seen": 810776576
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003809729187562688,
      "loss": 2.815,
      "theoretical_loss": 3.7241915095459213,
      "tokens_seen": 810842112
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.000380962888665998,
      "loss": 2.7922,
      "theoretical_loss": 3.724161744830035,
      "tokens_seen": 810907648
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003809528585757272,
      "loss": 2.9728,
      "theoretical_loss": 3.7241319831930646,
      "tokens_seen": 810973184
    },
    {
      "epoch": 2.07,
      "objective/train/docs_used": 924257,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.479335069656372,
      "objective/train/theoretical_loss": 3.724117103528995,
      "objective/train/tokens_used": 831465952,
      "theoretical_loss": 3.724117103528995,
      "tokens_seen": 811005952
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00038094282848545635,
      "loss": 2.5841,
      "theoretical_loss": 3.724102224634442,
      "tokens_seen": 811038720
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003809327983951856,
      "loss": 2.7625,
      "theoretical_loss": 3.7240724691536005,
      "tokens_seen": 811104256
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00038092276830491476,
      "loss": 2.6284,
      "theoretical_loss": 3.7240427167499734,
      "tokens_seen": 811169792
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00038091273821464394,
      "loss": 2.8429,
      "theoretical_loss": 3.724012967422994,
      "tokens_seen": 811235328
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003809027081243731,
      "loss": 2.6949,
      "theoretical_loss": 3.7239832211720953,
      "tokens_seen": 811300864
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003808926780341023,
      "loss": 2.8882,
      "theoretical_loss": 3.7239534779967105,
      "tokens_seen": 811366400
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003808826479438315,
      "loss": 2.8984,
      "theoretical_loss": 3.723923737896274,
      "tokens_seen": 811431936
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003808726178535607,
      "loss": 2.851,
      "theoretical_loss": 3.723894000870219,
      "tokens_seen": 811497472
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00038086258776328985,
      "loss": 2.8951,
      "theoretical_loss": 3.72386426691798,
      "tokens_seen": 811563008
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003808525576730191,
      "loss": 2.6595,
      "theoretical_loss": 3.72383453603899,
      "tokens_seen": 811628544
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003808425275827482,
      "loss": 2.7869,
      "theoretical_loss": 3.7238048082326847,
      "tokens_seen": 811694080
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00038083249749247745,
      "loss": 2.7446,
      "theoretical_loss": 3.7237750834984977,
      "tokens_seen": 811759616
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003808224674022066,
      "loss": 2.9705,
      "theoretical_loss": 3.7237453618358636,
      "tokens_seen": 811825152
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003808124373119358,
      "loss": 2.7296,
      "theoretical_loss": 3.723715643244217,
      "tokens_seen": 811890688
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.000380802407221665,
      "loss": 2.7897,
      "theoretical_loss": 3.7236859277229923,
      "tokens_seen": 811956224
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00038079237713139417,
      "loss": 2.8115,
      "theoretical_loss": 3.7236562152716255,
      "tokens_seen": 812021760
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00038078234704112335,
      "loss": 2.9943,
      "theoretical_loss": 3.723626505889551,
      "tokens_seen": 812087296
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003807723169508526,
      "loss": 2.7147,
      "theoretical_loss": 3.723596799576205,
      "tokens_seen": 812152832
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003807622868605817,
      "loss": 2.8725,
      "theoretical_loss": 3.7235670963310215,
      "tokens_seen": 812218368
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00038075225677031095,
      "loss": 2.9766,
      "theoretical_loss": 3.7235373961534375,
      "tokens_seen": 812283904
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00038074222668004013,
      "loss": 2.7214,
      "theoretical_loss": 3.7235076990428877,
      "tokens_seen": 812349440
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003807321965897693,
      "loss": 2.7914,
      "theoretical_loss": 3.7234780049988085,
      "tokens_seen": 812414976
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003807221664994985,
      "loss": 2.7851,
      "theoretical_loss": 3.7234483140206356,
      "tokens_seen": 812480512
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003807121364092277,
      "loss": 2.92,
      "theoretical_loss": 3.7234186261078057,
      "tokens_seen": 812546048
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00038070210631895685,
      "loss": 2.7906,
      "theoretical_loss": 3.7233889412597545,
      "tokens_seen": 812611584
    },
    {
      "epoch": 2.07,
      "objective/train/docs_used": 924620,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.1071670055389404,
      "objective/train/theoretical_loss": 3.723374099984845,
      "objective/train/tokens_used": 833104352,
      "theoretical_loss": 3.723374099984845,
      "tokens_seen": 812644352
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003806920762286861,
      "loss": 2.8908,
      "theoretical_loss": 3.7233592594759193,
      "tokens_seen": 812677120
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003806820461384152,
      "loss": 2.8797,
      "theoretical_loss": 3.7233295807557356,
      "tokens_seen": 812742656
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00038067201604814445,
      "loss": 2.65,
      "theoretical_loss": 3.7232999050986413,
      "tokens_seen": 812808192
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003806619859578736,
      "loss": 2.9167,
      "theoretical_loss": 3.7232702325040727,
      "tokens_seen": 812873728
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003806519558676028,
      "loss": 2.9298,
      "theoretical_loss": 3.7232405629714673,
      "tokens_seen": 812939264
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.000380641925777332,
      "loss": 2.8462,
      "theoretical_loss": 3.7232108965002615,
      "tokens_seen": 813004800
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003806318956870612,
      "loss": 2.7202,
      "theoretical_loss": 3.7231812330898935,
      "tokens_seen": 813070336
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003806218655967904,
      "loss": 2.8769,
      "theoretical_loss": 3.7231515727398006,
      "tokens_seen": 813135872
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00038061183550651954,
      "loss": 2.5064,
      "theoretical_loss": 3.72312191544942,
      "tokens_seen": 813201408
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003806018054162488,
      "loss": 2.8703,
      "theoretical_loss": 3.7230922612181905,
      "tokens_seen": 813266944
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00038059177532597796,
      "loss": 2.68,
      "theoretical_loss": 3.7230626100455497,
      "tokens_seen": 813332480
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00038058174523570714,
      "loss": 2.6174,
      "theoretical_loss": 3.723032961930935,
      "tokens_seen": 813398016
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003805717151454363,
      "loss": 2.8521,
      "theoretical_loss": 3.7230033168737853,
      "tokens_seen": 813463552
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00038056168505516555,
      "loss": 2.9735,
      "theoretical_loss": 3.722973674873539,
      "tokens_seen": 813529088
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003805516549648947,
      "loss": 2.9433,
      "theoretical_loss": 3.7229440359296344,
      "tokens_seen": 813594624
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003805416248746239,
      "loss": 2.9525,
      "theoretical_loss": 3.7229144000415104,
      "tokens_seen": 813660160
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00038053159478435304,
      "loss": 3.0126,
      "theoretical_loss": 3.722884767208606,
      "tokens_seen": 813725696
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003805215646940823,
      "loss": 2.7167,
      "theoretical_loss": 3.7228551374303596,
      "tokens_seen": 813791232
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00038051153460381146,
      "loss": 2.8659,
      "theoretical_loss": 3.7228255107062114,
      "tokens_seen": 813856768
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00038050150451354064,
      "loss": 2.7003,
      "theoretical_loss": 3.7227958870356,
      "tokens_seen": 813922304
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003804914744232698,
      "loss": 2.6967,
      "theoretical_loss": 3.722766266417965,
      "tokens_seen": 813987840
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.000380481444332999,
      "loss": 2.6027,
      "theoretical_loss": 3.7227366488527456,
      "tokens_seen": 814053376
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003804714142427282,
      "loss": 2.7122,
      "theoretical_loss": 3.722707034339382,
      "tokens_seen": 814118912
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003804613841524574,
      "loss": 2.9424,
      "theoretical_loss": 3.7226774228773145,
      "tokens_seen": 814184448
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00038045135406218655,
      "loss": 2.5289,
      "theoretical_loss": 3.722647814465982,
      "tokens_seen": 814249984
    },
    {
      "epoch": 2.07,
      "objective/train/docs_used": 925886,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.111829996109009,
      "objective/train/theoretical_loss": 3.7226330114041666,
      "objective/train/tokens_used": 834742752,
      "theoretical_loss": 3.7226330114041666,
      "tokens_seen": 814282752
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003804413239719158,
      "loss": 2.7899,
      "theoretical_loss": 3.7226182091048257,
      "tokens_seen": 814315520
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00038043129388164496,
      "loss": 2.6081,
      "theoretical_loss": 3.7225886067932854,
      "tokens_seen": 814381056
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00038042126379137414,
      "loss": 2.8281,
      "theoretical_loss": 3.722559007530802,
      "tokens_seen": 814446592
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003804112337011033,
      "loss": 2.8378,
      "theoretical_loss": 3.722529411316816,
      "tokens_seen": 814512128
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003804012036108325,
      "loss": 2.812,
      "theoretical_loss": 3.7224998181507676,
      "tokens_seen": 814577664
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003803911735205617,
      "loss": 3.027,
      "theoretical_loss": 3.7224702280320985,
      "tokens_seen": 814643200
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003803811434302909,
      "loss": 2.8126,
      "theoretical_loss": 3.722440640960249,
      "tokens_seen": 814708736
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00038037111334002005,
      "loss": 2.8249,
      "theoretical_loss": 3.7224110569346616,
      "tokens_seen": 814774272
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003803610832497493,
      "loss": 2.6414,
      "theoretical_loss": 3.7223814759547764,
      "tokens_seen": 814839808
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003803510531594784,
      "loss": 2.712,
      "theoretical_loss": 3.722351898020036,
      "tokens_seen": 814905344
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00038034102306920765,
      "loss": 2.7911,
      "theoretical_loss": 3.7223223231298808,
      "tokens_seen": 814970880
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00038033099297893683,
      "loss": 2.5521,
      "theoretical_loss": 3.722292751283754,
      "tokens_seen": 815036416
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.000380320962888666,
      "loss": 2.783,
      "theoretical_loss": 3.722263182481096,
      "tokens_seen": 815101952
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003803109327983952,
      "loss": 2.7707,
      "theoretical_loss": 3.72223361672135,
      "tokens_seen": 815167488
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00038030090270812437,
      "loss": 2.722,
      "theoretical_loss": 3.7222040540039583,
      "tokens_seen": 815233024
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00038029087261785355,
      "loss": 2.8075,
      "theoretical_loss": 3.722174494328363,
      "tokens_seen": 815298560
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003802808425275828,
      "loss": 2.805,
      "theoretical_loss": 3.7221449376940066,
      "tokens_seen": 815364096
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003802708124373119,
      "loss": 2.7188,
      "theoretical_loss": 3.7221153841003316,
      "tokens_seen": 815429632
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00038026078234704115,
      "loss": 2.6442,
      "theoretical_loss": 3.7220858335467817,
      "tokens_seen": 815495168
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00038025075225677033,
      "loss": 2.8672,
      "theoretical_loss": 3.722056286032799,
      "tokens_seen": 815560704
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003802407221664995,
      "loss": 2.8103,
      "theoretical_loss": 3.7220267415578268,
      "tokens_seen": 815626240
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003802306920762287,
      "loss": 2.626,
      "theoretical_loss": 3.721997200121309,
      "tokens_seen": 815691776
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003802206619859579,
      "loss": 2.7419,
      "theoretical_loss": 3.7219676617226876,
      "tokens_seen": 815757312
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00038021063189568705,
      "loss": 2.554,
      "theoretical_loss": 3.721938126361408,
      "tokens_seen": 815822848
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003802006018054163,
      "loss": 2.4705,
      "theoretical_loss": 3.721908594036913,
      "tokens_seen": 815888384
    },
    {
      "epoch": 2.07,
      "objective/train/docs_used": 927338,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.0634372234344482,
      "objective/train/theoretical_loss": 3.7218938290132857,
      "objective/train/tokens_used": 836381152,
      "theoretical_loss": 3.7218938290132857,
      "tokens_seen": 815921152
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003801905717151454,
      "loss": 2.8934,
      "theoretical_loss": 3.721879064748646,
      "tokens_seen": 815953920
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00038018054162487465,
      "loss": 2.9032,
      "theoretical_loss": 3.721849538496052,
      "tokens_seen": 816019456
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003801705115346038,
      "loss": 2.8216,
      "theoretical_loss": 3.721820015278575,
      "tokens_seen": 816084992
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.000380160481444333,
      "loss": 2.7976,
      "theoretical_loss": 3.721790495095658,
      "tokens_seen": 816150528
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003801504513540622,
      "loss": 2.4449,
      "theoretical_loss": 3.7217609779467473,
      "tokens_seen": 816216064
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003801404212637914,
      "loss": 2.8176,
      "theoretical_loss": 3.7217314638312864,
      "tokens_seen": 816281600
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00038013039117352056,
      "loss": 2.4909,
      "theoretical_loss": 3.7217019527487203,
      "tokens_seen": 816347136
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00038012036108324974,
      "loss": 2.7071,
      "theoretical_loss": 3.7216724446984943,
      "tokens_seen": 816412672
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003801103309929789,
      "loss": 2.6279,
      "theoretical_loss": 3.7216429396800526,
      "tokens_seen": 816478208
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00038010030090270816,
      "loss": 2.7475,
      "theoretical_loss": 3.721613437692841,
      "tokens_seen": 816543744
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003800902708124373,
      "loss": 2.8536,
      "theoretical_loss": 3.7215839387363046,
      "tokens_seen": 816609280
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003800802407221665,
      "loss": 2.5333,
      "theoretical_loss": 3.7215544428098895,
      "tokens_seen": 816674816
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003800702106318957,
      "loss": 2.5012,
      "theoretical_loss": 3.7215249499130403,
      "tokens_seen": 816740352
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003800601805416249,
      "loss": 2.8211,
      "theoretical_loss": 3.721495460045203,
      "tokens_seen": 816805888
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00038005015045135406,
      "loss": 2.6342,
      "theoretical_loss": 3.7214659732058246,
      "tokens_seen": 816871424
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00038004012036108324,
      "loss": 2.7294,
      "theoretical_loss": 3.7214364893943497,
      "tokens_seen": 816936960
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003800300902708124,
      "loss": 2.7032,
      "theoretical_loss": 3.721407008610225,
      "tokens_seen": 817002496
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00038002006018054166,
      "loss": 2.947,
      "theoretical_loss": 3.7213775308528976,
      "tokens_seen": 817068032
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003800100300902708,
      "loss": 2.8428,
      "theoretical_loss": 3.721348056121813,
      "tokens_seen": 817133568
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00038,
      "loss": 2.6575,
      "theoretical_loss": 3.7213185844164185,
      "tokens_seen": 817199104
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037998996990972915,
      "loss": 2.7767,
      "theoretical_loss": 3.7212891157361607,
      "tokens_seen": 817264640
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003799799398194584,
      "loss": 2.8218,
      "theoretical_loss": 3.721259650080486,
      "tokens_seen": 817330176
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037996990972918756,
      "loss": 2.652,
      "theoretical_loss": 3.7212301874488425,
      "tokens_seen": 817395712
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037995987963891675,
      "loss": 2.7455,
      "theoretical_loss": 3.7212007278406767,
      "tokens_seen": 817461248
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003799498495486459,
      "loss": 2.5604,
      "theoretical_loss": 3.721171271255436,
      "tokens_seen": 817526784
    },
    {
      "epoch": 2.07,
      "objective/train/docs_used": 928048,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.899970769882202,
      "objective/train/theoretical_loss": 3.7211565440962397,
      "objective/train/tokens_used": 838019552,
      "theoretical_loss": 3.7211565440962397,
      "tokens_seen": 817559552
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037993981945837516,
      "loss": 2.7355,
      "theoretical_loss": 3.721141817692568,
      "tokens_seen": 817592320
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003799297893681043,
      "loss": 2.7331,
      "theoretical_loss": 3.7211123671515214,
      "tokens_seen": 817657856
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003799197592778335,
      "loss": 2.8934,
      "theoretical_loss": 3.7210829196317423,
      "tokens_seen": 817723392
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037990972918756265,
      "loss": 2.9739,
      "theoretical_loss": 3.721053475132679,
      "tokens_seen": 817788928
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003798996990972919,
      "loss": 2.5838,
      "theoretical_loss": 3.7210240336537805,
      "tokens_seen": 817854464
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037988966900702107,
      "loss": 2.8182,
      "theoretical_loss": 3.7209945951944943,
      "tokens_seen": 817920000
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037987963891675025,
      "loss": 2.7168,
      "theoretical_loss": 3.7209651597542694,
      "tokens_seen": 817985536
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003798696088264795,
      "loss": 2.8628,
      "theoretical_loss": 3.7209357273325536,
      "tokens_seen": 818051072
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003798595787362086,
      "loss": 2.9665,
      "theoretical_loss": 3.7209062979287966,
      "tokens_seen": 818116608
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037984954864593785,
      "loss": 2.6236,
      "theoretical_loss": 3.720876871542446,
      "tokens_seen": 818182144
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037983951855566703,
      "loss": 2.7174,
      "theoretical_loss": 3.720847448172951,
      "tokens_seen": 818247680
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003798294884653962,
      "loss": 2.6124,
      "theoretical_loss": 3.720818027819762,
      "tokens_seen": 818313216
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003798194583751254,
      "loss": 3.0636,
      "theoretical_loss": 3.7207886104823267,
      "tokens_seen": 818378752
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037980942828485457,
      "loss": 2.8202,
      "theoretical_loss": 3.7207591961600954,
      "tokens_seen": 818444288
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037979939819458375,
      "loss": 2.7901,
      "theoretical_loss": 3.720729784852517,
      "tokens_seen": 818509824
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.000379789368104313,
      "loss": 2.6593,
      "theoretical_loss": 3.720700376559042,
      "tokens_seen": 818575360
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003797793380140421,
      "loss": 2.7329,
      "theoretical_loss": 3.7206709712791195,
      "tokens_seen": 818640896
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037976930792377135,
      "loss": 2.5465,
      "theoretical_loss": 3.7206415690121997,
      "tokens_seen": 818706432
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037975927783350053,
      "loss": 2.8261,
      "theoretical_loss": 3.7206121697577332,
      "tokens_seen": 818771968
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003797492477432297,
      "loss": 2.6284,
      "theoretical_loss": 3.7205827735151695,
      "tokens_seen": 818837504
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003797392176529589,
      "loss": 2.6674,
      "theoretical_loss": 3.7205533802839597,
      "tokens_seen": 818903040
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003797291875626881,
      "loss": 2.696,
      "theoretical_loss": 3.720523990063554,
      "tokens_seen": 818968576
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037971915747241725,
      "loss": 2.8248,
      "theoretical_loss": 3.720494602853403,
      "tokens_seen": 819034112
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003797091273821465,
      "loss": 2.8485,
      "theoretical_loss": 3.7204652186529574,
      "tokens_seen": 819099648
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003796990972918756,
      "loss": 2.8094,
      "theoretical_loss": 3.720435837461669,
      "tokens_seen": 819165184
    },
    {
      "debugging/Self-BLEU-5": 0.5761146013011956,
      "debugging/distinct-1-grams": 0.7462239175050402,
      "debugging/distinct-2-grams": 0.9529290102306198,
      "debugging/entropy-1-grams": 6.250000994807308,
      "debugging/entropy-2-grams": 7.4283795811873965,
      "debugging/length": 523.9565217391304,
      "debugging/num_segments": 23,
      "debugging/score": 0.003311065605696948,
      "debugging/score_std": 0.0050287850561047285,
      "epoch": 2.07,
      "objective/train/docs_used": 929355,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8574650287628174,
      "objective/train/theoretical_loss": 3.720421147994287,
      "objective/train/tokens_used": 839657952,
      "theoretical_loss": 3.720421147994287,
      "tokens_seen": 819197952
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037968906720160485,
      "loss": 2.6181,
      "theoretical_loss": 3.720406459278988,
      "tokens_seen": 819230720
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.000379679037111334,
      "loss": 2.6622,
      "theoretical_loss": 3.720377084104366,
      "tokens_seen": 819296256
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003796690070210632,
      "loss": 2.7908,
      "theoretical_loss": 3.720347711937255,
      "tokens_seen": 819361792
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003796589769307924,
      "loss": 2.7659,
      "theoretical_loss": 3.720318342777106,
      "tokens_seen": 819427328
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003796489468405216,
      "loss": 2.5301,
      "theoretical_loss": 3.7202889766233707,
      "tokens_seen": 819492864
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037963891675025076,
      "loss": 2.7743,
      "theoretical_loss": 3.720259613475501,
      "tokens_seen": 819558400
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037962888665997994,
      "loss": 2.6859,
      "theoretical_loss": 3.720230253332949,
      "tokens_seen": 819623936
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003796188565697091,
      "loss": 2.9187,
      "theoretical_loss": 3.7202008961951667,
      "tokens_seen": 819689472
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037960882647943836,
      "loss": 2.6287,
      "theoretical_loss": 3.720171542061607,
      "tokens_seen": 819755008
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003795987963891675,
      "loss": 2.8665,
      "theoretical_loss": 3.720142190931721,
      "tokens_seen": 819820544
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003795887662988967,
      "loss": 2.4396,
      "theoretical_loss": 3.720112842804962,
      "tokens_seen": 819886080
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003795787362086259,
      "loss": 2.5603,
      "theoretical_loss": 3.7200834976807835,
      "tokens_seen": 819951616
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003795687061183551,
      "loss": 2.7133,
      "theoretical_loss": 3.7200541555586373,
      "tokens_seen": 820017152
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037955867602808426,
      "loss": 2.7159,
      "theoretical_loss": 3.720024816437977,
      "tokens_seen": 820082688
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037954864593781344,
      "loss": 2.5554,
      "theoretical_loss": 3.7199954803182553,
      "tokens_seen": 820148224
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003795386158475426,
      "loss": 2.7215,
      "theoretical_loss": 3.7199661471989254,
      "tokens_seen": 820213760
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037952858575727186,
      "loss": 2.7678,
      "theoretical_loss": 3.7199368170794416,
      "tokens_seen": 820279296
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.000379518555667001,
      "loss": 2.7882,
      "theoretical_loss": 3.7199074899592564,
      "tokens_seen": 820344832
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003795085255767302,
      "loss": 2.8856,
      "theoretical_loss": 3.7198781658378244,
      "tokens_seen": 820410368
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037949849548645935,
      "loss": 2.8051,
      "theoretical_loss": 3.719848844714599,
      "tokens_seen": 820475904
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003794884653961886,
      "loss": 2.5591,
      "theoretical_loss": 3.719819526589034,
      "tokens_seen": 820541440
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037947843530591776,
      "loss": 2.6928,
      "theoretical_loss": 3.719790211460584,
      "tokens_seen": 820606976
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037946840521564695,
      "loss": 2.7898,
      "theoretical_loss": 3.719760899328703,
      "tokens_seen": 820672512
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003794583751253761,
      "loss": 2.6418,
      "theoretical_loss": 3.719731590192846,
      "tokens_seen": 820738048
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037944834503510536,
      "loss": 2.6323,
      "theoretical_loss": 3.7197022840524667,
      "tokens_seen": 820803584
    },
    {
      "epoch": 2.07,
      "objective/train/docs_used": 929749,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.016310930252075,
      "objective/train/theoretical_loss": 3.719687632105411,
      "objective/train/tokens_used": 841296352,
      "theoretical_loss": 3.719687632105411,
      "tokens_seen": 820836352
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003794383149448345,
      "loss": 2.5313,
      "theoretical_loss": 3.7196729809070206,
      "tokens_seen": 820869120
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003794282848545637,
      "loss": 2.6753,
      "theoretical_loss": 3.7196436807559623,
      "tokens_seen": 820934656
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037941825476429285,
      "loss": 2.6919,
      "theoretical_loss": 3.719614383598746,
      "tokens_seen": 821000192
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003794082246740221,
      "loss": 3.0039,
      "theoretical_loss": 3.719585089434828,
      "tokens_seen": 821065728
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037939819458375127,
      "loss": 2.8825,
      "theoretical_loss": 3.719555798263663,
      "tokens_seen": 821131264
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037938816449348045,
      "loss": 2.9654,
      "theoretical_loss": 3.719526510084707,
      "tokens_seen": 821196800
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037937813440320963,
      "loss": 2.8574,
      "theoretical_loss": 3.719497224897415,
      "tokens_seen": 821262336
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003793681043129388,
      "loss": 2.7252,
      "theoretical_loss": 3.719467942701243,
      "tokens_seen": 821327872
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.000379358074222668,
      "loss": 2.876,
      "theoretical_loss": 3.7194386634956462,
      "tokens_seen": 821393408
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037934804413239723,
      "loss": 2.9689,
      "theoretical_loss": 3.7194093872800815,
      "tokens_seen": 821458944
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037933801404212635,
      "loss": 2.6865,
      "theoretical_loss": 3.719380114054005,
      "tokens_seen": 821524480
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003793279839518556,
      "loss": 2.6635,
      "theoretical_loss": 3.7193508438168728,
      "tokens_seen": 821590016
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003793179538615847,
      "loss": 2.9359,
      "theoretical_loss": 3.7193215765681407,
      "tokens_seen": 821655552
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037930792377131395,
      "loss": 2.6683,
      "theoretical_loss": 3.7192923123072656,
      "tokens_seen": 821721088
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037929789368104313,
      "loss": 2.7919,
      "theoretical_loss": 3.719263051033705,
      "tokens_seen": 821786624
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003792878635907723,
      "loss": 2.549,
      "theoretical_loss": 3.7192337927469143,
      "tokens_seen": 821852160
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003792778335005015,
      "loss": 2.543,
      "theoretical_loss": 3.7192045374463527,
      "tokens_seen": 821917696
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037926780341023073,
      "loss": 2.7604,
      "theoretical_loss": 3.7191752851314748,
      "tokens_seen": 821983232
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037925777331995986,
      "loss": 2.9516,
      "theoretical_loss": 3.7191460358017396,
      "tokens_seen": 822048768
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003792477432296891,
      "loss": 2.7581,
      "theoretical_loss": 3.719116789456604,
      "tokens_seen": 822114304
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003792377131394182,
      "loss": 2.814,
      "theoretical_loss": 3.7190875460955257,
      "tokens_seen": 822179840
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037922768304914746,
      "loss": 2.5764,
      "theoretical_loss": 3.719058305717962,
      "tokens_seen": 822245376
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037921765295887664,
      "loss": 2.7264,
      "theoretical_loss": 3.7190290683233713,
      "tokens_seen": 822310912
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003792076228686058,
      "loss": 2.7514,
      "theoretical_loss": 3.7189998339112114,
      "tokens_seen": 822376448
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.000379197592778335,
      "loss": 2.6486,
      "theoretical_loss": 3.7189706024809404,
      "tokens_seen": 822441984
    },
    {
      "epoch": 2.07,
      "objective/train/docs_used": 930378,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.722386360168457,
      "objective/train/theoretical_loss": 3.7189559878838434,
      "objective/train/tokens_used": 842934752,
      "theoretical_loss": 3.7189559878838434,
      "tokens_seen": 822474752
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003791875626880642,
      "loss": 2.8354,
      "theoretical_loss": 3.718941374032016,
      "tokens_seen": 822507520
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037917753259779336,
      "loss": 3.0609,
      "theoretical_loss": 3.7189121485638976,
      "tokens_seen": 822573056
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003791675025075226,
      "loss": 2.7041,
      "theoretical_loss": 3.718882926076043,
      "tokens_seen": 822638592
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003791574724172517,
      "loss": 2.865,
      "theoretical_loss": 3.7188537065679115,
      "tokens_seen": 822704128
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037914744232698096,
      "loss": 3.0404,
      "theoretical_loss": 3.7188244900389615,
      "tokens_seen": 822769664
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003791374122367101,
      "loss": 2.7809,
      "theoretical_loss": 3.718795276488652,
      "tokens_seen": 822835200
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003791273821464393,
      "loss": 2.6324,
      "theoretical_loss": 3.7187660659164425,
      "tokens_seen": 822900736
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037911735205616856,
      "loss": 3.104,
      "theoretical_loss": 3.718736858321792,
      "tokens_seen": 822966272
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003791073219658977,
      "loss": 2.9184,
      "theoretical_loss": 3.7187076537041595,
      "tokens_seen": 823031808
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003790972918756269,
      "loss": 2.9075,
      "theoretical_loss": 3.7186784520630054,
      "tokens_seen": 823097344
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003790872617853561,
      "loss": 2.9344,
      "theoretical_loss": 3.718649253397789,
      "tokens_seen": 823162880
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003790772316950853,
      "loss": 2.6255,
      "theoretical_loss": 3.7186200577079704,
      "tokens_seen": 823228416
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037906720160481446,
      "loss": 2.7619,
      "theoretical_loss": 3.7185908649930086,
      "tokens_seen": 823293952
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037905717151454364,
      "loss": 2.7724,
      "theoretical_loss": 3.7185616752523645,
      "tokens_seen": 823359488
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003790471414242728,
      "loss": 2.7481,
      "theoretical_loss": 3.7185324884854984,
      "tokens_seen": 823425024
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037903711133400206,
      "loss": 2.8423,
      "theoretical_loss": 3.718503304691871,
      "tokens_seen": 823490560
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003790270812437312,
      "loss": 2.6441,
      "theoretical_loss": 3.7184741238709416,
      "tokens_seen": 823556096
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003790170511534604,
      "loss": 2.6508,
      "theoretical_loss": 3.718444946022172,
      "tokens_seen": 823621632
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037900702106318955,
      "loss": 2.8063,
      "theoretical_loss": 3.7184157711450228,
      "tokens_seen": 823687168
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003789969909729188,
      "loss": 2.9624,
      "theoretical_loss": 3.718386599238954,
      "tokens_seen": 823752704
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037898696088264796,
      "loss": 2.7201,
      "theoretical_loss": 3.7183574303034286,
      "tokens_seen": 823818240
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037897693079237715,
      "loss": 2.7259,
      "theoretical_loss": 3.718328264337906,
      "tokens_seen": 823883776
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003789669007021063,
      "loss": 2.824,
      "theoretical_loss": 3.7182991013418483,
      "tokens_seen": 823949312
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037895687061183556,
      "loss": 2.9522,
      "theoretical_loss": 3.718269941314717,
      "tokens_seen": 824014848
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003789468405215647,
      "loss": 2.7737,
      "theoretical_loss": 3.7182407842559746,
      "tokens_seen": 824080384
    },
    {
      "epoch": 2.07,
      "objective/train/docs_used": 931616,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7472262382507324,
      "objective/train/theoretical_loss": 3.71822620683958,
      "objective/train/tokens_used": 844573152,
      "theoretical_loss": 3.71822620683958,
      "tokens_seen": 824113152
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003789368104312939,
      "loss": 2.5734,
      "theoretical_loss": 3.718211630165081,
      "tokens_seen": 824145920
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037892678034102305,
      "loss": 2.7369,
      "theoretical_loss": 3.7181824790415003,
      "tokens_seen": 824211456
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003789167502507523,
      "loss": 2.887,
      "theoretical_loss": 3.718153330884693,
      "tokens_seen": 824276992
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037890672016048147,
      "loss": 2.9061,
      "theoretical_loss": 3.718124185694122,
      "tokens_seen": 824342528
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037889669007021065,
      "loss": 3.0517,
      "theoretical_loss": 3.7180950434692495,
      "tokens_seen": 824408064
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037888665997993983,
      "loss": 2.7274,
      "theoretical_loss": 3.7180659042095376,
      "tokens_seen": 824473600
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.000378876629889669,
      "loss": 2.8292,
      "theoretical_loss": 3.7180367679144495,
      "tokens_seen": 824539136
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003788665997993982,
      "loss": 2.8566,
      "theoretical_loss": 3.718007634583448,
      "tokens_seen": 824604672
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037885656970912743,
      "loss": 2.6371,
      "theoretical_loss": 3.7179785042159956,
      "tokens_seen": 824670208
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037884653961885655,
      "loss": 2.8622,
      "theoretical_loss": 3.717949376811556,
      "tokens_seen": 824735744
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003788365095285858,
      "loss": 2.69,
      "theoretical_loss": 3.7179202523695913,
      "tokens_seen": 824801280
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003788264794383149,
      "loss": 2.7501,
      "theoretical_loss": 3.7178911308895666,
      "tokens_seen": 824866816
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037881644934804415,
      "loss": 2.8124,
      "theoretical_loss": 3.7178620123709436,
      "tokens_seen": 824932352
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037880641925777333,
      "loss": 2.9955,
      "theoretical_loss": 3.717832896813187,
      "tokens_seen": 824997888
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003787963891675025,
      "loss": 2.8538,
      "theoretical_loss": 3.71780378421576,
      "tokens_seen": 825063424
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003787863590772317,
      "loss": 2.6721,
      "theoretical_loss": 3.7177746745781266,
      "tokens_seen": 825128960
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037877632898696093,
      "loss": 3.0316,
      "theoretical_loss": 3.717745567899751,
      "tokens_seen": 825194496
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037876629889669006,
      "loss": 2.7912,
      "theoretical_loss": 3.7177164641800973,
      "tokens_seen": 825260032
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003787562688064193,
      "loss": 2.8639,
      "theoretical_loss": 3.71768736341863,
      "tokens_seen": 825325568
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003787462387161484,
      "loss": 2.6579,
      "theoretical_loss": 3.717658265614814,
      "tokens_seen": 825391104
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037873620862587766,
      "loss": 2.9474,
      "theoretical_loss": 3.717629170768112,
      "tokens_seen": 825456640
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037872617853560684,
      "loss": 2.6777,
      "theoretical_loss": 3.7176000788779913,
      "tokens_seen": 825522176
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.000378716148445336,
      "loss": 2.766,
      "theoretical_loss": 3.7175709899439147,
      "tokens_seen": 825587712
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003787061183550652,
      "loss": 2.6364,
      "theoretical_loss": 3.7175419039653486,
      "tokens_seen": 825653248
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003786960882647944,
      "loss": 3.0823,
      "theoretical_loss": 3.7175128209417574,
      "tokens_seen": 825718784
    },
    {
      "epoch": 2.07,
      "objective/train/docs_used": 932123,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9297609329223633,
      "objective/train/theoretical_loss": 3.71749828053791,
      "objective/train/tokens_used": 846211552,
      "theoretical_loss": 3.71749828053791,
      "tokens_seen": 825751552
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037868605817452356,
      "loss": 3.0315,
      "theoretical_loss": 3.7174837408726065,
      "tokens_seen": 825784320
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003786760280842528,
      "loss": 2.9329,
      "theoretical_loss": 3.717454663757361,
      "tokens_seen": 825849856
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003786659979939819,
      "loss": 2.8325,
      "theoretical_loss": 3.7174255895954875,
      "tokens_seen": 825915392
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037865596790371116,
      "loss": 2.9202,
      "theoretical_loss": 3.717396518386451,
      "tokens_seen": 825980928
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003786459378134403,
      "loss": 2.709,
      "theoretical_loss": 3.717367450129718,
      "tokens_seen": 826046464
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003786359077231695,
      "loss": 2.7433,
      "theoretical_loss": 3.7173383848247528,
      "tokens_seen": 826112000
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003786258776328987,
      "loss": 2.7836,
      "theoretical_loss": 3.7173093224710234,
      "tokens_seen": 826177536
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003786158475426279,
      "loss": 3.0468,
      "theoretical_loss": 3.717280263067995,
      "tokens_seen": 826243072
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037860581745235706,
      "loss": 3.0001,
      "theoretical_loss": 3.7172512066151344,
      "tokens_seen": 826308608
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003785957873620863,
      "loss": 2.9357,
      "theoretical_loss": 3.7172221531119085,
      "tokens_seen": 826374144
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003785857572718154,
      "loss": 2.5982,
      "theoretical_loss": 3.7171931025577836,
      "tokens_seen": 826439680
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037857572718154466,
      "loss": 2.8929,
      "theoretical_loss": 3.7171640549522262,
      "tokens_seen": 826505216
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003785656970912738,
      "loss": 2.8073,
      "theoretical_loss": 3.717135010294703,
      "tokens_seen": 826570752
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.000378555667001003,
      "loss": 2.7072,
      "theoretical_loss": 3.717105968584683,
      "tokens_seen": 826636288
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003785456369107322,
      "loss": 2.9043,
      "theoretical_loss": 3.717076929821631,
      "tokens_seen": 826701824
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003785356068204614,
      "loss": 3.0974,
      "theoretical_loss": 3.717047894005016,
      "tokens_seen": 826767360
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037852557673019057,
      "loss": 2.5399,
      "theoretical_loss": 3.7170188611343047,
      "tokens_seen": 826832896
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037851554663991975,
      "loss": 3.0706,
      "theoretical_loss": 3.7169898312089655,
      "tokens_seen": 826898432
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037850551654964893,
      "loss": 3.1045,
      "theoretical_loss": 3.7169608042284654,
      "tokens_seen": 826963968
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037849548645937816,
      "loss": 2.8803,
      "theoretical_loss": 3.716931780192273,
      "tokens_seen": 827029504
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003784854563691073,
      "loss": 2.7085,
      "theoretical_loss": 3.7169027590998556,
      "tokens_seen": 827095040
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003784754262788365,
      "loss": 2.8001,
      "theoretical_loss": 3.7168737409506822,
      "tokens_seen": 827160576
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037846539618856565,
      "loss": 3.1734,
      "theoretical_loss": 3.7168447257442208,
      "tokens_seen": 827226112
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003784553660982949,
      "loss": 2.9276,
      "theoretical_loss": 3.7168157134799396,
      "tokens_seen": 827291648
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037844533600802407,
      "loss": 2.7264,
      "theoretical_loss": 3.716786704157308,
      "tokens_seen": 827357184
    },
    {
      "epoch": 2.07,
      "objective/train/docs_used": 933422,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9304966926574707,
      "objective/train/theoretical_loss": 3.7167722005989443,
      "objective/train/tokens_used": 847849952,
      "theoretical_loss": 3.7167722005989443,
      "tokens_seen": 827389952
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037843530591775325,
      "loss": 2.806,
      "theoretical_loss": 3.716757697775794,
      "tokens_seen": 827422720
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037842527582748243,
      "loss": 2.8655,
      "theoretical_loss": 3.7167286943348667,
      "tokens_seen": 827488256
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037841524573721167,
      "loss": 2.9854,
      "theoretical_loss": 3.7166996938339953,
      "tokens_seen": 827553792
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003784052156469408,
      "loss": 2.8557,
      "theoretical_loss": 3.716670696272649,
      "tokens_seen": 827619328
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037839518555667003,
      "loss": 2.8762,
      "theoretical_loss": 3.7166417016502966,
      "tokens_seen": 827684864
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037838515546639916,
      "loss": 2.9352,
      "theoretical_loss": 3.7166127099664084,
      "tokens_seen": 827750400
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003783751253761284,
      "loss": 2.6815,
      "theoretical_loss": 3.716583721220453,
      "tokens_seen": 827815936
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037836509528585763,
      "loss": 2.6254,
      "theoretical_loss": 3.7165547354119015,
      "tokens_seen": 827881472
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037835506519558675,
      "loss": 2.8917,
      "theoretical_loss": 3.716525752540222,
      "tokens_seen": 827947008
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.000378345035105316,
      "loss": 3.027,
      "theoretical_loss": 3.716496772604886,
      "tokens_seen": 828012544
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003783350050150451,
      "loss": 2.8621,
      "theoretical_loss": 3.7164677956053627,
      "tokens_seen": 828078080
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037832497492477435,
      "loss": 3.0463,
      "theoretical_loss": 3.7164388215411233,
      "tokens_seen": 828143616
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037831494483450353,
      "loss": 2.7848,
      "theoretical_loss": 3.7164098504116367,
      "tokens_seen": 828209152
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003783049147442327,
      "loss": 2.8613,
      "theoretical_loss": 3.7163808822163755,
      "tokens_seen": 828274688
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003782948846539619,
      "loss": 2.9271,
      "theoretical_loss": 3.7163519169548085,
      "tokens_seen": 828340224
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037828485456369113,
      "loss": 2.6843,
      "theoretical_loss": 3.7163229546264076,
      "tokens_seen": 828405760
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037827482447342026,
      "loss": 2.8605,
      "theoretical_loss": 3.7162939952306435,
      "tokens_seen": 828471296
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003782647943831495,
      "loss": 2.9175,
      "theoretical_loss": 3.7162650387669873,
      "tokens_seen": 828536832
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003782547642928786,
      "loss": 2.7707,
      "theoretical_loss": 3.71623608523491,
      "tokens_seen": 828602368
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037824473420260786,
      "loss": 2.676,
      "theoretical_loss": 3.7162071346338834,
      "tokens_seen": 828667904
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037823470411233704,
      "loss": 2.845,
      "theoretical_loss": 3.7161781869633788,
      "tokens_seen": 828733440
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003782246740220662,
      "loss": 2.9766,
      "theoretical_loss": 3.7161492422228677,
      "tokens_seen": 828798976
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003782146439317954,
      "loss": 2.9741,
      "theoretical_loss": 3.7161203004118217,
      "tokens_seen": 828864512
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003782046138415246,
      "loss": 2.9232,
      "theoretical_loss": 3.716091361529714,
      "tokens_seen": 828930048
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037819458375125376,
      "loss": 2.7427,
      "theoretical_loss": 3.716062425576015,
      "tokens_seen": 828995584
    },
    {
      "epoch": 2.07,
      "objective/train/docs_used": 934816,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.1100308895111084,
      "objective/train/theoretical_loss": 3.716047958697154,
      "objective/train/tokens_used": 849488352,
      "theoretical_loss": 3.716047958697154,
      "tokens_seen": 829028352
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.000378184553660983,
      "loss": 3.0321,
      "theoretical_loss": 3.7160334925501974,
      "tokens_seen": 829061120
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003781745235707121,
      "loss": 2.8202,
      "theoretical_loss": 3.7160045624517344,
      "tokens_seen": 829126656
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037816449348044136,
      "loss": 2.9412,
      "theoretical_loss": 3.7159756352800977,
      "tokens_seen": 829192192
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003781544633901705,
      "loss": 2.7031,
      "theoretical_loss": 3.7159467110347597,
      "tokens_seen": 829257728
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003781444332998997,
      "loss": 3.0871,
      "theoretical_loss": 3.715917789715194,
      "tokens_seen": 829323264
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003781344032096289,
      "loss": 2.8528,
      "theoretical_loss": 3.715888871320872,
      "tokens_seen": 829388800
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003781243731193581,
      "loss": 2.8641,
      "theoretical_loss": 3.7158599558512684,
      "tokens_seen": 829454336
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037811434302908726,
      "loss": 3.02,
      "theoretical_loss": 3.7158310433058555,
      "tokens_seen": 829519872
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003781043129388165,
      "loss": 2.925,
      "theoretical_loss": 3.7158021336841065,
      "tokens_seen": 829585408
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003780942828485456,
      "loss": 2.9142,
      "theoretical_loss": 3.715773226985495,
      "tokens_seen": 829650944
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037808425275827486,
      "loss": 2.7949,
      "theoretical_loss": 3.7157443232094947,
      "tokens_seen": 829716480
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.000378074222668004,
      "loss": 2.8045,
      "theoretical_loss": 3.7157154223555793,
      "tokens_seen": 829782016
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003780641925777332,
      "loss": 3.2325,
      "theoretical_loss": 3.7156865244232224,
      "tokens_seen": 829847552
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003780541624874624,
      "loss": 2.7547,
      "theoretical_loss": 3.7156576294118984,
      "tokens_seen": 829913088
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003780441323971916,
      "loss": 2.9053,
      "theoretical_loss": 3.715628737321081,
      "tokens_seen": 829978624
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037803410230692077,
      "loss": 2.928,
      "theoretical_loss": 3.7155998481502444,
      "tokens_seen": 830044160
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037802407221664995,
      "loss": 2.8699,
      "theoretical_loss": 3.715570961898863,
      "tokens_seen": 830109696
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037801404212637913,
      "loss": 2.8143,
      "theoretical_loss": 3.7155420785664113,
      "tokens_seen": 830175232
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037800401203610836,
      "loss": 2.7866,
      "theoretical_loss": 3.7155131981523644,
      "tokens_seen": 830240768
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003779939819458375,
      "loss": 2.6816,
      "theoretical_loss": 3.7154843206561967,
      "tokens_seen": 830306304
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037798395185556673,
      "loss": 2.9595,
      "theoretical_loss": 3.7154554460773834,
      "tokens_seen": 830371840
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037797392176529585,
      "loss": 2.9089,
      "theoretical_loss": 3.7154265744153996,
      "tokens_seen": 830437376
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003779638916750251,
      "loss": 2.8497,
      "theoretical_loss": 3.71539770566972,
      "tokens_seen": 830502912
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037795386158475427,
      "loss": 2.9463,
      "theoretical_loss": 3.7153688398398197,
      "tokens_seen": 830568448
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037794383149448345,
      "loss": 2.688,
      "theoretical_loss": 3.715339976925175,
      "tokens_seen": 830633984
    },
    {
      "epoch": 2.07,
      "objective/train/docs_used": 935526,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.734133005142212,
      "objective/train/theoretical_loss": 3.71532554656091,
      "objective/train/tokens_used": 851126752,
      "theoretical_loss": 3.71532554656091,
      "tokens_seen": 830666752
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037793380140421263,
      "loss": 2.9231,
      "theoretical_loss": 3.7153111169252613,
      "tokens_seen": 830699520
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037792377131394187,
      "loss": 2.7234,
      "theoretical_loss": 3.7152822598395545,
      "tokens_seen": 830765056
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.000377913741223671,
      "loss": 2.9191,
      "theoretical_loss": 3.7152534056675295,
      "tokens_seen": 830830592
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037790371113340023,
      "loss": 2.7125,
      "theoretical_loss": 3.715224554408664,
      "tokens_seen": 830896128
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037789368104312936,
      "loss": 2.7919,
      "theoretical_loss": 3.715195706062432,
      "tokens_seen": 830961664
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003778836509528586,
      "loss": 2.8418,
      "theoretical_loss": 3.7151668606283117,
      "tokens_seen": 831027200
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003778736208625878,
      "loss": 2.7902,
      "theoretical_loss": 3.7151380181057783,
      "tokens_seen": 831092736
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037786359077231695,
      "loss": 2.7545,
      "theoretical_loss": 3.715109178494309,
      "tokens_seen": 831158272
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037785356068204614,
      "loss": 2.7125,
      "theoretical_loss": 3.7150803417933798,
      "tokens_seen": 831223808
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003778435305917753,
      "loss": 2.9552,
      "theoretical_loss": 3.7150515080024684,
      "tokens_seen": 831289344
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003778335005015045,
      "loss": 2.764,
      "theoretical_loss": 3.715022677121052,
      "tokens_seen": 831354880
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037782347041123373,
      "loss": 3.0012,
      "theoretical_loss": 3.7149938491486063,
      "tokens_seen": 831420416
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037781344032096286,
      "loss": 2.675,
      "theoretical_loss": 3.7149650240846093,
      "tokens_seen": 831485952
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003778034102306921,
      "loss": 2.7746,
      "theoretical_loss": 3.7149362019285386,
      "tokens_seen": 831551488
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003777933801404213,
      "loss": 2.5407,
      "theoretical_loss": 3.714907382679871,
      "tokens_seen": 831617024
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037778335005015046,
      "loss": 2.8533,
      "theoretical_loss": 3.7148785663380854,
      "tokens_seen": 831682560
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037777331995987964,
      "loss": 2.7901,
      "theoretical_loss": 3.7148497529026585,
      "tokens_seen": 831748096
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003777632898696088,
      "loss": 2.8641,
      "theoretical_loss": 3.714820942373068,
      "tokens_seen": 831813632
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.000377753259779338,
      "loss": 2.8166,
      "theoretical_loss": 3.7147921347487927,
      "tokens_seen": 831879168
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037774322968906724,
      "loss": 2.9293,
      "theoretical_loss": 3.71476333002931,
      "tokens_seen": 831944704
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037773319959879636,
      "loss": 2.9647,
      "theoretical_loss": 3.714734528214099,
      "tokens_seen": 832010240
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003777231695085256,
      "loss": 2.8858,
      "theoretical_loss": 3.714705729302638,
      "tokens_seen": 832075776
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003777131394182547,
      "loss": 2.9028,
      "theoretical_loss": 3.714676933294405,
      "tokens_seen": 832141312
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037770310932798396,
      "loss": 2.7915,
      "theoretical_loss": 3.71464814018888,
      "tokens_seen": 832206848
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037769307923771314,
      "loss": 2.9318,
      "theoretical_loss": 3.71461934998554,
      "tokens_seen": 832272384
    },
    {
      "epoch": 2.07,
      "objective/train/docs_used": 936975,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9853835105895996,
      "objective/train/theoretical_loss": 3.7146049559720273,
      "objective/train/tokens_used": 852765152,
      "theoretical_loss": 3.7146049559720273,
      "tokens_seen": 832305152
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003776830491474423,
      "loss": 2.9336,
      "theoretical_loss": 3.7145905626838656,
      "tokens_seen": 832337920
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003776730190571715,
      "loss": 2.6135,
      "theoretical_loss": 3.714561778283335,
      "tokens_seen": 832403456
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003776629889669007,
      "loss": 2.6917,
      "theoretical_loss": 3.7145329967834275,
      "tokens_seen": 832468992
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037765295887662987,
      "loss": 2.6006,
      "theoretical_loss": 3.714504218183623,
      "tokens_seen": 832534528
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003776429287863591,
      "loss": 2.6902,
      "theoretical_loss": 3.7144754424834003,
      "tokens_seen": 832600064
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037763289869608823,
      "loss": 2.7044,
      "theoretical_loss": 3.71444666968224,
      "tokens_seen": 832665600
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037762286860581746,
      "loss": 2.6894,
      "theoretical_loss": 3.714417899779621,
      "tokens_seen": 832731136
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003776128385155467,
      "loss": 2.7718,
      "theoretical_loss": 3.7143891327750236,
      "tokens_seen": 832796672
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003776028084252758,
      "loss": 2.8134,
      "theoretical_loss": 3.7143603686679283,
      "tokens_seen": 832862208
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037759277833500506,
      "loss": 2.5345,
      "theoretical_loss": 3.7143316074578143,
      "tokens_seen": 832927744
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003775827482447342,
      "loss": 2.7188,
      "theoretical_loss": 3.7143028491441625,
      "tokens_seen": 832993280
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003775727181544634,
      "loss": 2.8647,
      "theoretical_loss": 3.7142740937264533,
      "tokens_seen": 833058816
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003775626880641926,
      "loss": 2.874,
      "theoretical_loss": 3.7142453412041676,
      "tokens_seen": 833124352
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003775526579739218,
      "loss": 2.9832,
      "theoretical_loss": 3.7142165915767853,
      "tokens_seen": 833189888
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037754262788365097,
      "loss": 2.7881,
      "theoretical_loss": 3.7141878448437886,
      "tokens_seen": 833255424
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037753259779338015,
      "loss": 2.7939,
      "theoretical_loss": 3.7141591010046575,
      "tokens_seen": 833320960
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037752256770310933,
      "loss": 2.9099,
      "theoretical_loss": 3.714130360058873,
      "tokens_seen": 833386496
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037751253761283857,
      "loss": 2.9208,
      "theoretical_loss": 3.7141016220059164,
      "tokens_seen": 833452032
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003775025075225677,
      "loss": 2.8116,
      "theoretical_loss": 3.71407288684527,
      "tokens_seen": 833517568
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037749247743229693,
      "loss": 2.8814,
      "theoretical_loss": 3.7140441545764142,
      "tokens_seen": 833583104
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037748244734202605,
      "loss": 2.5721,
      "theoretical_loss": 3.7140154251988315,
      "tokens_seen": 833648640
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003774724172517553,
      "loss": 2.5953,
      "theoretical_loss": 3.713986698712003,
      "tokens_seen": 833714176
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037746238716148447,
      "loss": 2.6858,
      "theoretical_loss": 3.7139579751154117,
      "tokens_seen": 833779712
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037745235707121365,
      "loss": 2.8436,
      "theoretical_loss": 3.7139292544085385,
      "tokens_seen": 833845248
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037744232698094283,
      "loss": 2.9229,
      "theoretical_loss": 3.713900536590866,
      "tokens_seen": 833910784
    },
    {
      "epoch": 2.07,
      "objective/train/docs_used": 937728,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8705952167510986,
      "objective/train/theoretical_loss": 3.7138861787653186,
      "objective/train/tokens_used": 854403552,
      "theoretical_loss": 3.7138861787653186,
      "tokens_seen": 833943552
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037743229689067207,
      "loss": 2.8177,
      "theoretical_loss": 3.713871821661877,
      "tokens_seen": 833976320
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003774222668004012,
      "loss": 2.7473,
      "theoretical_loss": 3.7138431096210534,
      "tokens_seen": 834041856
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037741223671013043,
      "loss": 2.7667,
      "theoretical_loss": 3.7138144004678777,
      "tokens_seen": 834107392
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037740220661985956,
      "loss": 2.8914,
      "theoretical_loss": 3.7137856942018335,
      "tokens_seen": 834172928
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003773921765295888,
      "loss": 2.9923,
      "theoretical_loss": 3.7137569908224024,
      "tokens_seen": 834238464
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.000377382146439318,
      "loss": 2.697,
      "theoretical_loss": 3.713728290329068,
      "tokens_seen": 834304000
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037737211634904715,
      "loss": 2.9992,
      "theoretical_loss": 3.7136995927213143,
      "tokens_seen": 834369536
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037736208625877634,
      "loss": 2.7865,
      "theoretical_loss": 3.7136708979986235,
      "tokens_seen": 834435072
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003773520561685055,
      "loss": 2.7856,
      "theoretical_loss": 3.713642206160479,
      "tokens_seen": 834500608
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003773420260782347,
      "loss": 2.9712,
      "theoretical_loss": 3.7136135172063645,
      "tokens_seen": 834566144
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037733199598796393,
      "loss": 2.6482,
      "theoretical_loss": 3.7135848311357638,
      "tokens_seen": 834631680
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037732196589769306,
      "loss": 2.7742,
      "theoretical_loss": 3.713556147948161,
      "tokens_seen": 834697216
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003773119358074223,
      "loss": 2.7187,
      "theoretical_loss": 3.7135274676430394,
      "tokens_seen": 834762752
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003773019057171515,
      "loss": 2.7803,
      "theoretical_loss": 3.7134987902198833,
      "tokens_seen": 834828288
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037729187562688066,
      "loss": 2.8031,
      "theoretical_loss": 3.7134701156781773,
      "tokens_seen": 834893824
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037728184553660984,
      "loss": 2.7147,
      "theoretical_loss": 3.713441444017405,
      "tokens_seen": 834959360
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.000377271815446339,
      "loss": 2.5068,
      "theoretical_loss": 3.7134127752370514,
      "tokens_seen": 835024896
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003772617853560682,
      "loss": 2.7243,
      "theoretical_loss": 3.713384109336601,
      "tokens_seen": 835090432
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037725175526579744,
      "loss": 2.7841,
      "theoretical_loss": 3.7133554463155383,
      "tokens_seen": 835155968
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037724172517552656,
      "loss": 2.8078,
      "theoretical_loss": 3.713326786173348,
      "tokens_seen": 835221504
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003772316950852558,
      "loss": 2.5191,
      "theoretical_loss": 3.713298128909516,
      "tokens_seen": 835287040
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003772216649949849,
      "loss": 2.6828,
      "theoretical_loss": 3.7132694745235266,
      "tokens_seen": 835352576
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037721163490471416,
      "loss": 2.8356,
      "theoretical_loss": 3.713240823014866,
      "tokens_seen": 835418112
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037720160481444334,
      "loss": 2.8376,
      "theoretical_loss": 3.7132121743830178,
      "tokens_seen": 835483648
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003771915747241725,
      "loss": 2.9586,
      "theoretical_loss": 3.713183528627469,
      "tokens_seen": 835549184
    },
    {
      "epoch": 2.07,
      "objective/train/docs_used": 939195,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6481375694274902,
      "objective/train/theoretical_loss": 3.713169206828146,
      "objective/train/tokens_used": 856041952,
      "theoretical_loss": 3.713169206828146,
      "tokens_seen": 835581952
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003771815446339017,
      "loss": 2.8213,
      "theoretical_loss": 3.713154885747705,
      "tokens_seen": 835614720
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003771715145436309,
      "loss": 2.7886,
      "theoretical_loss": 3.713126245743211,
      "tokens_seen": 835680256
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037716148445336007,
      "loss": 2.6125,
      "theoretical_loss": 3.7130976086134737,
      "tokens_seen": 835745792
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003771514543630893,
      "loss": 2.9789,
      "theoretical_loss": 3.713068974357979,
      "tokens_seen": 835811328
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037714142427281843,
      "loss": 2.8882,
      "theoretical_loss": 3.7130403429762127,
      "tokens_seen": 835876864
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037713139418254766,
      "loss": 2.7156,
      "theoretical_loss": 3.713011714467661,
      "tokens_seen": 835942400
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037712136409227685,
      "loss": 2.8133,
      "theoretical_loss": 3.7129830888318107,
      "tokens_seen": 836007936
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.000377111334002006,
      "loss": 2.8748,
      "theoretical_loss": 3.7129544660681493,
      "tokens_seen": 836073472
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003771013039117352,
      "loss": 2.7447,
      "theoretical_loss": 3.7129258461761614,
      "tokens_seen": 836139008
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003770912738214644,
      "loss": 2.4627,
      "theoretical_loss": 3.7128972291553355,
      "tokens_seen": 836204544
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037708124373119357,
      "loss": 2.8791,
      "theoretical_loss": 3.712868615005158,
      "tokens_seen": 836270080
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003770712136409228,
      "loss": 2.9345,
      "theoretical_loss": 3.712840003725116,
      "tokens_seen": 836335616
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037706118355065193,
      "loss": 2.6005,
      "theoretical_loss": 3.7128113953146973,
      "tokens_seen": 836401152
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037705115346038117,
      "loss": 2.7668,
      "theoretical_loss": 3.7127827897733887,
      "tokens_seen": 836466688
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003770411233701103,
      "loss": 2.6267,
      "theoretical_loss": 3.712754187100677,
      "tokens_seen": 836532224
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037703109327983953,
      "loss": 2.8772,
      "theoretical_loss": 3.712725587296051,
      "tokens_seen": 836597760
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003770210631895687,
      "loss": 2.9275,
      "theoretical_loss": 3.7126969903589986,
      "tokens_seen": 836663296
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003770110330992979,
      "loss": 2.8701,
      "theoretical_loss": 3.712668396289007,
      "tokens_seen": 836728832
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003770010030090271,
      "loss": 2.8012,
      "theoretical_loss": 3.712639805085564,
      "tokens_seen": 836794368
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037699097291875625,
      "loss": 2.7351,
      "theoretical_loss": 3.7126112167481584,
      "tokens_seen": 836859904
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037698094282848544,
      "loss": 2.7632,
      "theoretical_loss": 3.712582631276278,
      "tokens_seen": 836925440
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037697091273821467,
      "loss": 2.9011,
      "theoretical_loss": 3.712554048669412,
      "tokens_seen": 836990976
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003769608826479438,
      "loss": 2.8888,
      "theoretical_loss": 3.7125254689270486,
      "tokens_seen": 837056512
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037695085255767303,
      "loss": 2.9281,
      "theoretical_loss": 3.7124968920486756,
      "tokens_seen": 837122048
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003769408224674022,
      "loss": 2.8598,
      "theoretical_loss": 3.7124683180337827,
      "tokens_seen": 837187584
    },
    {
      "epoch": 2.07,
      "objective/train/docs_used": 939764,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9354538917541504,
      "objective/train/theoretical_loss": 3.712454032099982,
      "objective/train/tokens_used": 857680352,
      "theoretical_loss": 3.712454032099982,
      "tokens_seen": 837220352
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003769307923771314,
      "loss": 2.8442,
      "theoretical_loss": 3.7124397468818593,
      "tokens_seen": 837253120
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003769207622868606,
      "loss": 2.6892,
      "theoretical_loss": 3.7124111785923937,
      "tokens_seen": 837318656
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037691073219658976,
      "loss": 2.6348,
      "theoretical_loss": 3.712382613164875,
      "tokens_seen": 837384192
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037690070210631894,
      "loss": 2.7621,
      "theoretical_loss": 3.7123540505987926,
      "tokens_seen": 837449728
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003768906720160482,
      "loss": 2.8155,
      "theoretical_loss": 3.7123254908936367,
      "tokens_seen": 837515264
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003768806419257773,
      "loss": 2.8675,
      "theoretical_loss": 3.7122969340488963,
      "tokens_seen": 837580800
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037687061183550654,
      "loss": 2.7471,
      "theoretical_loss": 3.712268380064061,
      "tokens_seen": 837646336
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003768605817452357,
      "loss": 2.804,
      "theoretical_loss": 3.712239828938621,
      "tokens_seen": 837711872
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003768505516549649,
      "loss": 2.8195,
      "theoretical_loss": 3.712211280672067,
      "tokens_seen": 837777408
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037684052156469413,
      "loss": 2.9182,
      "theoretical_loss": 3.7121827352638874,
      "tokens_seen": 837842944
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037683049147442326,
      "loss": 2.6297,
      "theoretical_loss": 3.712154192713574,
      "tokens_seen": 837908480
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003768204613841525,
      "loss": 2.5391,
      "theoretical_loss": 3.7121256530206157,
      "tokens_seen": 837974016
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003768104312938817,
      "loss": 2.7109,
      "theoretical_loss": 3.7120971161845047,
      "tokens_seen": 838039552
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037680040120361086,
      "loss": 2.8297,
      "theoretical_loss": 3.712068582204731,
      "tokens_seen": 838105088
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037679037111334004,
      "loss": 3.0213,
      "theoretical_loss": 3.712040051080785,
      "tokens_seen": 838170624
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003767803410230692,
      "loss": 2.8937,
      "theoretical_loss": 3.7120115228121575,
      "tokens_seen": 838236160
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003767703109327984,
      "loss": 2.82,
      "theoretical_loss": 3.71198299739834,
      "tokens_seen": 838301696
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037676028084252764,
      "loss": 2.6802,
      "theoretical_loss": 3.711954474838824,
      "tokens_seen": 838367232
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037675025075225676,
      "loss": 2.8553,
      "theoretical_loss": 3.7119259551330996,
      "tokens_seen": 838432768
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.000376740220661986,
      "loss": 2.6405,
      "theoretical_loss": 3.7118974382806598,
      "tokens_seen": 838498304
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003767301905717151,
      "loss": 2.9445,
      "theoretical_loss": 3.711868924280995,
      "tokens_seen": 838563840
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037672016048144436,
      "loss": 2.845,
      "theoretical_loss": 3.7118404131335976,
      "tokens_seen": 838629376
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037671013039117354,
      "loss": 2.9059,
      "theoretical_loss": 3.7118119048379588,
      "tokens_seen": 838694912
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003767001003009027,
      "loss": 2.8623,
      "theoretical_loss": 3.711783399393571,
      "tokens_seen": 838760448
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003766900702106319,
      "loss": 3.0017,
      "theoretical_loss": 3.711754896799926,
      "tokens_seen": 838825984
    },
    {
      "epoch": 2.07,
      "objective/train/docs_used": 940402,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.893378257751465,
      "objective/train/theoretical_loss": 3.7117406465719736,
      "objective/train/tokens_used": 859318752,
      "theoretical_loss": 3.7117406465719736,
      "tokens_seen": 838858752
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003766800401203611,
      "loss": 2.7411,
      "theoretical_loss": 3.711726397056516,
      "tokens_seen": 838891520
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037667001003009027,
      "loss": 2.524,
      "theoretical_loss": 3.711697900162834,
      "tokens_seen": 838957056
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003766599799398195,
      "loss": 2.8174,
      "theoretical_loss": 3.7116694061183715,
      "tokens_seen": 839022592
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037664994984954863,
      "loss": 2.7206,
      "theoretical_loss": 3.7116409149226213,
      "tokens_seen": 839088128
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037663991975927786,
      "loss": 2.7177,
      "theoretical_loss": 3.711612426575077,
      "tokens_seen": 839153664
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037662988966900705,
      "loss": 2.7343,
      "theoretical_loss": 3.711583941075231,
      "tokens_seen": 839219200
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003766198595787362,
      "loss": 3.0564,
      "theoretical_loss": 3.711555458422576,
      "tokens_seen": 839284736
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003766098294884654,
      "loss": 2.5675,
      "theoretical_loss": 3.711526978616605,
      "tokens_seen": 839350272
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003765997993981946,
      "loss": 2.6686,
      "theoretical_loss": 3.711498501656812,
      "tokens_seen": 839415808
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037658976930792377,
      "loss": 2.8999,
      "theoretical_loss": 3.7114700275426893,
      "tokens_seen": 839481344
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.000376579739217653,
      "loss": 2.5858,
      "theoretical_loss": 3.7114415562737317,
      "tokens_seen": 839546880
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037656970912738213,
      "loss": 2.8555,
      "theoretical_loss": 3.711413087849432,
      "tokens_seen": 839612416
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037655967903711137,
      "loss": 2.7198,
      "theoretical_loss": 3.711384622269284,
      "tokens_seen": 839677952
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003765496489468405,
      "loss": 2.9233,
      "theoretical_loss": 3.711356159532782,
      "tokens_seen": 839743488
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037653961885656973,
      "loss": 2.7966,
      "theoretical_loss": 3.71132769963942,
      "tokens_seen": 839809024
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003765295887662989,
      "loss": 2.702,
      "theoretical_loss": 3.7112992425886917,
      "tokens_seen": 839874560
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003765195586760281,
      "loss": 2.7411,
      "theoretical_loss": 3.7112707883800917,
      "tokens_seen": 839940096
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003765095285857573,
      "loss": 2.5153,
      "theoretical_loss": 3.7112423370131142,
      "tokens_seen": 840005632
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037649949849548645,
      "loss": 2.5897,
      "theoretical_loss": 3.7112138884872543,
      "tokens_seen": 840071168
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037648946840521564,
      "loss": 2.5425,
      "theoretical_loss": 3.711185442802006,
      "tokens_seen": 840136704
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037647943831494487,
      "loss": 2.8974,
      "theoretical_loss": 3.7111569999568648,
      "tokens_seen": 840202240
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.000376469408224674,
      "loss": 2.8294,
      "theoretical_loss": 3.711128559951325,
      "tokens_seen": 840267776
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037645937813440323,
      "loss": 2.9157,
      "theoretical_loss": 3.711100122784882,
      "tokens_seen": 840333312
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003764493480441324,
      "loss": 3.2182,
      "theoretical_loss": 3.7110716884570305,
      "tokens_seen": 840398848
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003764393179538616,
      "loss": 2.7897,
      "theoretical_loss": 3.711043256967267,
      "tokens_seen": 840464384
    },
    {
      "epoch": 2.07,
      "objective/train/docs_used": 941411,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.780778169631958,
      "objective/train/theoretical_loss": 3.71102904228651,
      "objective/train/tokens_used": 860957152,
      "theoretical_loss": 3.71102904228651,
      "tokens_seen": 840497152
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003764292878635908,
      "loss": 2.8305,
      "theoretical_loss": 3.7110148283150854,
      "tokens_seen": 840529920
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037641925777331996,
      "loss": 2.7123,
      "theoretical_loss": 3.710986402499983,
      "tokens_seen": 840595456
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037640922768304914,
      "loss": 3.029,
      "theoretical_loss": 3.710957979521454,
      "tokens_seen": 840660992
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003763991975927784,
      "loss": 2.761,
      "theoretical_loss": 3.7109295593789953,
      "tokens_seen": 840726528
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003763891675025075,
      "loss": 2.6939,
      "theoretical_loss": 3.7109011420721023,
      "tokens_seen": 840792064
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037637913741223674,
      "loss": 2.6892,
      "theoretical_loss": 3.710872727600271,
      "tokens_seen": 840857600
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.00037636910732196586,
      "loss": 2.8136,
      "theoretical_loss": 3.710844315962998,
      "tokens_seen": 840923136
    },
    {
      "epoch": 2.07,
      "learning_rate": 0.0003763590772316951,
      "loss": 2.6677,
      "theoretical_loss": 3.7108159071597795,
      "tokens_seen": 840988672
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003763490471414243,
      "loss": 2.7426,
      "theoretical_loss": 3.710787501190112,
      "tokens_seen": 841054208
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037633901705115346,
      "loss": 2.7228,
      "theoretical_loss": 3.710759098053492,
      "tokens_seen": 841119744
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037632898696088264,
      "loss": 2.8451,
      "theoretical_loss": 3.7107306977494163,
      "tokens_seen": 841185280
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003763189568706119,
      "loss": 2.7124,
      "theoretical_loss": 3.710702300277382,
      "tokens_seen": 841250816
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.000376308926780341,
      "loss": 2.8676,
      "theoretical_loss": 3.7106739056368863,
      "tokens_seen": 841316352
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037629889669007024,
      "loss": 2.812,
      "theoretical_loss": 3.7106455138274255,
      "tokens_seen": 841381888
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037628886659979937,
      "loss": 2.8864,
      "theoretical_loss": 3.710617124848497,
      "tokens_seen": 841447424
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003762788365095286,
      "loss": 2.5823,
      "theoretical_loss": 3.710588738699599,
      "tokens_seen": 841512960
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003762688064192578,
      "loss": 2.6779,
      "theoretical_loss": 3.7105603553802284,
      "tokens_seen": 841578496
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037625877632898696,
      "loss": 2.6513,
      "theoretical_loss": 3.710531974889883,
      "tokens_seen": 841644032
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037624874623871615,
      "loss": 2.9036,
      "theoretical_loss": 3.710503597228061,
      "tokens_seen": 841709568
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003762387161484453,
      "loss": 2.9422,
      "theoretical_loss": 3.710475222394259,
      "tokens_seen": 841775104
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003762286860581745,
      "loss": 2.722,
      "theoretical_loss": 3.710446850387976,
      "tokens_seen": 841840640
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037621865596790374,
      "loss": 2.8682,
      "theoretical_loss": 3.7104184812087104,
      "tokens_seen": 841906176
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037620862587763287,
      "loss": 2.5776,
      "theoretical_loss": 3.71039011485596,
      "tokens_seen": 841971712
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003761985957873621,
      "loss": 2.5632,
      "theoretical_loss": 3.7103617513292235,
      "tokens_seen": 842037248
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037618856569709123,
      "loss": 2.5813,
      "theoretical_loss": 3.7103333906279987,
      "tokens_seen": 842102784
    },
    {
      "epoch": 2.08,
      "objective/train/docs_used": 942076,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.738717794418335,
      "objective/train/theoretical_loss": 3.710319211336797,
      "objective/train/tokens_used": 862595552,
      "theoretical_loss": 3.710319211336797,
      "tokens_seen": 842135552
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037617853560682047,
      "loss": 2.8054,
      "theoretical_loss": 3.7103050327517852,
      "tokens_seen": 842168320
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037616850551654965,
      "loss": 2.7416,
      "theoretical_loss": 3.7102766777000813,
      "tokens_seen": 842233856
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037615847542627883,
      "loss": 3.0039,
      "theoretical_loss": 3.7102483254723864,
      "tokens_seen": 842299392
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.000376148445336008,
      "loss": 2.9361,
      "theoretical_loss": 3.710219976068199,
      "tokens_seen": 842364928
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037613841524573725,
      "loss": 2.7569,
      "theoretical_loss": 3.710191629487018,
      "tokens_seen": 842430464
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037612838515546637,
      "loss": 2.5037,
      "theoretical_loss": 3.7101632857283438,
      "tokens_seen": 842496000
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003761183550651956,
      "loss": 2.8678,
      "theoretical_loss": 3.710134944791675,
      "tokens_seen": 842561536
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003761083249749248,
      "loss": 2.7399,
      "theoretical_loss": 3.7101066066765114,
      "tokens_seen": 842627072
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037609829488465397,
      "loss": 3.0493,
      "theoretical_loss": 3.710078271382353,
      "tokens_seen": 842692608
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003760882647943832,
      "loss": 2.9969,
      "theoretical_loss": 3.7100499389086985,
      "tokens_seen": 842758144
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037607823470411233,
      "loss": 2.8912,
      "theoretical_loss": 3.7100216092550493,
      "tokens_seen": 842823680
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037606820461384157,
      "loss": 2.9114,
      "theoretical_loss": 3.7099932824209043,
      "tokens_seen": 842889216
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003760581745235707,
      "loss": 2.6763,
      "theoretical_loss": 3.709964958405765,
      "tokens_seen": 842954752
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037604814443329993,
      "loss": 2.7973,
      "theoretical_loss": 3.7099366372091303,
      "tokens_seen": 843020288
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003760381143430291,
      "loss": 3.0025,
      "theoretical_loss": 3.709908318830501,
      "tokens_seen": 843085824
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003760280842527583,
      "loss": 3.03,
      "theoretical_loss": 3.7098800032693786,
      "tokens_seen": 843151360
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003760180541624875,
      "loss": 2.791,
      "theoretical_loss": 3.709851690525263,
      "tokens_seen": 843216896
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037600802407221665,
      "loss": 2.9545,
      "theoretical_loss": 3.7098233805976557,
      "tokens_seen": 843282432
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037599799398194584,
      "loss": 2.5329,
      "theoretical_loss": 3.7097950734860565,
      "tokens_seen": 843347968
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037598796389167507,
      "loss": 2.914,
      "theoretical_loss": 3.7097667691899674,
      "tokens_seen": 843413504
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003759779338014042,
      "loss": 2.8874,
      "theoretical_loss": 3.709738467708889,
      "tokens_seen": 843479040
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037596790371113343,
      "loss": 2.6411,
      "theoretical_loss": 3.7097101690423235,
      "tokens_seen": 843544576
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003759578736208626,
      "loss": 2.6696,
      "theoretical_loss": 3.7096818731897723,
      "tokens_seen": 843610112
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003759478435305918,
      "loss": 2.7156,
      "theoretical_loss": 3.709653580150736,
      "tokens_seen": 843675648
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.000375937813440321,
      "loss": 2.7071,
      "theoretical_loss": 3.7096252899247175,
      "tokens_seen": 843741184
    },
    {
      "epoch": 2.08,
      "objective/train/docs_used": 943510,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8786866664886475,
      "objective/train/theoretical_loss": 3.709611145866434,
      "objective/train/tokens_used": 864233952,
      "theoretical_loss": 3.709611145866434,
      "tokens_seen": 843773952
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037592778335005016,
      "loss": 2.8376,
      "theoretical_loss": 3.7095970025112175,
      "tokens_seen": 843806720
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037591775325977934,
      "loss": 2.8534,
      "theoretical_loss": 3.7095687179097387,
      "tokens_seen": 843872256
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003759077231695086,
      "loss": 2.9706,
      "theoretical_loss": 3.7095404361197835,
      "tokens_seen": 843937792
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003758976930792377,
      "loss": 2.6948,
      "theoretical_loss": 3.7095121571408534,
      "tokens_seen": 844003328
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037588766298896694,
      "loss": 2.7853,
      "theoretical_loss": 3.709483880972451,
      "tokens_seen": 844068864
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037587763289869606,
      "loss": 2.8065,
      "theoretical_loss": 3.709455607614079,
      "tokens_seen": 844134400
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003758676028084253,
      "loss": 2.7958,
      "theoretical_loss": 3.7094273370652404,
      "tokens_seen": 844199936
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003758575727181545,
      "loss": 2.8232,
      "theoretical_loss": 3.709399069325437,
      "tokens_seen": 844265472
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037584754262788366,
      "loss": 3.139,
      "theoretical_loss": 3.7093708043941716,
      "tokens_seen": 844331008
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037583751253761284,
      "loss": 2.8477,
      "theoretical_loss": 3.7093425422709485,
      "tokens_seen": 844396544
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003758274824473421,
      "loss": 2.8708,
      "theoretical_loss": 3.70931428295527,
      "tokens_seen": 844462080
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003758174523570712,
      "loss": 2.5674,
      "theoretical_loss": 3.7092860264466387,
      "tokens_seen": 844527616
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037580742226680044,
      "loss": 2.7413,
      "theoretical_loss": 3.709257772744559,
      "tokens_seen": 844593152
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037579739217652957,
      "loss": 2.7255,
      "theoretical_loss": 3.709229521848534,
      "tokens_seen": 844658688
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003757873620862588,
      "loss": 2.78,
      "theoretical_loss": 3.7092012737580675,
      "tokens_seen": 844724224
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.000375777331995988,
      "loss": 2.5741,
      "theoretical_loss": 3.709173028472663,
      "tokens_seen": 844789760
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037576730190571716,
      "loss": 2.8994,
      "theoretical_loss": 3.7091447859918247,
      "tokens_seen": 844855296
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037575727181544635,
      "loss": 2.7549,
      "theoretical_loss": 3.709116546315056,
      "tokens_seen": 844920832
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003757472417251755,
      "loss": 2.8349,
      "theoretical_loss": 3.709088309441862,
      "tokens_seen": 844986368
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003757372116349047,
      "loss": 2.9799,
      "theoretical_loss": 3.7090600753717458,
      "tokens_seen": 845051904
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037572718154463394,
      "loss": 2.9412,
      "theoretical_loss": 3.7090318441042127,
      "tokens_seen": 845117440
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037571715145436307,
      "loss": 2.6342,
      "theoretical_loss": 3.7090036156387667,
      "tokens_seen": 845182976
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003757071213640923,
      "loss": 3.0338,
      "theoretical_loss": 3.7089753899749134,
      "tokens_seen": 845248512
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037569709127382143,
      "loss": 2.7319,
      "theoretical_loss": 3.708947167112156,
      "tokens_seen": 845314048
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037568706118355067,
      "loss": 2.6745,
      "theoretical_loss": 3.70891894705,
      "tokens_seen": 845379584
    },
    {
      "epoch": 2.08,
      "objective/train/docs_used": 944052,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.632538318634033,
      "objective/train/theoretical_loss": 3.7089048380689933,
      "objective/train/tokens_used": 865872352,
      "theoretical_loss": 3.7089048380689933,
      "tokens_seen": 845412352
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037567703109327985,
      "loss": 2.8133,
      "theoretical_loss": 3.708890729787951,
      "tokens_seen": 845445120
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037566700100300903,
      "loss": 2.7545,
      "theoretical_loss": 3.7088625153255137,
      "tokens_seen": 845510656
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003756569709127382,
      "loss": 2.8038,
      "theoretical_loss": 3.7088343036621936,
      "tokens_seen": 845576192
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037564694082246745,
      "loss": 2.8285,
      "theoretical_loss": 3.7088060947974957,
      "tokens_seen": 845641728
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003756369107321966,
      "loss": 2.6281,
      "theoretical_loss": 3.7087778887309257,
      "tokens_seen": 845707264
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003756268806419258,
      "loss": 2.9724,
      "theoretical_loss": 3.7087496854619895,
      "tokens_seen": 845772800
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037561685055165494,
      "loss": 2.9071,
      "theoretical_loss": 3.708721484990192,
      "tokens_seen": 845838336
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037560682046138417,
      "loss": 2.8209,
      "theoretical_loss": 3.7086932873150404,
      "tokens_seen": 845903872
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037559679037111335,
      "loss": 2.725,
      "theoretical_loss": 3.7086650924360396,
      "tokens_seen": 845969408
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037558676028084253,
      "loss": 2.7387,
      "theoretical_loss": 3.7086369003526967,
      "tokens_seen": 846034944
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003755767301905717,
      "loss": 2.6394,
      "theoretical_loss": 3.7086087110645165,
      "tokens_seen": 846100480
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003755667001003009,
      "loss": 2.9535,
      "theoretical_loss": 3.708580524571007,
      "tokens_seen": 846166016
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003755566700100301,
      "loss": 2.7447,
      "theoretical_loss": 3.708552340871674,
      "tokens_seen": 846231552
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003755466399197593,
      "loss": 2.691,
      "theoretical_loss": 3.708524159966024,
      "tokens_seen": 846297088
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037553660982948844,
      "loss": 2.8282,
      "theoretical_loss": 3.7084959818535648,
      "tokens_seen": 846362624
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003755265797392177,
      "loss": 2.7942,
      "theoretical_loss": 3.7084678065338013,
      "tokens_seen": 846428160
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003755165496489468,
      "loss": 2.687,
      "theoretical_loss": 3.7084396340062424,
      "tokens_seen": 846493696
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037550651955867604,
      "loss": 2.9069,
      "theoretical_loss": 3.7084114642703945,
      "tokens_seen": 846559232
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003754964894684052,
      "loss": 2.7304,
      "theoretical_loss": 3.7083832973257653,
      "tokens_seen": 846624768
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003754864593781344,
      "loss": 2.9416,
      "theoretical_loss": 3.708355133171861,
      "tokens_seen": 846690304
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003754764292878636,
      "loss": 2.7376,
      "theoretical_loss": 3.708326971808191,
      "tokens_seen": 846755840
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003754663991975928,
      "loss": 2.8387,
      "theoretical_loss": 3.7082988132342605,
      "tokens_seen": 846821376
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037545636910732194,
      "loss": 2.7832,
      "theoretical_loss": 3.70827065744958,
      "tokens_seen": 846886912
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003754463390170512,
      "loss": 2.7474,
      "theoretical_loss": 3.708242504453655,
      "tokens_seen": 846952448
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003754363089267803,
      "loss": 2.9122,
      "theoretical_loss": 3.708214354245995,
      "tokens_seen": 847017984
    },
    {
      "epoch": 2.08,
      "objective/train/docs_used": 945481,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.310624122619629,
      "objective/train/theoretical_loss": 3.7082002801876106,
      "objective/train/tokens_used": 867510752,
      "theoretical_loss": 3.7082002801876106,
      "tokens_seen": 847050752
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037542627883650954,
      "loss": 2.6795,
      "theoretical_loss": 3.708186206826108,
      "tokens_seen": 847083520
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003754162487462387,
      "loss": 2.6511,
      "theoretical_loss": 3.708158062193502,
      "tokens_seen": 847149056
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003754062186559679,
      "loss": 2.9887,
      "theoretical_loss": 3.7081299203476847,
      "tokens_seen": 847214592
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003753961885656971,
      "loss": 2.9141,
      "theoretical_loss": 3.708101781288166,
      "tokens_seen": 847280128
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037538615847542626,
      "loss": 3.0222,
      "theoretical_loss": 3.708073645014454,
      "tokens_seen": 847345664
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037537612838515544,
      "loss": 2.6518,
      "theoretical_loss": 3.708045511526057,
      "tokens_seen": 847411200
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003753660982948847,
      "loss": 2.5696,
      "theoretical_loss": 3.7080173808224846,
      "tokens_seen": 847476736
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037535606820461386,
      "loss": 2.8286,
      "theoretical_loss": 3.707989252903245,
      "tokens_seen": 847542272
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037534603811434304,
      "loss": 2.7683,
      "theoretical_loss": 3.707961127767848,
      "tokens_seen": 847607808
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003753360080240723,
      "loss": 2.7759,
      "theoretical_loss": 3.7079330054158026,
      "tokens_seen": 847673344
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003753259779338014,
      "loss": 3.1806,
      "theoretical_loss": 3.7079048858466184,
      "tokens_seen": 847738880
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037531594784353064,
      "loss": 2.7589,
      "theoretical_loss": 3.707876769059805,
      "tokens_seen": 847804416
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037530591775325977,
      "loss": 2.6025,
      "theoretical_loss": 3.7078486550548715,
      "tokens_seen": 847869952
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.000375295887662989,
      "loss": 2.7559,
      "theoretical_loss": 3.707820543831328,
      "tokens_seen": 847935488
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003752858575727182,
      "loss": 2.5894,
      "theoretical_loss": 3.7077924353886846,
      "tokens_seen": 848001024
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037527582748244736,
      "loss": 2.745,
      "theoretical_loss": 3.707764329726451,
      "tokens_seen": 848066560
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037526579739217655,
      "loss": 2.6036,
      "theoretical_loss": 3.7077362268441374,
      "tokens_seen": 848132096
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003752557673019057,
      "loss": 2.531,
      "theoretical_loss": 3.707708126741254,
      "tokens_seen": 848197632
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003752457372116349,
      "loss": 2.9765,
      "theoretical_loss": 3.7076800294173116,
      "tokens_seen": 848263168
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037523570712136414,
      "loss": 2.7127,
      "theoretical_loss": 3.70765193487182,
      "tokens_seen": 848328704
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037522567703109327,
      "loss": 2.587,
      "theoretical_loss": 3.707623843104291,
      "tokens_seen": 848394240
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003752156469408225,
      "loss": 2.6971,
      "theoretical_loss": 3.707595754114234,
      "tokens_seen": 848459776
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037520561685055163,
      "loss": 2.6927,
      "theoretical_loss": 3.707567667901161,
      "tokens_seen": 848525312
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037519558676028087,
      "loss": 2.6176,
      "theoretical_loss": 3.707539584464582,
      "tokens_seen": 848590848
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037518555667001005,
      "loss": 2.6565,
      "theoretical_loss": 3.7075115038040085,
      "tokens_seen": 848656384
    },
    {
      "epoch": 2.08,
      "objective/train/docs_used": 946246,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.916497230529785,
      "objective/train/theoretical_loss": 3.7074974645145717,
      "objective/train/tokens_used": 869149152,
      "theoretical_loss": 3.7074974645145717,
      "tokens_seen": 848689152
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037517552657973923,
      "loss": 2.8534,
      "theoretical_loss": 3.7074834259189524,
      "tokens_seen": 848721920
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003751654964894684,
      "loss": 2.5884,
      "theoretical_loss": 3.7074553508089245,
      "tokens_seen": 848787456
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037515546639919765,
      "loss": 2.5969,
      "theoretical_loss": 3.7074272784734363,
      "tokens_seen": 848852992
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003751454363089268,
      "loss": 2.7696,
      "theoretical_loss": 3.7073992089119994,
      "tokens_seen": 848918528
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.000375135406218656,
      "loss": 2.6844,
      "theoretical_loss": 3.7073711421241256,
      "tokens_seen": 848984064
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037512537612838514,
      "loss": 2.6811,
      "theoretical_loss": 3.707343078109327,
      "tokens_seen": 849049600
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037511534603811437,
      "loss": 2.9713,
      "theoretical_loss": 3.707315016867115,
      "tokens_seen": 849115136
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037510531594784355,
      "loss": 2.8274,
      "theoretical_loss": 3.7072869583970025,
      "tokens_seen": 849180672
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037509528585757273,
      "loss": 2.7507,
      "theoretical_loss": 3.7072589026985012,
      "tokens_seen": 849246208
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003750852557673019,
      "loss": 3.0336,
      "theoretical_loss": 3.707230849771124,
      "tokens_seen": 849311744
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003750752256770311,
      "loss": 2.9991,
      "theoretical_loss": 3.707202799614383,
      "tokens_seen": 849377280
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003750651955867603,
      "loss": 2.6436,
      "theoretical_loss": 3.7071747522277905,
      "tokens_seen": 849442816
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003750551654964895,
      "loss": 2.7344,
      "theoretical_loss": 3.70714670761086,
      "tokens_seen": 849508352
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037504513540621864,
      "loss": 2.7969,
      "theoretical_loss": 3.7071186657631037,
      "tokens_seen": 849573888
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003750351053159479,
      "loss": 2.8591,
      "theoretical_loss": 3.7070906266840353,
      "tokens_seen": 849639424
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.000375025075225677,
      "loss": 3.018,
      "theoretical_loss": 3.707062590373167,
      "tokens_seen": 849704960
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037501504513540624,
      "loss": 2.6358,
      "theoretical_loss": 3.7070345568300125,
      "tokens_seen": 849770496
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003750050150451354,
      "loss": 2.9031,
      "theoretical_loss": 3.707006526054085,
      "tokens_seen": 849836032
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003749949849548646,
      "loss": 2.6426,
      "theoretical_loss": 3.706978498044898,
      "tokens_seen": 849901568
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003749849548645938,
      "loss": 2.6169,
      "theoretical_loss": 3.7069504728019655,
      "tokens_seen": 849967104
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.000374974924774323,
      "loss": 2.8302,
      "theoretical_loss": 3.706922450324801,
      "tokens_seen": 850032640
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037496489468405214,
      "loss": 2.7208,
      "theoretical_loss": 3.706894430612919,
      "tokens_seen": 850098176
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003749548645937814,
      "loss": 2.7541,
      "theoretical_loss": 3.7068664136658316,
      "tokens_seen": 850163712
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003749448345035105,
      "loss": 2.7727,
      "theoretical_loss": 3.7068383994830545,
      "tokens_seen": 850229248
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037493480441323974,
      "loss": 2.6672,
      "theoretical_loss": 3.7068103880641017,
      "tokens_seen": 850294784
    },
    {
      "epoch": 2.08,
      "objective/train/docs_used": 947576,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4943060874938965,
      "objective/train/theoretical_loss": 3.7067963833909072,
      "objective/train/tokens_used": 870787552,
      "theoretical_loss": 3.7067963833909072,
      "tokens_seen": 850327552
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003749247743229689,
      "loss": 2.6507,
      "theoretical_loss": 3.706782379408487,
      "tokens_seen": 850360320
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003749147442326981,
      "loss": 2.79,
      "theoretical_loss": 3.706754373515725,
      "tokens_seen": 850425856
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003749047141424273,
      "loss": 2.9062,
      "theoretical_loss": 3.7067263703853306,
      "tokens_seen": 850491392
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037489468405215646,
      "loss": 2.8495,
      "theoretical_loss": 3.7066983700168183,
      "tokens_seen": 850556928
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037488465396188564,
      "loss": 2.6916,
      "theoretical_loss": 3.706670372409703,
      "tokens_seen": 850622464
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003748746238716149,
      "loss": 2.7941,
      "theoretical_loss": 3.7066423775635,
      "tokens_seen": 850688000
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.000374864593781344,
      "loss": 2.8913,
      "theoretical_loss": 3.7066143854777236,
      "tokens_seen": 850753536
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037485456369107324,
      "loss": 3.0094,
      "theoretical_loss": 3.7065863961518897,
      "tokens_seen": 850819072
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037484453360080237,
      "loss": 2.9763,
      "theoretical_loss": 3.7065584095855133,
      "tokens_seen": 850884608
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003748345035105316,
      "loss": 2.9927,
      "theoretical_loss": 3.7065304257781095,
      "tokens_seen": 850950144
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003748244734202608,
      "loss": 2.7568,
      "theoretical_loss": 3.7065024447291943,
      "tokens_seen": 851015680
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037481444332998997,
      "loss": 2.721,
      "theoretical_loss": 3.7064744664382836,
      "tokens_seen": 851081216
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037480441323971915,
      "loss": 2.5999,
      "theoretical_loss": 3.706446490904893,
      "tokens_seen": 851146752
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003747943831494484,
      "loss": 2.7467,
      "theoretical_loss": 3.7064185181285376,
      "tokens_seen": 851212288
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003747843530591775,
      "loss": 2.6778,
      "theoretical_loss": 3.706390548108735,
      "tokens_seen": 851277824
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037477432296890675,
      "loss": 2.6205,
      "theoretical_loss": 3.706362580845,
      "tokens_seen": 851343360
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037476429287863587,
      "loss": 2.888,
      "theoretical_loss": 3.70633461633685,
      "tokens_seen": 851408896
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003747542627883651,
      "loss": 2.4779,
      "theoretical_loss": 3.706306654583801,
      "tokens_seen": 851474432
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003747442326980943,
      "loss": 2.6989,
      "theoretical_loss": 3.706278695585369,
      "tokens_seen": 851539968
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037473420260782347,
      "loss": 2.9415,
      "theoretical_loss": 3.706250739341071,
      "tokens_seen": 851605504
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037472417251755265,
      "loss": 2.7205,
      "theoretical_loss": 3.706222785850424,
      "tokens_seen": 851671040
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037471414242728183,
      "loss": 2.6934,
      "theoretical_loss": 3.706194835112945,
      "tokens_seen": 851736576
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.000374704112337011,
      "loss": 2.6926,
      "theoretical_loss": 3.706166887128151,
      "tokens_seen": 851802112
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037469408224674025,
      "loss": 2.8434,
      "theoretical_loss": 3.7061389418955586,
      "tokens_seen": 851867648
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003746840521564694,
      "loss": 2.5986,
      "theoretical_loss": 3.706110999414686,
      "tokens_seen": 851933184
    },
    {
      "epoch": 2.08,
      "objective/train/docs_used": 948304,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5480265617370605,
      "objective/train/theoretical_loss": 3.706097029205993,
      "objective/train/tokens_used": 872425952,
      "theoretical_loss": 3.706097029205993,
      "tokens_seen": 851965952
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003746740220661986,
      "loss": 2.8042,
      "theoretical_loss": 3.7060830596850494,
      "tokens_seen": 851998720
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037466399197592774,
      "loss": 2.5988,
      "theoretical_loss": 3.706055122706167,
      "tokens_seen": 852064256
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.000374653961885657,
      "loss": 2.5353,
      "theoretical_loss": 3.706027188477557,
      "tokens_seen": 852129792
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037464393179538615,
      "loss": 2.8022,
      "theoretical_loss": 3.7059992569987354,
      "tokens_seen": 852195328
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037463390170511534,
      "loss": 2.8748,
      "theoretical_loss": 3.7059713282692224,
      "tokens_seen": 852260864
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037462387161484457,
      "loss": 2.779,
      "theoretical_loss": 3.7059434022885345,
      "tokens_seen": 852326400
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037461384152457375,
      "loss": 2.7233,
      "theoretical_loss": 3.7059154790561895,
      "tokens_seen": 852391936
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037460381143430293,
      "loss": 2.7193,
      "theoretical_loss": 3.705887558571707,
      "tokens_seen": 852457472
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003745937813440321,
      "loss": 2.6575,
      "theoretical_loss": 3.7058596408346043,
      "tokens_seen": 852523008
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003745837512537613,
      "loss": 2.738,
      "theoretical_loss": 3.705831725844401,
      "tokens_seen": 852588544
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003745737211634905,
      "loss": 2.6115,
      "theoretical_loss": 3.7058038136006144,
      "tokens_seen": 852654080
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003745636910732197,
      "loss": 2.5825,
      "theoretical_loss": 3.705775904102764,
      "tokens_seen": 852719616
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037455366098294884,
      "loss": 2.6179,
      "theoretical_loss": 3.705747997350368,
      "tokens_seen": 852785152
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003745436308926781,
      "loss": 2.7327,
      "theoretical_loss": 3.7057200933429466,
      "tokens_seen": 852850688
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003745336008024072,
      "loss": 2.8464,
      "theoretical_loss": 3.7056921920800177,
      "tokens_seen": 852916224
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037452357071213644,
      "loss": 2.8869,
      "theoretical_loss": 3.7056642935611013,
      "tokens_seen": 852981760
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003745135406218656,
      "loss": 2.6987,
      "theoretical_loss": 3.7056363977857156,
      "tokens_seen": 853047296
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003745035105315948,
      "loss": 2.8166,
      "theoretical_loss": 3.7056085047533815,
      "tokens_seen": 853112832
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.000374493480441324,
      "loss": 2.4584,
      "theoretical_loss": 3.705580614463618,
      "tokens_seen": 853178368
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003744834503510532,
      "loss": 2.803,
      "theoretical_loss": 3.7055527269159443,
      "tokens_seen": 853243904
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037447342026078234,
      "loss": 2.7901,
      "theoretical_loss": 3.705524842109881,
      "tokens_seen": 853309440
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003744633901705116,
      "loss": 2.9767,
      "theoretical_loss": 3.7054969600449477,
      "tokens_seen": 853374976
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003744533600802407,
      "loss": 2.7746,
      "theoretical_loss": 3.705469080720664,
      "tokens_seen": 853440512
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037444332998996994,
      "loss": 2.8606,
      "theoretical_loss": 3.7054412041365508,
      "tokens_seen": 853506048
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003744332998996991,
      "loss": 2.6336,
      "theoretical_loss": 3.705413330292128,
      "tokens_seen": 853571584
    },
    {
      "epoch": 2.08,
      "objective/train/docs_used": 949423,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8431296348571777,
      "objective/train/theoretical_loss": 3.7053993943971513,
      "objective/train/tokens_used": 874064352,
      "theoretical_loss": 3.7053993943971513,
      "tokens_seen": 853604352
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003744232698094283,
      "loss": 2.6529,
      "theoretical_loss": 3.7053854591869166,
      "tokens_seen": 853637120
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003744132397191575,
      "loss": 3.0092,
      "theoretical_loss": 3.7053575908204364,
      "tokens_seen": 853702656
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037440320962888666,
      "loss": 2.6759,
      "theoretical_loss": 3.705329725192209,
      "tokens_seen": 853768192
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037439317953861585,
      "loss": 2.5965,
      "theoretical_loss": 3.705301862301754,
      "tokens_seen": 853833728
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003743831494483451,
      "loss": 2.7665,
      "theoretical_loss": 3.7052740021485926,
      "tokens_seen": 853899264
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003743731193580742,
      "loss": 3.0544,
      "theoretical_loss": 3.7052461447322464,
      "tokens_seen": 853964800
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037436308926780344,
      "loss": 2.8876,
      "theoretical_loss": 3.705218290052237,
      "tokens_seen": 854030336
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037435305917753257,
      "loss": 2.7202,
      "theoretical_loss": 3.705190438108084,
      "tokens_seen": 854095872
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003743430290872618,
      "loss": 2.7702,
      "theoretical_loss": 3.7051625888993103,
      "tokens_seen": 854161408
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.000374332998996991,
      "loss": 2.811,
      "theoretical_loss": 3.7051347424254364,
      "tokens_seen": 854226944
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037432296890672017,
      "loss": 3.0093,
      "theoretical_loss": 3.705106898685985,
      "tokens_seen": 854292480
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037431293881644935,
      "loss": 2.5793,
      "theoretical_loss": 3.7050790576804764,
      "tokens_seen": 854358016
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003743029087261786,
      "loss": 2.5169,
      "theoretical_loss": 3.7050512194084337,
      "tokens_seen": 854423552
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003742928786359077,
      "loss": 2.7798,
      "theoretical_loss": 3.7050233838693787,
      "tokens_seen": 854489088
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037428284854563695,
      "loss": 2.5137,
      "theoretical_loss": 3.704995551062833,
      "tokens_seen": 854554624
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037427281845536607,
      "loss": 2.6821,
      "theoretical_loss": 3.7049677209883196,
      "tokens_seen": 854620160
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003742627883650953,
      "loss": 2.8419,
      "theoretical_loss": 3.70493989364536,
      "tokens_seen": 854685696
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003742527582748245,
      "loss": 2.8814,
      "theoretical_loss": 3.704912069033477,
      "tokens_seen": 854751232
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037424272818455367,
      "loss": 2.6939,
      "theoretical_loss": 3.7048842471521937,
      "tokens_seen": 854816768
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037423269809428285,
      "loss": 2.8317,
      "theoretical_loss": 3.7048564280010323,
      "tokens_seen": 854882304
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037422266800401203,
      "loss": 2.9259,
      "theoretical_loss": 3.704828611579515,
      "tokens_seen": 854947840
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003742126379137412,
      "loss": 2.9123,
      "theoretical_loss": 3.704800797887166,
      "tokens_seen": 855013376
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037420260782347045,
      "loss": 2.8707,
      "theoretical_loss": 3.7047729869235084,
      "tokens_seen": 855078912
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003741925777331996,
      "loss": 2.7544,
      "theoretical_loss": 3.7047451786880643,
      "tokens_seen": 855144448
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003741825476429288,
      "loss": 2.8984,
      "theoretical_loss": 3.7047173731803573,
      "tokens_seen": 855209984
    },
    {
      "epoch": 2.08,
      "objective/train/docs_used": 949852,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5512642860412598,
      "objective/train/theoretical_loss": 3.704703471449257,
      "objective/train/tokens_used": 875702752,
      "theoretical_loss": 3.704703471449257,
      "tokens_seen": 855242752
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037417251755265794,
      "loss": 2.6332,
      "theoretical_loss": 3.7046895703999114,
      "tokens_seen": 855275520
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003741624874623872,
      "loss": 2.7775,
      "theoretical_loss": 3.7046617703462497,
      "tokens_seen": 855341056
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037415245737211635,
      "loss": 2.828,
      "theoretical_loss": 3.7046339730188964,
      "tokens_seen": 855406592
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037414242728184554,
      "loss": 2.6766,
      "theoretical_loss": 3.7046061784173743,
      "tokens_seen": 855472128
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003741323971915747,
      "loss": 2.531,
      "theoretical_loss": 3.7045783865412085,
      "tokens_seen": 855537664
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037412236710130395,
      "loss": 2.6946,
      "theoretical_loss": 3.704550597389922,
      "tokens_seen": 855603200
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003741123370110331,
      "loss": 2.7786,
      "theoretical_loss": 3.7045228109630397,
      "tokens_seen": 855668736
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003741023069207623,
      "loss": 2.8376,
      "theoretical_loss": 3.7044950272600854,
      "tokens_seen": 855734272
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037409227683049144,
      "loss": 2.8049,
      "theoretical_loss": 3.704467246280584,
      "tokens_seen": 855799808
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003740822467402207,
      "loss": 2.8418,
      "theoretical_loss": 3.7044394680240593,
      "tokens_seen": 855865344
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037407221664994986,
      "loss": 2.5376,
      "theoretical_loss": 3.7044116924900363,
      "tokens_seen": 855930880
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037406218655967904,
      "loss": 2.565,
      "theoretical_loss": 3.7043839196780404,
      "tokens_seen": 855996416
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003740521564694082,
      "loss": 2.7114,
      "theoretical_loss": 3.7043561495875954,
      "tokens_seen": 856061952
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003740421263791374,
      "loss": 2.8716,
      "theoretical_loss": 3.7043283822182262,
      "tokens_seen": 856127488
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003740320962888666,
      "loss": 2.7345,
      "theoretical_loss": 3.704300617569459,
      "tokens_seen": 856193024
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003740220661985958,
      "loss": 2.7921,
      "theoretical_loss": 3.704272855640818,
      "tokens_seen": 856258560
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037401203610832494,
      "loss": 2.8099,
      "theoretical_loss": 3.70424509643183,
      "tokens_seen": 856324096
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003740020060180542,
      "loss": 2.6542,
      "theoretical_loss": 3.7042173399420184,
      "tokens_seen": 856389632
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037399197592778336,
      "loss": 2.5126,
      "theoretical_loss": 3.7041895861709104,
      "tokens_seen": 856455168
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037398194583751254,
      "loss": 2.8828,
      "theoretical_loss": 3.7041618351180308,
      "tokens_seen": 856520704
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003739719157472417,
      "loss": 2.5162,
      "theoretical_loss": 3.704134086782906,
      "tokens_seen": 856586240
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003739618856569709,
      "loss": 2.7155,
      "theoretical_loss": 3.704106341165062,
      "tokens_seen": 856651776
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003739518555667001,
      "loss": 2.8485,
      "theoretical_loss": 3.7040785982640236,
      "tokens_seen": 856717312
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003739418254764293,
      "loss": 2.7267,
      "theoretical_loss": 3.7040508580793188,
      "tokens_seen": 856782848
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037393179538615845,
      "loss": 2.8417,
      "theoretical_loss": 3.704023120610473,
      "tokens_seen": 856848384
    },
    {
      "epoch": 2.08,
      "objective/train/docs_used": 951087,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8513641357421875,
      "objective/train/theoretical_loss": 3.7040092528943487,
      "objective/train/tokens_used": 877341152,
      "theoretical_loss": 3.7040092528943487,
      "tokens_seen": 856881152
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003739217652958877,
      "loss": 2.7939,
      "theoretical_loss": 3.703995385857012,
      "tokens_seen": 856913920
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003739117352056168,
      "loss": 2.8448,
      "theoretical_loss": 3.703967653818464,
      "tokens_seen": 856979456
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037390170511534605,
      "loss": 2.7197,
      "theoretical_loss": 3.703939924494353,
      "tokens_seen": 857044992
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003738916750250752,
      "loss": 2.5392,
      "theoretical_loss": 3.703912197884209,
      "tokens_seen": 857110528
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003738816449348044,
      "loss": 2.761,
      "theoretical_loss": 3.7038844739875567,
      "tokens_seen": 857176064
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037387161484453364,
      "loss": 2.6546,
      "theoretical_loss": 3.703856752803924,
      "tokens_seen": 857241600
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037386158475426277,
      "loss": 2.7587,
      "theoretical_loss": 3.7038290343328377,
      "tokens_seen": 857307136
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.000373851554663992,
      "loss": 2.7639,
      "theoretical_loss": 3.7038013185738246,
      "tokens_seen": 857372672
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003738415245737212,
      "loss": 2.6566,
      "theoretical_loss": 3.703773605526413,
      "tokens_seen": 857438208
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037383149448345037,
      "loss": 2.8388,
      "theoretical_loss": 3.7037458951901296,
      "tokens_seen": 857503744
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037382146439317955,
      "loss": 2.6906,
      "theoretical_loss": 3.703718187564503,
      "tokens_seen": 857569280
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003738114343029088,
      "loss": 2.6582,
      "theoretical_loss": 3.7036904826490598,
      "tokens_seen": 857634816
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003738014042126379,
      "loss": 2.8538,
      "theoretical_loss": 3.703662780443328,
      "tokens_seen": 857700352
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037379137412236715,
      "loss": 2.6117,
      "theoretical_loss": 3.7036350809468366,
      "tokens_seen": 857765888
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037378134403209627,
      "loss": 2.6217,
      "theoretical_loss": 3.7036073841591124,
      "tokens_seen": 857831424
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003737713139418255,
      "loss": 2.6182,
      "theoretical_loss": 3.7035796900796845,
      "tokens_seen": 857896960
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003737612838515547,
      "loss": 2.8175,
      "theoretical_loss": 3.7035519987080807,
      "tokens_seen": 857962496
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037375125376128387,
      "loss": 2.9811,
      "theoretical_loss": 3.7035243100438295,
      "tokens_seen": 858028032
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037374122367101305,
      "loss": 2.9455,
      "theoretical_loss": 3.7034966240864593,
      "tokens_seen": 858093568
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037373119358074223,
      "loss": 2.9907,
      "theoretical_loss": 3.7034689408354993,
      "tokens_seen": 858159104
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003737211634904714,
      "loss": 2.851,
      "theoretical_loss": 3.7034412602904783,
      "tokens_seen": 858224640
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037371113340020065,
      "loss": 2.6784,
      "theoretical_loss": 3.7034135824509242,
      "tokens_seen": 858290176
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003737011033099298,
      "loss": 2.8578,
      "theoretical_loss": 3.703385907316367,
      "tokens_seen": 858355712
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.000373691073219659,
      "loss": 2.8543,
      "theoretical_loss": 3.703358234886336,
      "tokens_seen": 858421248
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037368104312938814,
      "loss": 2.9281,
      "theoretical_loss": 3.703330565160359,
      "tokens_seen": 858486784
    },
    {
      "epoch": 2.08,
      "objective/train/docs_used": 951523,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.799268960952759,
      "objective/train/theoretical_loss": 3.7033167313112445,
      "objective/train/tokens_used": 878979552,
      "theoretical_loss": 3.7033167313112445,
      "tokens_seen": 858519552
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003736710130391174,
      "loss": 2.8842,
      "theoretical_loss": 3.7033028981379674,
      "tokens_seen": 858552320
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037366098294884655,
      "loss": 2.5017,
      "theoretical_loss": 3.703275233818689,
      "tokens_seen": 858617856
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037365095285857574,
      "loss": 2.7812,
      "theoretical_loss": 3.7032475722020544,
      "tokens_seen": 858683392
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003736409227683049,
      "loss": 3.0125,
      "theoretical_loss": 3.703219913287593,
      "tokens_seen": 858748928
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037363089267803415,
      "loss": 2.855,
      "theoretical_loss": 3.7031922570748343,
      "tokens_seen": 858814464
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003736208625877633,
      "loss": 2.7978,
      "theoretical_loss": 3.703164603563309,
      "tokens_seen": 858880000
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003736108324974925,
      "loss": 2.3875,
      "theoretical_loss": 3.7031369527525464,
      "tokens_seen": 858945536
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037360080240722164,
      "loss": 2.9526,
      "theoretical_loss": 3.7031093046420773,
      "tokens_seen": 859011072
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003735907723169509,
      "loss": 2.5221,
      "theoretical_loss": 3.703081659231432,
      "tokens_seen": 859076608
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037358074222668006,
      "loss": 2.9902,
      "theoretical_loss": 3.7030540165201407,
      "tokens_seen": 859142144
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037357071213640924,
      "loss": 2.5429,
      "theoretical_loss": 3.7030263765077334,
      "tokens_seen": 859207680
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003735606820461384,
      "loss": 2.673,
      "theoretical_loss": 3.702998739193742,
      "tokens_seen": 859273216
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003735506519558676,
      "loss": 3.012,
      "theoretical_loss": 3.7029711045776965,
      "tokens_seen": 859338752
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003735406218655968,
      "loss": 2.6181,
      "theoretical_loss": 3.7029434726591277,
      "tokens_seen": 859404288
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.000373530591775326,
      "loss": 2.7048,
      "theoretical_loss": 3.702915843437567,
      "tokens_seen": 859469824
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037352056168505514,
      "loss": 2.5463,
      "theoretical_loss": 3.7028882169125454,
      "tokens_seen": 859535360
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003735105315947844,
      "loss": 2.394,
      "theoretical_loss": 3.7028605930835945,
      "tokens_seen": 859600896
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037350050150451356,
      "loss": 2.7804,
      "theoretical_loss": 3.702832971950245,
      "tokens_seen": 859666432
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037349047141424274,
      "loss": 2.6185,
      "theoretical_loss": 3.702805353512029,
      "tokens_seen": 859731968
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003734804413239719,
      "loss": 2.8201,
      "theoretical_loss": 3.7027777377684776,
      "tokens_seen": 859797504
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003734704112337011,
      "loss": 3.0059,
      "theoretical_loss": 3.7027501247191226,
      "tokens_seen": 859863040
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003734603811434303,
      "loss": 2.8458,
      "theoretical_loss": 3.7027225143634963,
      "tokens_seen": 859928576
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003734503510531595,
      "loss": 2.6558,
      "theoretical_loss": 3.7026949067011303,
      "tokens_seen": 859994112
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037344032096288865,
      "loss": 2.7805,
      "theoretical_loss": 3.7026673017315566,
      "tokens_seen": 860059648
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003734302908726179,
      "loss": 2.6283,
      "theoretical_loss": 3.7026396994543074,
      "tokens_seen": 860125184
    },
    {
      "epoch": 2.08,
      "objective/train/docs_used": 952944,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6841440200805664,
      "objective/train/theoretical_loss": 3.7026258993251586,
      "objective/train/tokens_used": 880617952,
      "theoretical_loss": 3.7026258993251586,
      "tokens_seen": 860157952
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.000373420260782347,
      "loss": 2.8276,
      "theoretical_loss": 3.702612099868915,
      "tokens_seen": 860190720
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037341023069207625,
      "loss": 2.9939,
      "theoretical_loss": 3.7025845029749123,
      "tokens_seen": 860256256
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003734002006018054,
      "loss": 2.696,
      "theoretical_loss": 3.7025569087718315,
      "tokens_seen": 860321792
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003733901705115346,
      "loss": 2.8568,
      "theoretical_loss": 3.7025293172592053,
      "tokens_seen": 860387328
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003733801404212638,
      "loss": 2.6686,
      "theoretical_loss": 3.702501728436566,
      "tokens_seen": 860452864
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037337011033099297,
      "loss": 2.752,
      "theoretical_loss": 3.7024741423034477,
      "tokens_seen": 860518400
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037336008024072215,
      "loss": 2.6242,
      "theoretical_loss": 3.702446558859382,
      "tokens_seen": 860583936
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003733500501504514,
      "loss": 2.7184,
      "theoretical_loss": 3.702418978103903,
      "tokens_seen": 860649472
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003733400200601805,
      "loss": 2.8764,
      "theoretical_loss": 3.702391400036543,
      "tokens_seen": 860715008
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037332998996990975,
      "loss": 2.5713,
      "theoretical_loss": 3.7023638246568367,
      "tokens_seen": 860780544
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037331995987963893,
      "loss": 2.7039,
      "theoretical_loss": 3.702336251964317,
      "tokens_seen": 860846080
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003733099297893681,
      "loss": 2.7103,
      "theoretical_loss": 3.7023086819585167,
      "tokens_seen": 860911616
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003732998996990973,
      "loss": 2.808,
      "theoretical_loss": 3.7022811146389696,
      "tokens_seen": 860977152
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003732898696088265,
      "loss": 2.9242,
      "theoretical_loss": 3.702253550005211,
      "tokens_seen": 861042688
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037327983951855565,
      "loss": 2.7875,
      "theoretical_loss": 3.7022259880567736,
      "tokens_seen": 861108224
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003732698094282849,
      "loss": 2.8359,
      "theoretical_loss": 3.702198428793192,
      "tokens_seen": 861173760
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.000373259779338014,
      "loss": 2.7244,
      "theoretical_loss": 3.702170872214,
      "tokens_seen": 861239296
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037324974924774325,
      "loss": 2.8139,
      "theoretical_loss": 3.702143318318732,
      "tokens_seen": 861304832
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003732397191574724,
      "loss": 2.6474,
      "theoretical_loss": 3.702115767106922,
      "tokens_seen": 861370368
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003732296890672016,
      "loss": 2.7847,
      "theoretical_loss": 3.7020882185781057,
      "tokens_seen": 861435904
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003732196589769308,
      "loss": 2.8831,
      "theoretical_loss": 3.7020606727318164,
      "tokens_seen": 861501440
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037320962888666,
      "loss": 2.6749,
      "theoretical_loss": 3.7020331295675897,
      "tokens_seen": 861566976
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037319959879638916,
      "loss": 2.5583,
      "theoretical_loss": 3.70200558908496,
      "tokens_seen": 861632512
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037318956870611834,
      "loss": 2.8254,
      "theoretical_loss": 3.7019780512834632,
      "tokens_seen": 861698048
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003731795386158475,
      "loss": 2.7722,
      "theoretical_loss": 3.7019505161626327,
      "tokens_seen": 861763584
    },
    {
      "epoch": 2.08,
      "objective/train/docs_used": 953723,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7122349739074707,
      "objective/train/theoretical_loss": 3.7019367496073228,
      "objective/train/tokens_used": 882256352,
      "theoretical_loss": 3.7019367496073228,
      "tokens_seen": 861796352
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037316950852557675,
      "loss": 2.8407,
      "theoretical_loss": 3.701922983722006,
      "tokens_seen": 861829120
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003731594784353059,
      "loss": 2.7291,
      "theoretical_loss": 3.701895453961116,
      "tokens_seen": 861894656
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003731494483450351,
      "loss": 2.7609,
      "theoretical_loss": 3.7018679268794994,
      "tokens_seen": 861960192
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003731394182547643,
      "loss": 2.7356,
      "theoretical_loss": 3.701840402476692,
      "tokens_seen": 862025728
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003731293881644935,
      "loss": 2.8384,
      "theoretical_loss": 3.7018128807522297,
      "tokens_seen": 862091264
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003731193580742227,
      "loss": 2.7187,
      "theoretical_loss": 3.701785361705647,
      "tokens_seen": 862156800
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037310932798395184,
      "loss": 2.7171,
      "theoretical_loss": 3.7017578453364806,
      "tokens_seen": 862222336
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003730992978936811,
      "loss": 2.8718,
      "theoretical_loss": 3.701730331644267,
      "tokens_seen": 862287872
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037308926780341026,
      "loss": 2.8957,
      "theoretical_loss": 3.701702820628541,
      "tokens_seen": 862353408
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037307923771313944,
      "loss": 2.9734,
      "theoretical_loss": 3.701675312288841,
      "tokens_seen": 862418944
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003730692076228686,
      "loss": 2.8563,
      "theoretical_loss": 3.701647806624701,
      "tokens_seen": 862484480
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003730591775325978,
      "loss": 2.724,
      "theoretical_loss": 3.701620303635659,
      "tokens_seen": 862550016
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.000373049147442327,
      "loss": 2.676,
      "theoretical_loss": 3.701592803321251,
      "tokens_seen": 862615552
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003730391173520562,
      "loss": 2.7422,
      "theoretical_loss": 3.7015653056810143,
      "tokens_seen": 862681088
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037302908726178534,
      "loss": 2.8457,
      "theoretical_loss": 3.701537810714485,
      "tokens_seen": 862746624
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003730190571715146,
      "loss": 2.6289,
      "theoretical_loss": 3.7015103184212004,
      "tokens_seen": 862812160
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037300902708124376,
      "loss": 2.7261,
      "theoretical_loss": 3.7014828288006973,
      "tokens_seen": 862877696
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037299899699097294,
      "loss": 2.7542,
      "theoretical_loss": 3.7014553418525136,
      "tokens_seen": 862943232
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003729889669007021,
      "loss": 2.5174,
      "theoretical_loss": 3.701427857576186,
      "tokens_seen": 863008768
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003729789368104313,
      "loss": 2.8556,
      "theoretical_loss": 3.701400375971252,
      "tokens_seen": 863074304
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003729689067201605,
      "loss": 2.8079,
      "theoretical_loss": 3.701372897037249,
      "tokens_seen": 863139840
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003729588766298897,
      "loss": 2.7774,
      "theoretical_loss": 3.7013454207737153,
      "tokens_seen": 863205376
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037294884653961885,
      "loss": 2.8166,
      "theoretical_loss": 3.701317947180188,
      "tokens_seen": 863270912
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003729388164493481,
      "loss": 2.518,
      "theoretical_loss": 3.7012904762562053,
      "tokens_seen": 863336448
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003729287863590772,
      "loss": 2.8338,
      "theoretical_loss": 3.701263008001305,
      "tokens_seen": 863401984
    },
    {
      "epoch": 2.08,
      "objective/train/docs_used": 954800,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.510504961013794,
      "objective/train/theoretical_loss": 3.701249274874616,
      "objective/train/tokens_used": 883894752,
      "theoretical_loss": 3.701249274874616,
      "tokens_seen": 863434752
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037291875626880645,
      "loss": 2.7716,
      "theoretical_loss": 3.7012355424150254,
      "tokens_seen": 863467520
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003729087261785356,
      "loss": 2.7719,
      "theoretical_loss": 3.7012080794969044,
      "tokens_seen": 863533056
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003728986960882648,
      "loss": 2.7007,
      "theoretical_loss": 3.7011806192464807,
      "tokens_seen": 863598592
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.000372888665997994,
      "loss": 2.5711,
      "theoretical_loss": 3.7011531616632922,
      "tokens_seen": 863664128
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037287863590772317,
      "loss": 2.7918,
      "theoretical_loss": 3.701125706746878,
      "tokens_seen": 863729664
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037286860581745235,
      "loss": 2.8889,
      "theoretical_loss": 3.701098254496777,
      "tokens_seen": 863795200
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003728585757271816,
      "loss": 2.9669,
      "theoretical_loss": 3.7010708049125274,
      "tokens_seen": 863860736
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003728485456369107,
      "loss": 2.6103,
      "theoretical_loss": 3.701043357993668,
      "tokens_seen": 863926272
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037283851554663995,
      "loss": 2.7198,
      "theoretical_loss": 3.701015913739739,
      "tokens_seen": 863991808
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037282848545636913,
      "loss": 2.7146,
      "theoretical_loss": 3.7009884721502777,
      "tokens_seen": 864057344
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003728184553660983,
      "loss": 2.8361,
      "theoretical_loss": 3.700961033224825,
      "tokens_seen": 864122880
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003728084252758275,
      "loss": 2.7008,
      "theoretical_loss": 3.7009335969629196,
      "tokens_seen": 864188416
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003727983951855567,
      "loss": 2.7614,
      "theoretical_loss": 3.7009061633641007,
      "tokens_seen": 864253952
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037278836509528585,
      "loss": 2.7302,
      "theoretical_loss": 3.700878732427908,
      "tokens_seen": 864319488
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003727783350050151,
      "loss": 2.6105,
      "theoretical_loss": 3.7008513041538817,
      "tokens_seen": 864385024
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003727683049147442,
      "loss": 2.9334,
      "theoretical_loss": 3.700823878541562,
      "tokens_seen": 864450560
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037275827482447345,
      "loss": 2.7052,
      "theoretical_loss": 3.7007964555904875,
      "tokens_seen": 864516096
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003727482447342026,
      "loss": 2.8405,
      "theoretical_loss": 3.7007690353001985,
      "tokens_seen": 864581632
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003727382146439318,
      "loss": 2.6395,
      "theoretical_loss": 3.700741617670236,
      "tokens_seen": 864647168
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.000372728184553661,
      "loss": 2.6473,
      "theoretical_loss": 3.70071420270014,
      "tokens_seen": 864712704
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003727181544633902,
      "loss": 2.8752,
      "theoretical_loss": 3.700686790389451,
      "tokens_seen": 864778240
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037270812437311936,
      "loss": 2.852,
      "theoretical_loss": 3.700659380737709,
      "tokens_seen": 864843776
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037269809428284854,
      "loss": 2.9933,
      "theoretical_loss": 3.7006319737444553,
      "tokens_seen": 864909312
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003726880641925777,
      "loss": 2.5759,
      "theoretical_loss": 3.7006045694092298,
      "tokens_seen": 864974848
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037267803410230696,
      "loss": 2.9994,
      "theoretical_loss": 3.700577167731574,
      "tokens_seen": 865040384
    },
    {
      "epoch": 2.08,
      "objective/train/docs_used": 955501,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.222694158554077,
      "objective/train/theoretical_loss": 3.700563467889191,
      "objective/train/tokens_used": 885533152,
      "theoretical_loss": 3.700563467889191,
      "tokens_seen": 865073152
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003726680040120361,
      "loss": 2.7528,
      "theoretical_loss": 3.7005497687110287,
      "tokens_seen": 865105920
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003726579739217653,
      "loss": 2.7434,
      "theoretical_loss": 3.7005223723471348,
      "tokens_seen": 865171456
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003726479438314945,
      "loss": 2.5054,
      "theoretical_loss": 3.700494978639434,
      "tokens_seen": 865236992
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003726379137412237,
      "loss": 2.7522,
      "theoretical_loss": 3.700467587587467,
      "tokens_seen": 865302528
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037262788365095286,
      "loss": 2.593,
      "theoretical_loss": 3.7004401991907754,
      "tokens_seen": 865368064
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037261785356068204,
      "loss": 2.4979,
      "theoretical_loss": 3.700412813448901,
      "tokens_seen": 865433600
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003726078234704112,
      "loss": 2.7374,
      "theoretical_loss": 3.7003854303613855,
      "tokens_seen": 865499136
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037259779338014046,
      "loss": 2.7745,
      "theoretical_loss": 3.70035804992777,
      "tokens_seen": 865564672
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003725877632898696,
      "loss": 2.8962,
      "theoretical_loss": 3.7003306721475973,
      "tokens_seen": 865630208
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003725777331995988,
      "loss": 2.8755,
      "theoretical_loss": 3.700303297020409,
      "tokens_seen": 865695744
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037256770310932795,
      "loss": 2.744,
      "theoretical_loss": 3.700275924545747,
      "tokens_seen": 865761280
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003725576730190572,
      "loss": 2.8406,
      "theoretical_loss": 3.7002485547231534,
      "tokens_seen": 865826816
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037254764292878636,
      "loss": 2.3279,
      "theoretical_loss": 3.700221187552171,
      "tokens_seen": 865892352
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037253761283851554,
      "loss": 2.7378,
      "theoretical_loss": 3.7001938230323423,
      "tokens_seen": 865957888
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003725275827482447,
      "loss": 2.1945,
      "theoretical_loss": 3.7001664611632092,
      "tokens_seen": 866023424
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037251755265797396,
      "loss": 2.6143,
      "theoretical_loss": 3.700139101944315,
      "tokens_seen": 866088960
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003725075225677031,
      "loss": 2.816,
      "theoretical_loss": 3.700111745375202,
      "tokens_seen": 866154496
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003724974924774323,
      "loss": 2.4319,
      "theoretical_loss": 3.700084391455414,
      "tokens_seen": 866220032
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037248746238716145,
      "loss": 2.5547,
      "theoretical_loss": 3.7000570401844928,
      "tokens_seen": 866285568
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003724774322968907,
      "loss": 2.7006,
      "theoretical_loss": 3.7000296915619826,
      "tokens_seen": 866351104
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037246740220661987,
      "loss": 2.6894,
      "theoretical_loss": 3.700002345587426,
      "tokens_seen": 866416640
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037245737211634905,
      "loss": 2.7271,
      "theoretical_loss": 3.6999750022603664,
      "tokens_seen": 866482176
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037244734202607823,
      "loss": 2.7118,
      "theoretical_loss": 3.6999476615803473,
      "tokens_seen": 866547712
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003724373119358074,
      "loss": 2.7494,
      "theoretical_loss": 3.6999203235469125,
      "tokens_seen": 866613248
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003724272818455366,
      "loss": 2.8457,
      "theoretical_loss": 3.699892988159606,
      "tokens_seen": 866678784
    },
    {
      "epoch": 2.08,
      "objective/train/docs_used": 956034,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.010556221008301,
      "objective/train/theoretical_loss": 3.6998793214581074,
      "objective/train/tokens_used": 887171552,
      "theoretical_loss": 3.6998793214581074,
      "tokens_seen": 866711552
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003724172517552658,
      "loss": 2.8745,
      "theoretical_loss": 3.6998656554179705,
      "tokens_seen": 866744320
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037240722166499495,
      "loss": 2.7951,
      "theoretical_loss": 3.6998383253215508,
      "tokens_seen": 866809856
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003723971915747242,
      "loss": 2.5935,
      "theoretical_loss": 3.699810997869891,
      "tokens_seen": 866875392
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003723871614844533,
      "loss": 2.6498,
      "theoretical_loss": 3.6997836730625346,
      "tokens_seen": 866940928
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037237713139418255,
      "loss": 2.6788,
      "theoretical_loss": 3.6997563508990265,
      "tokens_seen": 867006464
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003723671013039118,
      "loss": 2.5869,
      "theoretical_loss": 3.6997290313789106,
      "tokens_seen": 867072000
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003723570712136409,
      "loss": 2.6665,
      "theoretical_loss": 3.6997017145017326,
      "tokens_seen": 867137536
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037234704112337015,
      "loss": 2.621,
      "theoretical_loss": 3.699674400267035,
      "tokens_seen": 867203072
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037233701103309933,
      "loss": 2.9424,
      "theoretical_loss": 3.699647088674364,
      "tokens_seen": 867268608
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003723269809428285,
      "loss": 2.6496,
      "theoretical_loss": 3.699619779723264,
      "tokens_seen": 867334144
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003723169508525577,
      "loss": 2.519,
      "theoretical_loss": 3.69959247341328,
      "tokens_seen": 867399680
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003723069207622869,
      "loss": 2.5122,
      "theoretical_loss": 3.6995651697439573,
      "tokens_seen": 867465216
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037229689067201605,
      "loss": 2.5631,
      "theoretical_loss": 3.699537868714841,
      "tokens_seen": 867530752
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003722868605817453,
      "loss": 2.5265,
      "theoretical_loss": 3.699510570325475,
      "tokens_seen": 867596288
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003722768304914744,
      "loss": 2.7926,
      "theoretical_loss": 3.699483274575407,
      "tokens_seen": 867661824
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037226680040120365,
      "loss": 2.9379,
      "theoretical_loss": 3.6994559814641814,
      "tokens_seen": 867727360
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003722567703109328,
      "loss": 2.8544,
      "theoretical_loss": 3.6994286909913434,
      "tokens_seen": 867792896
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.000372246740220662,
      "loss": 2.4664,
      "theoretical_loss": 3.699401403156439,
      "tokens_seen": 867858432
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003722367101303912,
      "loss": 2.628,
      "theoretical_loss": 3.6993741179590143,
      "tokens_seen": 867923968
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003722266800401204,
      "loss": 2.5972,
      "theoretical_loss": 3.6993468353986154,
      "tokens_seen": 867989504
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037221664994984956,
      "loss": 2.7708,
      "theoretical_loss": 3.699319555474788,
      "tokens_seen": 868055040
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037220661985957874,
      "loss": 2.5467,
      "theoretical_loss": 3.6992922781870776,
      "tokens_seen": 868120576
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003721965897693079,
      "loss": 2.7708,
      "theoretical_loss": 3.6992650035350323,
      "tokens_seen": 868186112
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037218655967903716,
      "loss": 2.4194,
      "theoretical_loss": 3.699237731518197,
      "tokens_seen": 868251648
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003721765295887663,
      "loss": 2.4483,
      "theoretical_loss": 3.6992104621361186,
      "tokens_seen": 868317184
    },
    {
      "epoch": 2.08,
      "objective/train/docs_used": 959586,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.304150342941284,
      "objective/train/theoretical_loss": 3.6991968284329717,
      "objective/train/tokens_used": 888809952,
      "theoretical_loss": 3.6991968284329717,
      "tokens_seen": 868349952
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003721664994984955,
      "loss": 2.646,
      "theoretical_loss": 3.699183195388344,
      "tokens_seen": 868382720
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003721564694082247,
      "loss": 2.6881,
      "theoretical_loss": 3.6991559312744196,
      "tokens_seen": 868448256
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003721464393179539,
      "loss": 2.7205,
      "theoretical_loss": 3.6991286697938923,
      "tokens_seen": 868513792
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037213640922768306,
      "loss": 2.9575,
      "theoretical_loss": 3.6991014109463096,
      "tokens_seen": 868579328
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037212637913741224,
      "loss": 2.7497,
      "theoretical_loss": 3.6990741547312176,
      "tokens_seen": 868644864
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003721163490471414,
      "loss": 2.6964,
      "theoretical_loss": 3.699046901148164,
      "tokens_seen": 868710400
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037210631895687066,
      "loss": 2.8692,
      "theoretical_loss": 3.6990196501966963,
      "tokens_seen": 868775936
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003720962888665998,
      "loss": 2.6819,
      "theoretical_loss": 3.6989924018763616,
      "tokens_seen": 868841472
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.000372086258776329,
      "loss": 2.6184,
      "theoretical_loss": 3.6989651561867074,
      "tokens_seen": 868907008
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037207622868605815,
      "loss": 2.6148,
      "theoretical_loss": 3.698937913127282,
      "tokens_seen": 868972544
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003720661985957874,
      "loss": 2.5731,
      "theoretical_loss": 3.6989106726976324,
      "tokens_seen": 869038080
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037205616850551656,
      "loss": 2.8067,
      "theoretical_loss": 3.6988834348973065,
      "tokens_seen": 869103616
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037204613841524575,
      "loss": 2.7593,
      "theoretical_loss": 3.698856199725853,
      "tokens_seen": 869169152
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003720361083249749,
      "loss": 2.609,
      "theoretical_loss": 3.698828967182819,
      "tokens_seen": 869234688
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037202607823470416,
      "loss": 2.7118,
      "theoretical_loss": 3.698801737267753,
      "tokens_seen": 869300224
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003720160481444333,
      "loss": 2.7569,
      "theoretical_loss": 3.698774509980203,
      "tokens_seen": 869365760
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003720060180541625,
      "loss": 2.7569,
      "theoretical_loss": 3.698747285319719,
      "tokens_seen": 869431296
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037199598796389165,
      "loss": 2.7691,
      "theoretical_loss": 3.6987200632858475,
      "tokens_seen": 869496832
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003719859578736209,
      "loss": 2.5376,
      "theoretical_loss": 3.698692843878139,
      "tokens_seen": 869562368
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037197592778335007,
      "loss": 2.8628,
      "theoretical_loss": 3.6986656270961404,
      "tokens_seen": 869627904
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037196589769307925,
      "loss": 2.5259,
      "theoretical_loss": 3.6986384129394017,
      "tokens_seen": 869693440
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037195586760280843,
      "loss": 2.6141,
      "theoretical_loss": 3.6986112014074717,
      "tokens_seen": 869758976
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003719458375125376,
      "loss": 2.789,
      "theoretical_loss": 3.6985839924998993,
      "tokens_seen": 869824512
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003719358074222668,
      "loss": 2.7556,
      "theoretical_loss": 3.698556786216234,
      "tokens_seen": 869890048
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.000371925777331996,
      "loss": 2.6361,
      "theoretical_loss": 3.6985295825560245,
      "tokens_seen": 869955584
    },
    {
      "epoch": 2.08,
      "objective/train/docs_used": 961818,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.836449146270752,
      "objective/train/theoretical_loss": 3.6985159817095754,
      "objective/train/tokens_used": 890267104,
      "theoretical_loss": 3.6985159817095754,
      "tokens_seen": 869988352
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037191574724172515,
      "loss": 2.8748,
      "theoretical_loss": 3.698502381518821,
      "tokens_seen": 870021120
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003719057171514544,
      "loss": 2.8179,
      "theoretical_loss": 3.6984751831041724,
      "tokens_seen": 870086656
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003718956870611835,
      "loss": 2.6425,
      "theoretical_loss": 3.6984479873116287,
      "tokens_seen": 870152192
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037188565697091275,
      "loss": 2.6881,
      "theoretical_loss": 3.6984207941407394,
      "tokens_seen": 870217728
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037187562688064193,
      "loss": 2.9222,
      "theoretical_loss": 3.698393603591055,
      "tokens_seen": 870283264
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003718655967903711,
      "loss": 2.7699,
      "theoretical_loss": 3.698366415662125,
      "tokens_seen": 870348800
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003718555667001003,
      "loss": 2.8366,
      "theoretical_loss": 3.698339230353499,
      "tokens_seen": 870414336
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037184553660982953,
      "loss": 2.8265,
      "theoretical_loss": 3.698312047664728,
      "tokens_seen": 870479872
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037183550651955866,
      "loss": 2.7096,
      "theoretical_loss": 3.6982848675953623,
      "tokens_seen": 870545408
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003718254764292879,
      "loss": 2.8957,
      "theoretical_loss": 3.6982576901449518,
      "tokens_seen": 870610944
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.000371815446339017,
      "loss": 2.8406,
      "theoretical_loss": 3.698230515313047,
      "tokens_seen": 870676480
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037180541624874625,
      "loss": 2.5983,
      "theoretical_loss": 3.698203343099199,
      "tokens_seen": 870742016
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037179538615847544,
      "loss": 2.5483,
      "theoretical_loss": 3.698176173502959,
      "tokens_seen": 870807552
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003717853560682046,
      "loss": 2.755,
      "theoretical_loss": 3.698149006523876,
      "tokens_seen": 870873088
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003717753259779338,
      "loss": 2.6726,
      "theoretical_loss": 3.6981218421615027,
      "tokens_seen": 870938624
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.000371765295887663,
      "loss": 2.6759,
      "theoretical_loss": 3.69809468041539,
      "tokens_seen": 871004160
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037175526579739216,
      "loss": 2.8533,
      "theoretical_loss": 3.698067521285089,
      "tokens_seen": 871069696
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003717452357071214,
      "loss": 2.6255,
      "theoretical_loss": 3.6980403647701503,
      "tokens_seen": 871135232
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003717352056168505,
      "loss": 2.589,
      "theoretical_loss": 3.698013210870126,
      "tokens_seen": 871200768
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037172517552657976,
      "loss": 2.7553,
      "theoretical_loss": 3.6979860595845673,
      "tokens_seen": 871266304
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003717151454363089,
      "loss": 2.3292,
      "theoretical_loss": 3.6979589109130258,
      "tokens_seen": 871331840
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003717051153460381,
      "loss": 2.6952,
      "theoretical_loss": 3.697931764855054,
      "tokens_seen": 871397376
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003716950852557673,
      "loss": 3.0049,
      "theoretical_loss": 3.697904621410203,
      "tokens_seen": 871462912
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003716850551654965,
      "loss": 2.7278,
      "theoretical_loss": 3.697877480578025,
      "tokens_seen": 871528448
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037167502507522566,
      "loss": 2.7773,
      "theoretical_loss": 3.6978503423580715,
      "tokens_seen": 871593984
    },
    {
      "epoch": 2.08,
      "objective/train/docs_used": 961818,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4541244506835938,
      "objective/train/theoretical_loss": 3.69783677422754,
      "objective/train/tokens_used": 890267104,
      "theoretical_loss": 3.69783677422754,
      "tokens_seen": 871626752
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003716649949849549,
      "loss": 2.7079,
      "theoretical_loss": 3.697823206749896,
      "tokens_seen": 871659520
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.000371654964894684,
      "loss": 2.6983,
      "theoretical_loss": 3.69779607375305,
      "tokens_seen": 871725056
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037164493480441326,
      "loss": 2.6245,
      "theoretical_loss": 3.6977689433670857,
      "tokens_seen": 871790592
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003716349047141424,
      "loss": 2.4994,
      "theoretical_loss": 3.6977418155915562,
      "tokens_seen": 871856128
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003716248746238716,
      "loss": 2.4224,
      "theoretical_loss": 3.6977146904260136,
      "tokens_seen": 871921664
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037161484453360086,
      "loss": 2.647,
      "theoretical_loss": 3.6976875678700107,
      "tokens_seen": 871987200
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037160481444333,
      "loss": 2.5049,
      "theoretical_loss": 3.697660447923101,
      "tokens_seen": 872052736
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003715947843530592,
      "loss": 3.0161,
      "theoretical_loss": 3.6976333305848366,
      "tokens_seen": 872118272
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037158475426278835,
      "loss": 2.8686,
      "theoretical_loss": 3.6976062158547713,
      "tokens_seen": 872183808
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003715747241725176,
      "loss": 2.3929,
      "theoretical_loss": 3.697579103732458,
      "tokens_seen": 872249344
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037156469408224676,
      "loss": 2.6487,
      "theoretical_loss": 3.6975519942174495,
      "tokens_seen": 872314880
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037155466399197595,
      "loss": 2.5502,
      "theoretical_loss": 3.6975248873093003,
      "tokens_seen": 872380416
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003715446339017051,
      "loss": 2.8222,
      "theoretical_loss": 3.6974977830075626,
      "tokens_seen": 872445952
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037153460381143436,
      "loss": 2.4791,
      "theoretical_loss": 3.697470681311791,
      "tokens_seen": 872511488
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003715245737211635,
      "loss": 2.5693,
      "theoretical_loss": 3.6974435822215392,
      "tokens_seen": 872577024
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003715145436308927,
      "loss": 2.6591,
      "theoretical_loss": 3.6974164857363605,
      "tokens_seen": 872642560
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037150451354062185,
      "loss": 2.5865,
      "theoretical_loss": 3.6973893918558094,
      "tokens_seen": 872708096
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003714944834503511,
      "loss": 2.414,
      "theoretical_loss": 3.6973623005794396,
      "tokens_seen": 872773632
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037148445336008027,
      "loss": 2.6693,
      "theoretical_loss": 3.6973352119068053,
      "tokens_seen": 872839168
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037147442326980945,
      "loss": 2.7309,
      "theoretical_loss": 3.697308125837461,
      "tokens_seen": 872904704
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037146439317953863,
      "loss": 2.6199,
      "theoretical_loss": 3.6972810423709603,
      "tokens_seen": 872970240
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003714543630892678,
      "loss": 2.5813,
      "theoretical_loss": 3.6972539615068594,
      "tokens_seen": 873035776
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.000371444332998997,
      "loss": 2.5225,
      "theoretical_loss": 3.6972268832447117,
      "tokens_seen": 873101312
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037143430290872623,
      "loss": 2.7103,
      "theoretical_loss": 3.6971998075840715,
      "tokens_seen": 873166848
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037142427281845535,
      "loss": 2.4854,
      "theoretical_loss": 3.6971727345244947,
      "tokens_seen": 873232384
    },
    {
      "epoch": 2.08,
      "objective/train/docs_used": 961818,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5809988975524902,
      "objective/train/theoretical_loss": 3.697159198969966,
      "objective/train/tokens_used": 890267104,
      "theoretical_loss": 3.697159198969966,
      "tokens_seen": 873265152
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003714142427281846,
      "loss": 2.58,
      "theoretical_loss": 3.697145664065536,
      "tokens_seen": 873297920
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003714042126379137,
      "loss": 2.6467,
      "theoretical_loss": 3.69711859620675,
      "tokens_seen": 873363456
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037139418254764295,
      "loss": 2.4065,
      "theoretical_loss": 3.6970915309476915,
      "tokens_seen": 873428992
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037138415245737213,
      "loss": 2.497,
      "theoretical_loss": 3.6970644682879175,
      "tokens_seen": 873494528
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003713741223671013,
      "loss": 2.7433,
      "theoretical_loss": 3.6970374082269815,
      "tokens_seen": 873560064
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003713640922768305,
      "loss": 2.782,
      "theoretical_loss": 3.6970103507644403,
      "tokens_seen": 873625600
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037135406218655973,
      "loss": 2.5894,
      "theoretical_loss": 3.6969832958998485,
      "tokens_seen": 873691136
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037134403209628886,
      "loss": 2.7581,
      "theoretical_loss": 3.6969562436327625,
      "tokens_seen": 873756672
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003713340020060181,
      "loss": 2.5189,
      "theoretical_loss": 3.696929193962738,
      "tokens_seen": 873822208
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.0003713239719157472,
      "loss": 2.8104,
      "theoretical_loss": 3.6969021468893306,
      "tokens_seen": 873887744
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037131394182547645,
      "loss": 2.5091,
      "theoretical_loss": 3.6968751024120965,
      "tokens_seen": 873953280
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.00037130391173520564,
      "loss": 2.8142,
      "theoretical_loss": 3.696848060530592,
      "tokens_seen": 874018816
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.0003712938816449348,
      "loss": 2.9372,
      "theoretical_loss": 3.6968210212443733,
      "tokens_seen": 874084352
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.000371283851554664,
      "loss": 2.5234,
      "theoretical_loss": 3.6967939845529965,
      "tokens_seen": 874149888
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.0003712738214643932,
      "loss": 2.489,
      "theoretical_loss": 3.696766950456018,
      "tokens_seen": 874215424
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.00037126379137412236,
      "loss": 2.6471,
      "theoretical_loss": 3.6967399189529955,
      "tokens_seen": 874280960
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.0003712537612838516,
      "loss": 2.6633,
      "theoretical_loss": 3.696712890043485,
      "tokens_seen": 874346496
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.0003712437311935807,
      "loss": 2.7678,
      "theoretical_loss": 3.6966858637270423,
      "tokens_seen": 874412032
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.00037123370110330996,
      "loss": 2.727,
      "theoretical_loss": 3.6966588400032254,
      "tokens_seen": 874477568
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.0003712236710130391,
      "loss": 2.9024,
      "theoretical_loss": 3.6966318188715914,
      "tokens_seen": 874543104
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.0003712136409227683,
      "loss": 2.6938,
      "theoretical_loss": 3.696604800331697,
      "tokens_seen": 874608640
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.0003712036108324975,
      "loss": 2.7446,
      "theoretical_loss": 3.6965777843830994,
      "tokens_seen": 874674176
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.0003711935807422267,
      "loss": 2.8905,
      "theoretical_loss": 3.6965507710253562,
      "tokens_seen": 874739712
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.00037118355065195586,
      "loss": 2.575,
      "theoretical_loss": 3.696523760258025,
      "tokens_seen": 874805248
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.0003711735205616851,
      "loss": 2.6711,
      "theoretical_loss": 3.696496752080663,
      "tokens_seen": 874870784
    },
    {
      "epoch": 2.09,
      "objective/train/docs_used": 961818,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8190174102783203,
      "objective/train/theoretical_loss": 3.696483248963082,
      "objective/train/tokens_used": 890267104,
      "theoretical_loss": 3.696483248963082,
      "tokens_seen": 874903552
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.0003711634904714142,
      "loss": 2.9269,
      "theoretical_loss": 3.696469746492828,
      "tokens_seen": 874936320
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.00037115346038114346,
      "loss": 2.7968,
      "theoretical_loss": 3.6964427434940776,
      "tokens_seen": 875001856
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.0003711434302908726,
      "loss": 2.6184,
      "theoretical_loss": 3.6964157430839704,
      "tokens_seen": 875067392
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.0003711334002006018,
      "loss": 2.5877,
      "theoretical_loss": 3.6963887452620634,
      "tokens_seen": 875132928
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.000371123370110331,
      "loss": 2.4639,
      "theoretical_loss": 3.696361750027915,
      "tokens_seen": 875198464
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.0003711133400200602,
      "loss": 2.9735,
      "theoretical_loss": 3.696334757381084,
      "tokens_seen": 875264000
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.00037110330992978937,
      "loss": 2.6998,
      "theoretical_loss": 3.696307767321128,
      "tokens_seen": 875329536
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.00037109327983951855,
      "loss": 2.6655,
      "theoretical_loss": 3.696280779847606,
      "tokens_seen": 875395072
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.00037108324974924773,
      "loss": 2.713,
      "theoretical_loss": 3.696253794960076,
      "tokens_seen": 875460608
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.00037107321965897696,
      "loss": 2.5791,
      "theoretical_loss": 3.696226812658097,
      "tokens_seen": 875526144
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.0003710631895687061,
      "loss": 2.7312,
      "theoretical_loss": 3.696199832941228,
      "tokens_seen": 875591680
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.0003710531594784353,
      "loss": 2.9171,
      "theoretical_loss": 3.696172855809027,
      "tokens_seen": 875657216
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.00037104312938816445,
      "loss": 2.7087,
      "theoretical_loss": 3.696145881261054,
      "tokens_seen": 875722752
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.0003710330992978937,
      "loss": 2.726,
      "theoretical_loss": 3.6961189092968674,
      "tokens_seen": 875788288
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.00037102306920762287,
      "loss": 2.5598,
      "theoretical_loss": 3.696091939916027,
      "tokens_seen": 875853824
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.00037101303911735205,
      "loss": 2.6229,
      "theoretical_loss": 3.696064973118091,
      "tokens_seen": 875919360
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.00037100300902708123,
      "loss": 2.7326,
      "theoretical_loss": 3.6960380089026197,
      "tokens_seen": 875984896
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.00037099297893681047,
      "loss": 2.5333,
      "theoretical_loss": 3.696011047269173,
      "tokens_seen": 876050432
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.0003709829488465396,
      "loss": 2.4307,
      "theoretical_loss": 3.695984088217309,
      "tokens_seen": 876115968
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.00037097291875626883,
      "loss": 2.5423,
      "theoretical_loss": 3.695957131746589,
      "tokens_seen": 876181504
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.00037096288866599796,
      "loss": 2.8121,
      "theoretical_loss": 3.695930177856572,
      "tokens_seen": 876247040
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.0003709528585757272,
      "loss": 2.4569,
      "theoretical_loss": 3.695903226546818,
      "tokens_seen": 876312576
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.0003709428284854564,
      "loss": 2.7681,
      "theoretical_loss": 3.695876277816888,
      "tokens_seen": 876378112
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.00037093279839518555,
      "loss": 2.6246,
      "theoretical_loss": 3.6958493316663406,
      "tokens_seen": 876443648
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.00037092276830491474,
      "loss": 2.5519,
      "theoretical_loss": 3.6958223880947374,
      "tokens_seen": 876509184
    },
    {
      "epoch": 2.09,
      "objective/train/docs_used": 961818,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.962693691253662,
      "objective/train/theoretical_loss": 3.695808917275902,
      "objective/train/tokens_used": 890267104,
      "theoretical_loss": 3.695808917275902,
      "tokens_seen": 876541952
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.0003709127382146439,
      "loss": 2.7102,
      "theoretical_loss": 3.695795447101638,
      "tokens_seen": 876574720
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.0003709027081243731,
      "loss": 2.8035,
      "theoretical_loss": 3.695768508686603,
      "tokens_seen": 876640256
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.00037089267803410233,
      "loss": 2.7445,
      "theoretical_loss": 3.695741572849193,
      "tokens_seen": 876705792
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.00037088264794383146,
      "loss": 2.8925,
      "theoretical_loss": 3.695714639588969,
      "tokens_seen": 876771328
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.0003708726178535607,
      "loss": 2.7924,
      "theoretical_loss": 3.695687708905491,
      "tokens_seen": 876836864
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.00037086258776328993,
      "loss": 2.5923,
      "theoretical_loss": 3.6956607807983213,
      "tokens_seen": 876902400
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.00037085255767301906,
      "loss": 2.6051,
      "theoretical_loss": 3.69563385526702,
      "tokens_seen": 876967936
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.0003708425275827483,
      "loss": 2.7864,
      "theoretical_loss": 3.695606932311148,
      "tokens_seen": 877033472
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.0003708324974924774,
      "loss": 2.9097,
      "theoretical_loss": 3.695580011930267,
      "tokens_seen": 877099008
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.00037082246740220665,
      "loss": 2.8712,
      "theoretical_loss": 3.6955530941239387,
      "tokens_seen": 877164544
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.00037081243731193584,
      "loss": 2.7235,
      "theoretical_loss": 3.6955261788917237,
      "tokens_seen": 877230080
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.000370802407221665,
      "loss": 2.7495,
      "theoretical_loss": 3.6954992662331847,
      "tokens_seen": 877295616
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.0003707923771313942,
      "loss": 2.7919,
      "theoretical_loss": 3.695472356147882,
      "tokens_seen": 877361152
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.0003707823470411234,
      "loss": 2.5835,
      "theoretical_loss": 3.695445448635378,
      "tokens_seen": 877426688
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.00037077231695085256,
      "loss": 2.6822,
      "theoretical_loss": 3.6954185436952347,
      "tokens_seen": 877492224
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.0003707622868605818,
      "loss": 2.658,
      "theoretical_loss": 3.6953916413270145,
      "tokens_seen": 877557760
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.0003707522567703109,
      "loss": 2.4346,
      "theoretical_loss": 3.6953647415302786,
      "tokens_seen": 877623296
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.00037074222668004016,
      "loss": 2.8774,
      "theoretical_loss": 3.6953378443045892,
      "tokens_seen": 877688832
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.0003707321965897693,
      "loss": 2.6141,
      "theoretical_loss": 3.6953109496495093,
      "tokens_seen": 877754368
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.0003707221664994985,
      "loss": 2.683,
      "theoretical_loss": 3.695284057564601,
      "tokens_seen": 877819904
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.0003707121364092277,
      "loss": 2.7027,
      "theoretical_loss": 3.6952571680494275,
      "tokens_seen": 877885440
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.0003707021063189569,
      "loss": 2.6527,
      "theoretical_loss": 3.6952302811035502,
      "tokens_seen": 877950976
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.00037069207622868606,
      "loss": 2.593,
      "theoretical_loss": 3.6952033967265328,
      "tokens_seen": 878016512
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.0003706820461384153,
      "loss": 2.7754,
      "theoretical_loss": 3.6951765149179376,
      "tokens_seen": 878082048
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.0003706720160481444,
      "loss": 2.7959,
      "theoretical_loss": 3.6951496356773275,
      "tokens_seen": 878147584
    },
    {
      "epoch": 2.09,
      "objective/train/docs_used": 961818,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8952300548553467,
      "objective/train/theoretical_loss": 3.6951361970198806,
      "objective/train/tokens_used": 890267104,
      "theoretical_loss": 3.6951361970198806,
      "tokens_seen": 878180352
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.00037066198595787366,
      "loss": 2.6569,
      "theoretical_loss": 3.695122759004266,
      "tokens_seen": 878213120
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.0003706519558676028,
      "loss": 2.885,
      "theoretical_loss": 3.695095884898316,
      "tokens_seen": 878278656
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.000370641925777332,
      "loss": 2.7,
      "theoretical_loss": 3.695069013359041,
      "tokens_seen": 878344192
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.0003706318956870612,
      "loss": 2.8532,
      "theoretical_loss": 3.6950421443860044,
      "tokens_seen": 878409728
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.0003706218655967904,
      "loss": 2.7147,
      "theoretical_loss": 3.6950152779787695,
      "tokens_seen": 878475264
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.00037061183550651957,
      "loss": 2.5623,
      "theoretical_loss": 3.6949884141369003,
      "tokens_seen": 878540800
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.00037060180541624875,
      "loss": 2.584,
      "theoretical_loss": 3.6949615528599593,
      "tokens_seen": 878606336
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.00037059177532597793,
      "loss": 2.778,
      "theoretical_loss": 3.694934694147512,
      "tokens_seen": 878671872
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.00037058174523570716,
      "loss": 2.8953,
      "theoretical_loss": 3.6949078379991205,
      "tokens_seen": 878737408
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.0003705717151454363,
      "loss": 2.7651,
      "theoretical_loss": 3.6948809844143504,
      "tokens_seen": 878802944
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.0003705616850551655,
      "loss": 2.6819,
      "theoretical_loss": 3.694854133392765,
      "tokens_seen": 878868480
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.00037055165496489465,
      "loss": 2.6763,
      "theoretical_loss": 3.694827284933929,
      "tokens_seen": 878934016
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.0003705416248746239,
      "loss": 2.4986,
      "theoretical_loss": 3.6948004390374063,
      "tokens_seen": 878999552
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.00037053159478435307,
      "loss": 2.6861,
      "theoretical_loss": 3.6947735957027614,
      "tokens_seen": 879065088
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.00037052156469408225,
      "loss": 2.6115,
      "theoretical_loss": 3.694746754929559,
      "tokens_seen": 879130624
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.00037051153460381143,
      "loss": 2.5333,
      "theoretical_loss": 3.6947199167173643,
      "tokens_seen": 879196160
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.00037050150451354067,
      "loss": 2.8628,
      "theoretical_loss": 3.6946930810657412,
      "tokens_seen": 879261696
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.0003704914744232698,
      "loss": 2.8541,
      "theoretical_loss": 3.694666247974255,
      "tokens_seen": 879327232
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.00037048144433299903,
      "loss": 2.6677,
      "theoretical_loss": 3.6946394174424704,
      "tokens_seen": 879392768
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.00037047141424272816,
      "loss": 2.5643,
      "theoretical_loss": 3.694612589469953,
      "tokens_seen": 879458304
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.0003704613841524574,
      "loss": 2.6471,
      "theoretical_loss": 3.6945857640562676,
      "tokens_seen": 879523840
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.0003704513540621866,
      "loss": 2.5586,
      "theoretical_loss": 3.6945589412009796,
      "tokens_seen": 879589376
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.00037044132397191575,
      "loss": 2.6428,
      "theoretical_loss": 3.694532120903654,
      "tokens_seen": 879654912
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.00037043129388164494,
      "loss": 2.6342,
      "theoretical_loss": 3.694505303163857,
      "tokens_seen": 879720448
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.0003704212637913741,
      "loss": 2.6444,
      "theoretical_loss": 3.6944784879811543,
      "tokens_seen": 879785984
    },
    {
      "epoch": 2.09,
      "objective/train/docs_used": 961818,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.845834732055664,
      "objective/train/theoretical_loss": 3.6944650813485773,
      "objective/train/tokens_used": 890267104,
      "theoretical_loss": 3.6944650813485773,
      "tokens_seen": 879818752
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.0003704112337011033,
      "loss": 2.6323,
      "theoretical_loss": 3.694451675355111,
      "tokens_seen": 879851520
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.00037040120361083253,
      "loss": 2.6846,
      "theoretical_loss": 3.694424865285293,
      "tokens_seen": 879917056
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.00037039117352056166,
      "loss": 2.2876,
      "theoretical_loss": 3.6943980577712665,
      "tokens_seen": 879982592
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.0003703811434302909,
      "loss": 2.8253,
      "theoretical_loss": 3.694371252812597,
      "tokens_seen": 880048128
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.0003703711133400201,
      "loss": 2.8466,
      "theoretical_loss": 3.694344450408852,
      "tokens_seen": 880113664
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.00037036108324974926,
      "loss": 2.9858,
      "theoretical_loss": 3.6943176505595963,
      "tokens_seen": 880179200
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.00037035105315947844,
      "loss": 2.9263,
      "theoretical_loss": 3.6942908532643974,
      "tokens_seen": 880244736
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.0003703410230692076,
      "loss": 2.6523,
      "theoretical_loss": 3.6942640585228204,
      "tokens_seen": 880310272
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.0003703309929789368,
      "loss": 2.6214,
      "theoretical_loss": 3.694237266334433,
      "tokens_seen": 880375808
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.00037032096288866604,
      "loss": 2.4765,
      "theoretical_loss": 3.694210476698802,
      "tokens_seen": 880441344
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.00037031093279839516,
      "loss": 2.6534,
      "theoretical_loss": 3.6941836896154934,
      "tokens_seen": 880506880
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.0003703009027081244,
      "loss": 2.638,
      "theoretical_loss": 3.6941569050840743,
      "tokens_seen": 880572416
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.0003702908726178535,
      "loss": 2.7322,
      "theoretical_loss": 3.6941301231041117,
      "tokens_seen": 880637952
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.00037028084252758276,
      "loss": 2.7069,
      "theoretical_loss": 3.6941033436751733,
      "tokens_seen": 880703488
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.00037027081243731194,
      "loss": 2.6372,
      "theoretical_loss": 3.6940765667968254,
      "tokens_seen": 880769024
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.0003702607823470411,
      "loss": 2.6195,
      "theoretical_loss": 3.694049792468636,
      "tokens_seen": 880834560
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.0003702507522567703,
      "loss": 2.7658,
      "theoretical_loss": 3.6940230206901727,
      "tokens_seen": 880900096
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.0003702407221664995,
      "loss": 2.5632,
      "theoretical_loss": 3.693996251461002,
      "tokens_seen": 880965632
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.00037023069207622867,
      "loss": 2.4424,
      "theoretical_loss": 3.693969484780692,
      "tokens_seen": 881031168
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.0003702206619859579,
      "loss": 2.6033,
      "theoretical_loss": 3.693942720648811,
      "tokens_seen": 881096704
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.00037021063189568703,
      "loss": 2.6652,
      "theoretical_loss": 3.6939159590649266,
      "tokens_seen": 881162240
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.00037020060180541626,
      "loss": 2.773,
      "theoretical_loss": 3.693889200028606,
      "tokens_seen": 881227776
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.00037019057171514544,
      "loss": 2.6614,
      "theoretical_loss": 3.6938624435394183,
      "tokens_seen": 881293312
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.0003701805416248746,
      "loss": 2.4945,
      "theoretical_loss": 3.6938356895969306,
      "tokens_seen": 881358848
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.0003701705115346038,
      "loss": 2.6585,
      "theoretical_loss": 3.6938089382007124,
      "tokens_seen": 881424384
    },
    {
      "epoch": 2.09,
      "objective/train/docs_used": 961818,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6805853843688965,
      "objective/train/theoretical_loss": 3.6937955634573187,
      "objective/train/tokens_used": 890267104,
      "theoretical_loss": 3.6937955634573187,
      "tokens_seen": 881457152
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.000370160481444333,
      "loss": 2.7097,
      "theoretical_loss": 3.6937821893503306,
      "tokens_seen": 881489920
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.00037015045135406217,
      "loss": 2.8098,
      "theoretical_loss": 3.6937554430453554,
      "tokens_seen": 881555456
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.0003701404212637914,
      "loss": 2.4764,
      "theoretical_loss": 3.6937286992853537,
      "tokens_seen": 881620992
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.00037013039117352053,
      "loss": 2.767,
      "theoretical_loss": 3.6937019580698953,
      "tokens_seen": 881686528
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.00037012036108324977,
      "loss": 2.6104,
      "theoretical_loss": 3.6936752193985485,
      "tokens_seen": 881752064
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.00037011033099297895,
      "loss": 2.6776,
      "theoretical_loss": 3.6936484832708825,
      "tokens_seen": 881817600
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.00037010030090270813,
      "loss": 2.8724,
      "theoretical_loss": 3.693621749686466,
      "tokens_seen": 881883136
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.00037009027081243736,
      "loss": 2.5104,
      "theoretical_loss": 3.693595018644868,
      "tokens_seen": 881948672
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.0003700802407221665,
      "loss": 2.6103,
      "theoretical_loss": 3.6935682901456586,
      "tokens_seen": 882014208
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.0003700702106318957,
      "loss": 2.8333,
      "theoretical_loss": 3.6935415641884055,
      "tokens_seen": 882079744
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.00037006018054162485,
      "loss": 2.4745,
      "theoretical_loss": 3.6935148407726794,
      "tokens_seen": 882145280
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.0003700501504513541,
      "loss": 2.6831,
      "theoretical_loss": 3.69348811989805,
      "tokens_seen": 882210816
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.00037004012036108327,
      "loss": 2.5069,
      "theoretical_loss": 3.693461401564086,
      "tokens_seen": 882276352
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.00037003009027081245,
      "loss": 2.5407,
      "theoretical_loss": 3.6934346857703577,
      "tokens_seen": 882341888
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.00037002006018054163,
      "loss": 2.9309,
      "theoretical_loss": 3.693407972516434,
      "tokens_seen": 882407424
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.00037001003009027087,
      "loss": 2.5706,
      "theoretical_loss": 3.6933812618018864,
      "tokens_seen": 882472960
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.00037,
      "loss": 2.6596,
      "theoretical_loss": 3.6933545536262846,
      "tokens_seen": 882538496
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.00036998996990972923,
      "loss": 2.8532,
      "theoretical_loss": 3.693327847989198,
      "tokens_seen": 882604032
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.00036997993981945836,
      "loss": 2.6359,
      "theoretical_loss": 3.6933011448901967,
      "tokens_seen": 882669568
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.0003699699097291876,
      "loss": 2.7255,
      "theoretical_loss": 3.693274444328852,
      "tokens_seen": 882735104
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.0003699598796389168,
      "loss": 2.6571,
      "theoretical_loss": 3.693247746304734,
      "tokens_seen": 882800640
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.00036994984954864595,
      "loss": 2.5259,
      "theoretical_loss": 3.693221050817413,
      "tokens_seen": 882866176
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.00036993981945837514,
      "loss": 2.8501,
      "theoretical_loss": 3.69319435786646,
      "tokens_seen": 882931712
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.0003699297893681043,
      "loss": 2.466,
      "theoretical_loss": 3.693167667451445,
      "tokens_seen": 882997248
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.0003699197592778335,
      "loss": 2.8891,
      "theoretical_loss": 3.6931409795719405,
      "tokens_seen": 883062784
    },
    {
      "epoch": 2.09,
      "objective/train/docs_used": 961818,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.3581273555755615,
      "objective/train/theoretical_loss": 3.6931276365828696,
      "objective/train/tokens_used": 890267104,
      "theoretical_loss": 3.6931276365828696,
      "tokens_seen": 883095552
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.00036990972918756273,
      "loss": 2.6237,
      "theoretical_loss": 3.6931142942275157,
      "tokens_seen": 883128320
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.00036989969909729186,
      "loss": 2.4756,
      "theoretical_loss": 3.6930876114177433,
      "tokens_seen": 883193856
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.0003698896690070211,
      "loss": 2.5828,
      "theoretical_loss": 3.6930609311421936,
      "tokens_seen": 883259392
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.0003698796389167503,
      "loss": 2.7051,
      "theoretical_loss": 3.6930342534004374,
      "tokens_seen": 883324928
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.00036986960882647946,
      "loss": 2.4825,
      "theoretical_loss": 3.693007578192047,
      "tokens_seen": 883390464
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.00036985957873620864,
      "loss": 2.6811,
      "theoretical_loss": 3.6929809055165936,
      "tokens_seen": 883456000
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.0003698495486459378,
      "loss": 2.8016,
      "theoretical_loss": 3.6929542353736493,
      "tokens_seen": 883521536
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.000369839518555667,
      "loss": 2.6199,
      "theoretical_loss": 3.6929275677627853,
      "tokens_seen": 883587072
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.00036982948846539624,
      "loss": 2.6573,
      "theoretical_loss": 3.692900902683573,
      "tokens_seen": 883652608
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.00036981945837512536,
      "loss": 2.7305,
      "theoretical_loss": 3.692874240135585,
      "tokens_seen": 883718144
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.0003698094282848546,
      "loss": 2.7184,
      "theoretical_loss": 3.6928475801183933,
      "tokens_seen": 883783680
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.0003697993981945837,
      "loss": 2.5299,
      "theoretical_loss": 3.6928209226315696,
      "tokens_seen": 883849216
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.00036978936810431296,
      "loss": 2.5705,
      "theoretical_loss": 3.692794267674687,
      "tokens_seen": 883914752
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.00036977933801404214,
      "loss": 2.6542,
      "theoretical_loss": 3.692767615247317,
      "tokens_seen": 883980288
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.0003697693079237713,
      "loss": 2.7557,
      "theoretical_loss": 3.6927409653490324,
      "tokens_seen": 884045824
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.0003697592778335005,
      "loss": 2.5676,
      "theoretical_loss": 3.6927143179794055,
      "tokens_seen": 884111360
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.0003697492477432297,
      "loss": 2.7962,
      "theoretical_loss": 3.692687673138009,
      "tokens_seen": 884176896
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.00036973921765295887,
      "loss": 2.475,
      "theoretical_loss": 3.692661030824416,
      "tokens_seen": 884242432
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.0003697291875626881,
      "loss": 2.6934,
      "theoretical_loss": 3.6926343910381996,
      "tokens_seen": 884307968
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.00036971915747241723,
      "loss": 2.5649,
      "theoretical_loss": 3.692607753778932,
      "tokens_seen": 884373504
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.00036970912738214646,
      "loss": 2.825,
      "theoretical_loss": 3.692581119046186,
      "tokens_seen": 884439040
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.00036969909729187565,
      "loss": 2.7717,
      "theoretical_loss": 3.692554486839536,
      "tokens_seen": 884504576
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.0003696890672016048,
      "loss": 2.6135,
      "theoretical_loss": 3.692527857158555,
      "tokens_seen": 884570112
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.000369679037111334,
      "loss": 2.6465,
      "theoretical_loss": 3.692501230002815,
      "tokens_seen": 884635648
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.0003696690070210632,
      "loss": 2.7335,
      "theoretical_loss": 3.6924746053718915,
      "tokens_seen": 884701184
    },
    {
      "epoch": 2.09,
      "objective/train/docs_used": 961818,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.489288568496704,
      "objective/train/theoretical_loss": 3.692461294003102,
      "objective/train/tokens_used": 890267104,
      "theoretical_loss": 3.692461294003102,
      "tokens_seen": 884733952
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.00036965897693079237,
      "loss": 2.7261,
      "theoretical_loss": 3.6924479832653567,
      "tokens_seen": 884766720
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.0003696489468405216,
      "loss": 2.7215,
      "theoretical_loss": 3.6924213636827847,
      "tokens_seen": 884832256
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.00036963891675025073,
      "loss": 2.4819,
      "theoretical_loss": 3.6923947466237497,
      "tokens_seen": 884897792
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.00036962888665997997,
      "loss": 2.7193,
      "theoretical_loss": 3.6923681320878243,
      "tokens_seen": 884963328
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.0003696188565697091,
      "loss": 2.7264,
      "theoretical_loss": 3.692341520074584,
      "tokens_seen": 885028864
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.00036960882647943833,
      "loss": 2.6642,
      "theoretical_loss": 3.692314910583603,
      "tokens_seen": 885094400
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.0003695987963891675,
      "loss": 2.7689,
      "theoretical_loss": 3.692288303614454,
      "tokens_seen": 885159936
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.0003695887662988967,
      "loss": 2.7526,
      "theoretical_loss": 3.692261699166712,
      "tokens_seen": 885225472
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.00036957873620862587,
      "loss": 2.6084,
      "theoretical_loss": 3.692235097239952,
      "tokens_seen": 885291008
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.00036956870611835505,
      "loss": 2.6363,
      "theoretical_loss": 3.6922084978337475,
      "tokens_seen": 885356544
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.00036955867602808424,
      "loss": 2.6097,
      "theoretical_loss": 3.6921819009476744,
      "tokens_seen": 885422080
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.00036954864593781347,
      "loss": 2.4996,
      "theoretical_loss": 3.6921553065813066,
      "tokens_seen": 885487616
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.0003695386158475426,
      "loss": 2.6923,
      "theoretical_loss": 3.6921287147342188,
      "tokens_seen": 885553152
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.00036952858575727183,
      "loss": 2.6882,
      "theoretical_loss": 3.6921021254059863,
      "tokens_seen": 885618688
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.000369518555667001,
      "loss": 2.5709,
      "theoretical_loss": 3.6920755385961845,
      "tokens_seen": 885684224
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.0003695085255767302,
      "loss": 2.6799,
      "theoretical_loss": 3.6920489543043873,
      "tokens_seen": 885749760
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.0003694984954864594,
      "loss": 2.7231,
      "theoretical_loss": 3.692022372530171,
      "tokens_seen": 885815296
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.00036948846539618856,
      "loss": 2.5862,
      "theoretical_loss": 3.691995793273111,
      "tokens_seen": 885880832
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.00036947843530591774,
      "loss": 2.7793,
      "theoretical_loss": 3.691969216532782,
      "tokens_seen": 885946368
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.000369468405215647,
      "loss": 2.566,
      "theoretical_loss": 3.6919426423087605,
      "tokens_seen": 886011904
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.0003694583751253761,
      "loss": 2.577,
      "theoretical_loss": 3.6919160706006213,
      "tokens_seen": 886077440
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.00036944834503510534,
      "loss": 2.7,
      "theoretical_loss": 3.6918895014079407,
      "tokens_seen": 886142976
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.00036943831494483446,
      "loss": 2.5987,
      "theoretical_loss": 3.691862934730294,
      "tokens_seen": 886208512
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.0003694282848545637,
      "loss": 2.4723,
      "theoretical_loss": 3.691836370567257,
      "tokens_seen": 886274048
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.0003694182547642929,
      "loss": 2.4132,
      "theoretical_loss": 3.6918098089184075,
      "tokens_seen": 886339584
    },
    {
      "epoch": 2.09,
      "objective/train/docs_used": 961818,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.893305540084839,
      "objective/train/theoretical_loss": 3.6917965290366697,
      "objective/train/tokens_used": 890267104,
      "theoretical_loss": 3.6917965290366697,
      "tokens_seen": 886372352
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.00036940822467402206,
      "loss": 2.7739,
      "theoretical_loss": 3.6917832497833203,
      "tokens_seen": 886405120
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.00036939819458375124,
      "loss": 2.7017,
      "theoretical_loss": 3.691756693161571,
      "tokens_seen": 886470656
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.0003693881644934805,
      "loss": 2.8098,
      "theoretical_loss": 3.6917301390527375,
      "tokens_seen": 886536192
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.0003693781344032096,
      "loss": 2.6014,
      "theoretical_loss": 3.691703587456395,
      "tokens_seen": 886601728
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.00036936810431293884,
      "loss": 2.7644,
      "theoretical_loss": 3.691677038372121,
      "tokens_seen": 886667264
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.000369358074222668,
      "loss": 2.6346,
      "theoretical_loss": 3.691650491799492,
      "tokens_seen": 886732800
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.0003693480441323972,
      "loss": 2.534,
      "theoretical_loss": 3.6916239477380843,
      "tokens_seen": 886798336
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.00036933801404212644,
      "loss": 2.6581,
      "theoretical_loss": 3.6915974061874754,
      "tokens_seen": 886863872
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.00036932798395185556,
      "loss": 2.7481,
      "theoretical_loss": 3.6915708671472416,
      "tokens_seen": 886929408
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.0003693179538615848,
      "loss": 2.6258,
      "theoretical_loss": 3.691544330616961,
      "tokens_seen": 886994944
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.0003693079237713139,
      "loss": 2.6422,
      "theoretical_loss": 3.6915177965962096,
      "tokens_seen": 887060480
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.00036929789368104316,
      "loss": 2.7657,
      "theoretical_loss": 3.691491265084566,
      "tokens_seen": 887126016
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.00036928786359077234,
      "loss": 2.5757,
      "theoretical_loss": 3.691464736081606,
      "tokens_seen": 887191552
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.0003692778335005015,
      "loss": 2.7411,
      "theoretical_loss": 3.6914382095869085,
      "tokens_seen": 887257088
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.0003692678034102307,
      "loss": 2.9059,
      "theoretical_loss": 3.691411685600051,
      "tokens_seen": 887322624
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.0003692577733199599,
      "loss": 2.767,
      "theoretical_loss": 3.691385164120611,
      "tokens_seen": 887388160
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.00036924774322968907,
      "loss": 2.6977,
      "theoretical_loss": 3.6913586451481653,
      "tokens_seen": 887453696
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.0003692377131394183,
      "loss": 2.6306,
      "theoretical_loss": 3.6913321286822924,
      "tokens_seen": 887519232
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.00036922768304914743,
      "loss": 2.7088,
      "theoretical_loss": 3.6913056147225713,
      "tokens_seen": 887584768
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.00036921765295887666,
      "loss": 2.4745,
      "theoretical_loss": 3.691279103268579,
      "tokens_seen": 887650304
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.00036920762286860585,
      "loss": 2.7917,
      "theoretical_loss": 3.691252594319894,
      "tokens_seen": 887715840
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.000369197592778335,
      "loss": 2.7978,
      "theoretical_loss": 3.691226087876095,
      "tokens_seen": 887781376
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.0003691875626880642,
      "loss": 2.7152,
      "theoretical_loss": 3.6911995839367604,
      "tokens_seen": 887846912
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.0003691775325977934,
      "loss": 2.6977,
      "theoretical_loss": 3.691173082501468,
      "tokens_seen": 887912448
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.00036916750250752257,
      "loss": 2.769,
      "theoretical_loss": 3.6911465835697976,
      "tokens_seen": 887977984
    },
    {
      "epoch": 2.09,
      "objective/train/docs_used": 961818,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6037163734436035,
      "objective/train/theoretical_loss": 3.691133335042688,
      "objective/train/tokens_used": 890267104,
      "theoretical_loss": 3.691133335042688,
      "tokens_seen": 888010752
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.0003691574724172518,
      "loss": 2.5662,
      "theoretical_loss": 3.6911200871413263,
      "tokens_seen": 888043520
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.00036914744232698093,
      "loss": 2.7313,
      "theoretical_loss": 3.6910935932156343,
      "tokens_seen": 888109056
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.00036913741223671017,
      "loss": 2.8887,
      "theoretical_loss": 3.6910671017923002,
      "tokens_seen": 888174592
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.0003691273821464393,
      "loss": 2.7674,
      "theoretical_loss": 3.6910406128709026,
      "tokens_seen": 888240128
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.00036911735205616853,
      "loss": 2.5497,
      "theoretical_loss": 3.6910141264510212,
      "tokens_seen": 888305664
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.0003691073219658977,
      "loss": 2.6463,
      "theoretical_loss": 3.690987642532235,
      "tokens_seen": 888371200
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.0003690972918756269,
      "loss": 2.8052,
      "theoretical_loss": 3.690961161114123,
      "tokens_seen": 888436736
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.0003690872617853561,
      "loss": 2.906,
      "theoretical_loss": 3.690934682196265,
      "tokens_seen": 888502272
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.00036907723169508525,
      "loss": 2.7999,
      "theoretical_loss": 3.690908205778241,
      "tokens_seen": 888567808
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.00036906720160481444,
      "loss": 2.7024,
      "theoretical_loss": 3.6908817318596303,
      "tokens_seen": 888633344
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.00036905717151454367,
      "loss": 2.6924,
      "theoretical_loss": 3.6908552604400118,
      "tokens_seen": 888698880
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.0003690471414242728,
      "loss": 2.6747,
      "theoretical_loss": 3.6908287915189666,
      "tokens_seen": 888764416
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.00036903711133400203,
      "loss": 2.6567,
      "theoretical_loss": 3.690802325096074,
      "tokens_seen": 888829952
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.0003690270812437312,
      "loss": 2.4936,
      "theoretical_loss": 3.6907758611709145,
      "tokens_seen": 888895488
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.0003690170511534604,
      "loss": 2.4739,
      "theoretical_loss": 3.6907493997430674,
      "tokens_seen": 888961024
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.0003690070210631896,
      "loss": 2.5213,
      "theoretical_loss": 3.6907229408121136,
      "tokens_seen": 889026560
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.00036899699097291876,
      "loss": 2.549,
      "theoretical_loss": 3.690696484377634,
      "tokens_seen": 889092096
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.00036898696088264794,
      "loss": 2.7987,
      "theoretical_loss": 3.6906700304392075,
      "tokens_seen": 889157632
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.0003689769307923772,
      "loss": 2.4143,
      "theoretical_loss": 3.690643578996416,
      "tokens_seen": 889223168
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.0003689669007021063,
      "loss": 2.5284,
      "theoretical_loss": 3.6906171300488397,
      "tokens_seen": 889288704
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.00036895687061183554,
      "loss": 2.5884,
      "theoretical_loss": 3.690590683596059,
      "tokens_seen": 889354240
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.00036894684052156466,
      "loss": 2.5301,
      "theoretical_loss": 3.6905642396376557,
      "tokens_seen": 889419776
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.0003689368104312939,
      "loss": 2.7106,
      "theoretical_loss": 3.69053779817321,
      "tokens_seen": 889485312
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.0003689267803410231,
      "loss": 2.6738,
      "theoretical_loss": 3.6905113592023033,
      "tokens_seen": 889550848
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.00036891675025075226,
      "loss": 2.9077,
      "theoretical_loss": 3.690484922724517,
      "tokens_seen": 889616384
    },
    {
      "epoch": 2.09,
      "objective/train/docs_used": 961818,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.595710515975952,
      "objective/train/theoretical_loss": 3.6904717054204124,
      "objective/train/tokens_used": 890267104,
      "theoretical_loss": 3.6904717054204124,
      "tokens_seen": 889649152
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.00036890672016048144,
      "loss": 2.6372,
      "theoretical_loss": 3.690458488739431,
      "tokens_seen": 889681920
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.0003688966900702107,
      "loss": 2.6164,
      "theoretical_loss": 3.6904320572466283,
      "tokens_seen": 889747456
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.0003688866599799398,
      "loss": 2.6341,
      "theoretical_loss": 3.6904056282456903,
      "tokens_seen": 889812992
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.00036887662988966904,
      "loss": 2.7133,
      "theoretical_loss": 3.690379201736197,
      "tokens_seen": 889878528
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.00036886659979939817,
      "loss": 2.7007,
      "theoretical_loss": 3.690352777717732,
      "tokens_seen": 889944064
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.0003688565697091274,
      "loss": 2.6807,
      "theoretical_loss": 3.6903263561898756,
      "tokens_seen": 890009600
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.0003688465396188566,
      "loss": 2.5019,
      "theoretical_loss": 3.6902999371522105,
      "tokens_seen": 890075136
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.00036883650952858576,
      "loss": 2.8941,
      "theoretical_loss": 3.6902735206043182,
      "tokens_seen": 890140672
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.00036882647943831494,
      "loss": 2.4319,
      "theoretical_loss": 3.6902471065457814,
      "tokens_seen": 890206208
    },
    {
      "epoch": 2.09,
      "learning_rate": 0.0003688164493480441,
      "loss": 2.775,
      "theoretical_loss": 3.6902206949761815,
      "tokens_seen": 890271744
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.0003688064192577733,
      "loss": 3.5933,
      "theoretical_loss": 3.690188096626651,
      "tokens_seen": 890352640
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.00036879638916750254,
      "loss": 2.7673,
      "theoretical_loss": 3.690161690616762,
      "tokens_seen": 890418176
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.00036878635907723167,
      "loss": 2.9802,
      "theoretical_loss": 3.690135287094459,
      "tokens_seen": 890483712
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.0003687763289869609,
      "loss": 2.9272,
      "theoretical_loss": 3.690108886059325,
      "tokens_seen": 890549248
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.00036876629889669003,
      "loss": 2.8218,
      "theoretical_loss": 3.690082487510943,
      "tokens_seen": 890614784
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.00036875626880641927,
      "loss": 2.9439,
      "theoretical_loss": 3.690056091448896,
      "tokens_seen": 890680320
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.00036874623871614845,
      "loss": 2.8595,
      "theoretical_loss": 3.690029697872766,
      "tokens_seen": 890745856
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.00036873620862587763,
      "loss": 2.8154,
      "theoretical_loss": 3.6900033067821374,
      "tokens_seen": 890811392
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.0003687261785356068,
      "loss": 2.8048,
      "theoretical_loss": 3.6899769181765922,
      "tokens_seen": 890876928
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.00036871614844533605,
      "loss": 2.8396,
      "theoretical_loss": 3.6899505320557138,
      "tokens_seen": 890942464
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.00036870611835506517,
      "loss": 2.7181,
      "theoretical_loss": 3.689924148419086,
      "tokens_seen": 891008000
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.0003686960882647944,
      "loss": 2.9879,
      "theoretical_loss": 3.6898977672662916,
      "tokens_seen": 891073536
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.00036868605817452353,
      "loss": 2.7716,
      "theoretical_loss": 3.6898713885969148,
      "tokens_seen": 891139072
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.00036867602808425277,
      "loss": 2.7015,
      "theoretical_loss": 3.6898450124105384,
      "tokens_seen": 891204608
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.00036866599799398195,
      "loss": 2.7423,
      "theoretical_loss": 3.689818638706747,
      "tokens_seen": 891270144
    },
    {
      "epoch": 3.0,
      "objective/train/docs_used": 1026627,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.057016134262085,
      "objective/train/theoretical_loss": 3.6898120456686536,
      "objective/train/tokens_used": 911746528,
      "theoretical_loss": 3.6898120456686536,
      "tokens_seen": 891286528
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.00036865596790371113,
      "loss": 2.8914,
      "theoretical_loss": 3.689792267485123,
      "tokens_seen": 891335680
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.0003686459378134403,
      "loss": 2.7807,
      "theoretical_loss": 3.689765898745252,
      "tokens_seen": 891401216
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.0003686359077231695,
      "loss": 2.8683,
      "theoretical_loss": 3.689739532486717,
      "tokens_seen": 891466752
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.0003686258776328987,
      "loss": 2.6409,
      "theoretical_loss": 3.689713168709102,
      "tokens_seen": 891532288
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.0003686158475426279,
      "loss": 2.8472,
      "theoretical_loss": 3.6896868074119924,
      "tokens_seen": 891597824
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.0003686058174523571,
      "loss": 2.619,
      "theoretical_loss": 3.689660448594971,
      "tokens_seen": 891663360
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.0003685957873620863,
      "loss": 2.788,
      "theoretical_loss": 3.689634092257623,
      "tokens_seen": 891728896
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.00036858575727181545,
      "loss": 2.9168,
      "theoretical_loss": 3.689607738399533,
      "tokens_seen": 891794432
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.00036857572718154464,
      "loss": 2.8188,
      "theoretical_loss": 3.6895813870202856,
      "tokens_seen": 891859968
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.00036856569709127387,
      "loss": 2.744,
      "theoretical_loss": 3.6895550381194653,
      "tokens_seen": 891925504
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.000368555667001003,
      "loss": 2.8304,
      "theoretical_loss": 3.6895286916966565,
      "tokens_seen": 891991040
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.00036854563691073223,
      "loss": 2.7051,
      "theoretical_loss": 3.689502347751445,
      "tokens_seen": 892056576
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.0003685356068204614,
      "loss": 2.781,
      "theoretical_loss": 3.6894760062834155,
      "tokens_seen": 892122112
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.0003685255767301906,
      "loss": 2.7491,
      "theoretical_loss": 3.689449667292153,
      "tokens_seen": 892187648
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.0003685155466399198,
      "loss": 2.8092,
      "theoretical_loss": 3.689423330777242,
      "tokens_seen": 892253184
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.00036850551654964896,
      "loss": 2.8625,
      "theoretical_loss": 3.689396996738269,
      "tokens_seen": 892318720
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.00036849548645937814,
      "loss": 2.672,
      "theoretical_loss": 3.6893706651748195,
      "tokens_seen": 892384256
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.0003684854563691074,
      "loss": 2.8162,
      "theoretical_loss": 3.6893443360864775,
      "tokens_seen": 892449792
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.0003684754262788365,
      "loss": 2.8069,
      "theoretical_loss": 3.68931800947283,
      "tokens_seen": 892515328
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.00036846539618856574,
      "loss": 2.78,
      "theoretical_loss": 3.6892916853334627,
      "tokens_seen": 892580864
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.00036845536609829486,
      "loss": 2.7692,
      "theoretical_loss": 3.6892653636679604,
      "tokens_seen": 892646400
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.0003684453360080241,
      "loss": 2.6777,
      "theoretical_loss": 3.68923904447591,
      "tokens_seen": 892711936
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.0003684353059177533,
      "loss": 2.7668,
      "theoretical_loss": 3.689212727756897,
      "tokens_seen": 892777472
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.00036842527582748246,
      "loss": 2.8518,
      "theoretical_loss": 3.689186413510508,
      "tokens_seen": 892843008
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.00036841524573721164,
      "loss": 2.8783,
      "theoretical_loss": 3.689160101736328,
      "tokens_seen": 892908544
    },
    {
      "epoch": 3.0,
      "objective/train/docs_used": 1031633,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.0238943099975586,
      "objective/train/theoretical_loss": 3.6891535241790177,
      "objective/train/tokens_used": 913384928,
      "theoretical_loss": 3.6891535241790177,
      "tokens_seen": 892924928
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.0003684052156469409,
      "loss": 2.858,
      "theoretical_loss": 3.689133792433945,
      "tokens_seen": 892974080
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.00036839518555667,
      "loss": 2.9869,
      "theoretical_loss": 3.689107485602944,
      "tokens_seen": 893039616
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.00036838515546639924,
      "loss": 2.8007,
      "theoretical_loss": 3.689081181242912,
      "tokens_seen": 893105152
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.00036837512537612837,
      "loss": 2.9061,
      "theoretical_loss": 3.689054879353437,
      "tokens_seen": 893170688
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.0003683650952858576,
      "loss": 2.8327,
      "theoretical_loss": 3.689028579934103,
      "tokens_seen": 893236224
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.0003683550651955868,
      "loss": 2.8884,
      "theoretical_loss": 3.6890022829844984,
      "tokens_seen": 893301760
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.00036834503510531596,
      "loss": 2.538,
      "theoretical_loss": 3.6889759885042106,
      "tokens_seen": 893367296
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.00036833500501504514,
      "loss": 2.8365,
      "theoretical_loss": 3.6889496964928252,
      "tokens_seen": 893432832
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.0003683249749247743,
      "loss": 2.7163,
      "theoretical_loss": 3.6889234069499306,
      "tokens_seen": 893498368
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.0003683149448345035,
      "loss": 2.7978,
      "theoretical_loss": 3.6888971198751133,
      "tokens_seen": 893563904
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.00036830491474423274,
      "loss": 2.7502,
      "theoretical_loss": 3.688870835267961,
      "tokens_seen": 893629440
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.00036829488465396187,
      "loss": 2.7726,
      "theoretical_loss": 3.6888445531280603,
      "tokens_seen": 893694976
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.0003682848545636911,
      "loss": 2.9963,
      "theoretical_loss": 3.688818273455,
      "tokens_seen": 893760512
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.00036827482447342023,
      "loss": 2.7254,
      "theoretical_loss": 3.688791996248366,
      "tokens_seen": 893826048
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.00036826479438314947,
      "loss": 2.8003,
      "theoretical_loss": 3.688765721507748,
      "tokens_seen": 893891584
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.00036825476429287865,
      "loss": 2.8246,
      "theoretical_loss": 3.6887394492327323,
      "tokens_seen": 893957120
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.00036824473420260783,
      "loss": 2.5958,
      "theoretical_loss": 3.6887131794229076,
      "tokens_seen": 894022656
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.000368234704112337,
      "loss": 2.6973,
      "theoretical_loss": 3.688686912077861,
      "tokens_seen": 894088192
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.00036822467402206625,
      "loss": 2.7454,
      "theoretical_loss": 3.6886606471971817,
      "tokens_seen": 894153728
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.00036821464393179537,
      "loss": 2.6587,
      "theoretical_loss": 3.688634384780457,
      "tokens_seen": 894219264
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.0003682046138415246,
      "loss": 2.6925,
      "theoretical_loss": 3.688608124827276,
      "tokens_seen": 894284800
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.00036819458375125373,
      "loss": 2.8973,
      "theoretical_loss": 3.688581867337226,
      "tokens_seen": 894350336
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.00036818455366098297,
      "loss": 2.6081,
      "theoretical_loss": 3.688555612309897,
      "tokens_seen": 894415872
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.00036817452357071215,
      "loss": 2.8145,
      "theoretical_loss": 3.6885293597448765,
      "tokens_seen": 894481408
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.00036816449348044133,
      "loss": 2.7993,
      "theoretical_loss": 3.6885031096417533,
      "tokens_seen": 894546944
    },
    {
      "epoch": 3.0,
      "objective/train/docs_used": 1036652,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8129405975341797,
      "objective/train/theoretical_loss": 3.6884965475005957,
      "objective/train/tokens_used": 915023328,
      "theoretical_loss": 3.6884965475005957,
      "tokens_seen": 894563328
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.0003681544633901705,
      "loss": 2.7861,
      "theoretical_loss": 3.6884768620001163,
      "tokens_seen": 894612480
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.0003681444332998997,
      "loss": 2.7175,
      "theoretical_loss": 3.688450616819555,
      "tokens_seen": 894678016
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.0003681344032096289,
      "loss": 2.9492,
      "theoretical_loss": 3.6884243740996574,
      "tokens_seen": 894743552
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.0003681243731193581,
      "loss": 2.6274,
      "theoretical_loss": 3.688398133840013,
      "tokens_seen": 894809088
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.00036811434302908724,
      "loss": 2.7315,
      "theoretical_loss": 3.688371896040211,
      "tokens_seen": 894874624
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.0003681043129388165,
      "loss": 2.9269,
      "theoretical_loss": 3.6883456606998406,
      "tokens_seen": 894940160
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.0003680942828485456,
      "loss": 2.8358,
      "theoretical_loss": 3.6883194278184916,
      "tokens_seen": 895005696
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.00036808425275827484,
      "loss": 2.8829,
      "theoretical_loss": 3.688293197395753,
      "tokens_seen": 895071232
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.000368074222668004,
      "loss": 2.7785,
      "theoretical_loss": 3.6882669694312145,
      "tokens_seen": 895136768
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.0003680641925777332,
      "loss": 2.9305,
      "theoretical_loss": 3.688240743924466,
      "tokens_seen": 895202304
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.0003680541624874624,
      "loss": 2.7584,
      "theoretical_loss": 3.688214520875097,
      "tokens_seen": 895267840
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.0003680441323971916,
      "loss": 2.859,
      "theoretical_loss": 3.6881883002826976,
      "tokens_seen": 895333376
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.0003680441323971916,
      "loss": 2.8224,
      "theoretical_loss": 3.6881620821468575,
      "tokens_seen": 895398912
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.00036803410230692074,
      "loss": 2.8127,
      "theoretical_loss": 3.6881358664671673,
      "tokens_seen": 895464448
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.00036802407221665,
      "loss": 2.8278,
      "theoretical_loss": 3.6881096532432163,
      "tokens_seen": 895529984
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.0003680140421263791,
      "loss": 2.8223,
      "theoretical_loss": 3.6880834424745954,
      "tokens_seen": 895595520
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.00036800401203610834,
      "loss": 2.8507,
      "theoretical_loss": 3.688057234160895,
      "tokens_seen": 895661056
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.0003679939819458375,
      "loss": 2.9332,
      "theoretical_loss": 3.6880310283017055,
      "tokens_seen": 895726592
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.0003679839518555667,
      "loss": 2.8083,
      "theoretical_loss": 3.688004824896617,
      "tokens_seen": 895792128
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.0003679739217652959,
      "loss": 2.887,
      "theoretical_loss": 3.687978623945221,
      "tokens_seen": 895857664
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.00036796389167502506,
      "loss": 2.8633,
      "theoretical_loss": 3.687952425447108,
      "tokens_seen": 895923200
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.00036795386158475424,
      "loss": 2.7929,
      "theoretical_loss": 3.687926229401868,
      "tokens_seen": 895988736
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.0003679438314944835,
      "loss": 2.8538,
      "theoretical_loss": 3.687900035809093,
      "tokens_seen": 896054272
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.0003679338014042126,
      "loss": 2.8214,
      "theoretical_loss": 3.687873844668373,
      "tokens_seen": 896119808
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.00036792377131394184,
      "loss": 2.7678,
      "theoretical_loss": 3.6878476559793008,
      "tokens_seen": 896185344
    },
    {
      "epoch": 3.0,
      "objective/train/docs_used": 1041586,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.545072555541992,
      "objective/train/theoretical_loss": 3.687841109190055,
      "objective/train/tokens_used": 916661728,
      "theoretical_loss": 3.687841109190055,
      "tokens_seen": 896201728
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.00036791374122367097,
      "loss": 2.575,
      "theoretical_loss": 3.6878214697414666,
      "tokens_seen": 896250880
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.0003679037111334002,
      "loss": 2.7571,
      "theoretical_loss": 3.6877952859544614,
      "tokens_seen": 896316416
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.0003678936810431294,
      "loss": 2.8336,
      "theoretical_loss": 3.6877691046178773,
      "tokens_seen": 896381952
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.00036788365095285857,
      "loss": 2.8234,
      "theoretical_loss": 3.687742925731306,
      "tokens_seen": 896447488
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.00036787362086258775,
      "loss": 2.7132,
      "theoretical_loss": 3.687716749294338,
      "tokens_seen": 896513024
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.000367863590772317,
      "loss": 2.751,
      "theoretical_loss": 3.6876905753065667,
      "tokens_seen": 896578560
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.00036785356068204616,
      "loss": 2.8567,
      "theoretical_loss": 3.6876644037675823,
      "tokens_seen": 896644096
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.00036784353059177535,
      "loss": 2.8155,
      "theoretical_loss": 3.687638234676978,
      "tokens_seen": 896709632
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.0003678335005015045,
      "loss": 2.9743,
      "theoretical_loss": 3.687612068034345,
      "tokens_seen": 896775168
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.0003678234704112337,
      "loss": 2.7782,
      "theoretical_loss": 3.6875859038392758,
      "tokens_seen": 896840704
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.00036781344032096294,
      "loss": 2.618,
      "theoretical_loss": 3.687559742091363,
      "tokens_seen": 896906240
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.00036780341023069207,
      "loss": 2.8275,
      "theoretical_loss": 3.6875335827901985,
      "tokens_seen": 896971776
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.0003677933801404213,
      "loss": 2.7824,
      "theoretical_loss": 3.687507425935374,
      "tokens_seen": 897037312
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.0003677933801404213,
      "loss": 2.8379,
      "theoretical_loss": 3.687481271526484,
      "tokens_seen": 897102848
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.00036778335005015043,
      "loss": 2.8418,
      "theoretical_loss": 3.6874551195631193,
      "tokens_seen": 897168384
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.00036777331995987967,
      "loss": 2.8881,
      "theoretical_loss": 3.687428970044873,
      "tokens_seen": 897233920
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.00036776328986960885,
      "loss": 2.8368,
      "theoretical_loss": 3.6874028229713387,
      "tokens_seen": 897299456
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.00036775325977933803,
      "loss": 2.7317,
      "theoretical_loss": 3.6873766783421082,
      "tokens_seen": 897364992
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.0003677432296890672,
      "loss": 2.9576,
      "theoretical_loss": 3.6873505361567753,
      "tokens_seen": 897430528
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.00036773319959879645,
      "loss": 2.8593,
      "theoretical_loss": 3.687324396414933,
      "tokens_seen": 897496064
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.00036772316950852557,
      "loss": 2.8501,
      "theoretical_loss": 3.6872982591161745,
      "tokens_seen": 897561600
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.0003677131394182548,
      "loss": 2.9521,
      "theoretical_loss": 3.6872721242600925,
      "tokens_seen": 897627136
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.00036770310932798393,
      "loss": 2.6791,
      "theoretical_loss": 3.6872459918462814,
      "tokens_seen": 897692672
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.00036769307923771317,
      "loss": 2.6738,
      "theoretical_loss": 3.6872198618743344,
      "tokens_seen": 897758208
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.00036768304914744235,
      "loss": 3.0484,
      "theoretical_loss": 3.687193734343844,
      "tokens_seen": 897823744
    },
    {
      "epoch": 3.0,
      "objective/train/docs_used": 1046594,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.713426113128662,
      "objective/train/theoretical_loss": 3.687187202842652,
      "objective/train/tokens_used": 918300128,
      "theoretical_loss": 3.687187202842652,
      "tokens_seen": 897840128
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.00036767301905717153,
      "loss": 2.8236,
      "theoretical_loss": 3.6871676092544057,
      "tokens_seen": 897889280
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.0003676629889669007,
      "loss": 2.856,
      "theoretical_loss": 3.6871414866056123,
      "tokens_seen": 897954816
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.0003676529588766299,
      "loss": 2.9411,
      "theoretical_loss": 3.687115366397057,
      "tokens_seen": 898020352
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.0003676429287863591,
      "loss": 2.622,
      "theoretical_loss": 3.6870892486283355,
      "tokens_seen": 898085888
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.0003676328986960883,
      "loss": 2.8903,
      "theoretical_loss": 3.6870631332990405,
      "tokens_seen": 898151424
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.00036762286860581744,
      "loss": 2.7094,
      "theoretical_loss": 3.6870370204087664,
      "tokens_seen": 898216960
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.0003676128385155467,
      "loss": 2.5918,
      "theoretical_loss": 3.6870109099571082,
      "tokens_seen": 898282496
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.0003676028084252758,
      "loss": 3.0356,
      "theoretical_loss": 3.6869848019436597,
      "tokens_seen": 898348032
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.00036759277833500504,
      "loss": 2.8615,
      "theoretical_loss": 3.6869586963680154,
      "tokens_seen": 898413568
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.0003675827482447342,
      "loss": 2.7986,
      "theoretical_loss": 3.68693259322977,
      "tokens_seen": 898479104
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.0003675727181544634,
      "loss": 2.7854,
      "theoretical_loss": 3.686906492528518,
      "tokens_seen": 898544640
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.0003675626880641926,
      "loss": 2.7457,
      "theoretical_loss": 3.686880394263854,
      "tokens_seen": 898610176
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.0003675526579739218,
      "loss": 2.731,
      "theoretical_loss": 3.6868542984353736,
      "tokens_seen": 898675712
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.00036754262788365094,
      "loss": 2.8772,
      "theoretical_loss": 3.686828205042671,
      "tokens_seen": 898741248
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.0003675325977933802,
      "loss": 2.8269,
      "theoretical_loss": 3.6868021140853413,
      "tokens_seen": 898806784
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.0003675225677031093,
      "loss": 2.8921,
      "theoretical_loss": 3.6867760255629802,
      "tokens_seen": 898872320
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.00036751253761283854,
      "loss": 2.7749,
      "theoretical_loss": 3.686749939475183,
      "tokens_seen": 898937856
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.0003675025075225677,
      "loss": 2.6101,
      "theoretical_loss": 3.686723855821544,
      "tokens_seen": 899003392
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.0003674924774322969,
      "loss": 2.8056,
      "theoretical_loss": 3.68669777460166,
      "tokens_seen": 899068928
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.0003674824473420261,
      "loss": 2.8025,
      "theoretical_loss": 3.686671695815125,
      "tokens_seen": 899134464
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.00036747241725175526,
      "loss": 2.7852,
      "theoretical_loss": 3.686645619461536,
      "tokens_seen": 899200000
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.00036746238716148444,
      "loss": 2.5889,
      "theoretical_loss": 3.6866195455404878,
      "tokens_seen": 899265536
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.0003674523570712137,
      "loss": 2.9105,
      "theoretical_loss": 3.686593474051577,
      "tokens_seen": 899331072
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.0003674423269809428,
      "loss": 2.5764,
      "theoretical_loss": 3.686567404994399,
      "tokens_seen": 899396608
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.00036743229689067204,
      "loss": 2.6419,
      "theoretical_loss": 3.68654133836855,
      "tokens_seen": 899462144
    },
    {
      "epoch": 3.0,
      "objective/train/docs_used": 1051693,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.662870407104492,
      "objective/train/theoretical_loss": 3.686534822091936,
      "objective/train/tokens_used": 919938528,
      "theoretical_loss": 3.686534822091936,
      "tokens_seen": 899478528
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.00036742226680040117,
      "loss": 2.8781,
      "theoretical_loss": 3.686515274173626,
      "tokens_seen": 899527680
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.0003674122367101304,
      "loss": 2.8185,
      "theoretical_loss": 3.6864892124092234,
      "tokens_seen": 899593216
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.0003674022066198596,
      "loss": 2.8196,
      "theoretical_loss": 3.6864631530749388,
      "tokens_seen": 899658752
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.00036739217652958877,
      "loss": 2.69,
      "theoretical_loss": 3.686437096170368,
      "tokens_seen": 899724288
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.00036738214643931795,
      "loss": 2.7704,
      "theoretical_loss": 3.6864110416951075,
      "tokens_seen": 899789824
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.0003673721163490472,
      "loss": 2.8191,
      "theoretical_loss": 3.6863849896487544,
      "tokens_seen": 899855360
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.0003673620862587763,
      "loss": 2.8529,
      "theoretical_loss": 3.686358940030905,
      "tokens_seen": 899920896
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.00036735205616850555,
      "loss": 2.7385,
      "theoretical_loss": 3.686332892841156,
      "tokens_seen": 899986432
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.00036734202607823467,
      "loss": 2.9449,
      "theoretical_loss": 3.686306848079105,
      "tokens_seen": 900051968
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.0003673319959879639,
      "loss": 2.9013,
      "theoretical_loss": 3.686280805744348,
      "tokens_seen": 900117504
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.0003673219658976931,
      "loss": 2.87,
      "theoretical_loss": 3.686254765836483,
      "tokens_seen": 900183040
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.00036731193580742227,
      "loss": 2.7614,
      "theoretical_loss": 3.6862287283551067,
      "tokens_seen": 900248576
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.00036730190571715145,
      "loss": 2.7338,
      "theoretical_loss": 3.686202693299816,
      "tokens_seen": 900314112
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.00036729187562688063,
      "loss": 2.5981,
      "theoretical_loss": 3.686176660670209,
      "tokens_seen": 900379648
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.0003672818455366098,
      "loss": 2.7967,
      "theoretical_loss": 3.6861506304658826,
      "tokens_seen": 900445184
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.00036727181544633905,
      "loss": 2.9024,
      "theoretical_loss": 3.6861246026864354,
      "tokens_seen": 900510720
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.0003672617853560682,
      "loss": 2.8047,
      "theoretical_loss": 3.6860985773314634,
      "tokens_seen": 900576256
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.0003672517552657974,
      "loss": 2.6793,
      "theoretical_loss": 3.6860725544005657,
      "tokens_seen": 900641792
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.00036724172517552654,
      "loss": 2.9473,
      "theoretical_loss": 3.6860465338933395,
      "tokens_seen": 900707328
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.00036723169508525577,
      "loss": 2.6197,
      "theoretical_loss": 3.686020515809383,
      "tokens_seen": 900772864
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.00036722166499498495,
      "loss": 2.8145,
      "theoretical_loss": 3.6859945001482943,
      "tokens_seen": 900838400
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.00036721163490471414,
      "loss": 2.7004,
      "theoretical_loss": 3.685968486909671,
      "tokens_seen": 900903936
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.0003672016048144433,
      "loss": 2.9374,
      "theoretical_loss": 3.6859424760931123,
      "tokens_seen": 900969472
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.00036719157472417255,
      "loss": 2.7168,
      "theoretical_loss": 3.6859164676982155,
      "tokens_seen": 901035008
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.0003671815446339017,
      "loss": 2.7888,
      "theoretical_loss": 3.68589046172458,
      "tokens_seen": 901100544
    },
    {
      "epoch": 3.0,
      "objective/train/docs_used": 1056725,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.974947214126587,
      "objective/train/theoretical_loss": 3.685883960609446,
      "objective/train/tokens_used": 921576928,
      "theoretical_loss": 3.685883960609446,
      "tokens_seen": 901116928
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.0003671715145436309,
      "loss": 2.7201,
      "theoretical_loss": 3.685864458171803,
      "tokens_seen": 901166080
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.00036716148445336004,
      "loss": 2.7771,
      "theoretical_loss": 3.685838457039485,
      "tokens_seen": 901231616
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.0003671514543630893,
      "loss": 2.7884,
      "theoretical_loss": 3.6858124583272227,
      "tokens_seen": 901297152
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.00036714142427281846,
      "loss": 2.9321,
      "theoretical_loss": 3.685786462034616,
      "tokens_seen": 901362688
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.00036713139418254764,
      "loss": 2.7786,
      "theoretical_loss": 3.6857604681612646,
      "tokens_seen": 901428224
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.0003671213640922768,
      "loss": 2.7034,
      "theoretical_loss": 3.6857344767067657,
      "tokens_seen": 901493760
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.000367111334002006,
      "loss": 2.8064,
      "theoretical_loss": 3.6857084876707193,
      "tokens_seen": 901559296
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.00036710130391173524,
      "loss": 2.7424,
      "theoretical_loss": 3.6856825010527245,
      "tokens_seen": 901624832
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.0003670912738214644,
      "loss": 2.7533,
      "theoretical_loss": 3.685656516852381,
      "tokens_seen": 901690368
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.0003670812437311936,
      "loss": 2.9162,
      "theoretical_loss": 3.6856305350692873,
      "tokens_seen": 901755904
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.0003670712136409228,
      "loss": 2.7772,
      "theoretical_loss": 3.685604555703044,
      "tokens_seen": 901821440
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.000367061183550652,
      "loss": 2.8262,
      "theoretical_loss": 3.68557857875325,
      "tokens_seen": 901886976
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.00036705115346038114,
      "loss": 2.5888,
      "theoretical_loss": 3.6855526042195046,
      "tokens_seen": 901952512
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.0003670411233701104,
      "loss": 2.4701,
      "theoretical_loss": 3.685526632101408,
      "tokens_seen": 902018048
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.0003670310932798395,
      "loss": 3.0106,
      "theoretical_loss": 3.6855006623985602,
      "tokens_seen": 902083584
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.00036702106318956874,
      "loss": 2.5845,
      "theoretical_loss": 3.6854746951105612,
      "tokens_seen": 902149120
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.0003670110330992979,
      "loss": 2.738,
      "theoretical_loss": 3.6854487302370105,
      "tokens_seen": 902214656
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.0003670010030090271,
      "loss": 2.7965,
      "theoretical_loss": 3.685422767777509,
      "tokens_seen": 902280192
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.0003669909729187563,
      "loss": 2.5952,
      "theoretical_loss": 3.6853968077316566,
      "tokens_seen": 902345728
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.00036698094282848546,
      "loss": 2.7099,
      "theoretical_loss": 3.685370850099053,
      "tokens_seen": 902411264
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.00036697091273821464,
      "loss": 2.9008,
      "theoretical_loss": 3.6853448948792993,
      "tokens_seen": 902476800
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.0003669608826479439,
      "loss": 2.5409,
      "theoretical_loss": 3.685318942071996,
      "tokens_seen": 902542336
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.000366950852557673,
      "loss": 2.6951,
      "theoretical_loss": 3.685292991676744,
      "tokens_seen": 902607872
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.00036694082246740224,
      "loss": 2.9099,
      "theoretical_loss": 3.685267043693144,
      "tokens_seen": 902673408
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.00036693079237713137,
      "loss": 2.5979,
      "theoretical_loss": 3.6852410981207955,
      "tokens_seen": 902738944
    },
    {
      "epoch": 3.0,
      "objective/train/docs_used": 1061572,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.726848840713501,
      "objective/train/theoretical_loss": 3.68523461210442,
      "objective/train/tokens_used": 923215328,
      "theoretical_loss": 3.68523461210442,
      "tokens_seen": 902755328
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.0003669207622868606,
      "loss": 2.8534,
      "theoretical_loss": 3.685215154959301,
      "tokens_seen": 902804480
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.0003669107321965898,
      "loss": 2.8183,
      "theoretical_loss": 3.6851892142082607,
      "tokens_seen": 902870016
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.00036690070210631897,
      "loss": 2.7522,
      "theoretical_loss": 3.685163275867276,
      "tokens_seen": 902935552
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.00036689067201604815,
      "loss": 2.7613,
      "theoretical_loss": 3.685137339935948,
      "tokens_seen": 903001088
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.0003668806419257774,
      "loss": 2.7894,
      "theoretical_loss": 3.6851114064138777,
      "tokens_seen": 903066624
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.0003668706118355065,
      "loss": 2.8125,
      "theoretical_loss": 3.685085475300667,
      "tokens_seen": 903132160
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.00036686058174523575,
      "loss": 2.7282,
      "theoretical_loss": 3.685059546595917,
      "tokens_seen": 903197696
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.00036685055165496487,
      "loss": 2.8206,
      "theoretical_loss": 3.6850336202992295,
      "tokens_seen": 903263232
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.0003668405215646941,
      "loss": 2.7565,
      "theoretical_loss": 3.6850076964102056,
      "tokens_seen": 903328768
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.0003668304914744233,
      "loss": 2.7826,
      "theoretical_loss": 3.684981774928448,
      "tokens_seen": 903394304
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.00036682046138415247,
      "loss": 2.8357,
      "theoretical_loss": 3.684955855853558,
      "tokens_seen": 903459840
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.00036681043129388165,
      "loss": 2.8705,
      "theoretical_loss": 3.684929939185137,
      "tokens_seen": 903525376
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.00036680040120361083,
      "loss": 2.6921,
      "theoretical_loss": 3.6849040249227887,
      "tokens_seen": 903590912
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.00036679037111334,
      "loss": 2.8528,
      "theoretical_loss": 3.6848781130661137,
      "tokens_seen": 903656448
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.00036678034102306925,
      "loss": 2.7383,
      "theoretical_loss": 3.6848522036147147,
      "tokens_seen": 903721984
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.0003667703109327984,
      "loss": 2.7307,
      "theoretical_loss": 3.684826296568194,
      "tokens_seen": 903787520
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.0003667602808425276,
      "loss": 2.6868,
      "theoretical_loss": 3.6848003919261547,
      "tokens_seen": 903853056
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.00036675025075225674,
      "loss": 2.8841,
      "theoretical_loss": 3.6847744896881975,
      "tokens_seen": 903918592
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.000366740220661986,
      "loss": 2.8175,
      "theoretical_loss": 3.6847485898539274,
      "tokens_seen": 903984128
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.00036673019057171515,
      "loss": 2.577,
      "theoretical_loss": 3.6847226924229455,
      "tokens_seen": 904049664
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.00036672016048144434,
      "loss": 2.6837,
      "theoretical_loss": 3.684696797394855,
      "tokens_seen": 904115200
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.0003667101303911735,
      "loss": 2.7804,
      "theoretical_loss": 3.6846709047692587,
      "tokens_seen": 904180736
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.00036670010030090275,
      "loss": 2.847,
      "theoretical_loss": 3.68464501454576,
      "tokens_seen": 904246272
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.0003666900702106319,
      "loss": 2.7734,
      "theoretical_loss": 3.6846191267239616,
      "tokens_seen": 904311808
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.0003666800401203611,
      "loss": 2.7455,
      "theoretical_loss": 3.6845932413034665,
      "tokens_seen": 904377344
    },
    {
      "epoch": 3.0,
      "objective/train/docs_used": 1066630,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.873715877532959,
      "objective/train/theoretical_loss": 3.6845867703234996,
      "objective/train/tokens_used": 924853728,
      "theoretical_loss": 3.6845867703234996,
      "tokens_seen": 904393728
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.00036667001003009024,
      "loss": 2.8156,
      "theoretical_loss": 3.6845673582838785,
      "tokens_seen": 904442880
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.0003666599799398195,
      "loss": 2.7413,
      "theoretical_loss": 3.6845414776648004,
      "tokens_seen": 904508416
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.00036664994984954866,
      "loss": 2.7902,
      "theoretical_loss": 3.6845155994458363,
      "tokens_seen": 904573952
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.00036663991975927784,
      "loss": 2.6701,
      "theoretical_loss": 3.684489723626589,
      "tokens_seen": 904639488
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.000366629889669007,
      "loss": 2.6021,
      "theoretical_loss": 3.6844638502066633,
      "tokens_seen": 904705024
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.0003666198595787362,
      "loss": 2.6243,
      "theoretical_loss": 3.6844379791856614,
      "tokens_seen": 904770560
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.0003666098294884654,
      "loss": 2.8246,
      "theoretical_loss": 3.684412110563189,
      "tokens_seen": 904836096
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.0003665997993981946,
      "loss": 2.5495,
      "theoretical_loss": 3.6843862443388478,
      "tokens_seen": 904901632
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.00036658976930792374,
      "loss": 2.7142,
      "theoretical_loss": 3.684360380512244,
      "tokens_seen": 904967168
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.000366579739217653,
      "loss": 2.7124,
      "theoretical_loss": 3.68433451908298,
      "tokens_seen": 905032704
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.00036656970912738216,
      "loss": 2.8297,
      "theoretical_loss": 3.684308660050661,
      "tokens_seen": 905098240
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.00036655967903711134,
      "loss": 2.625,
      "theoretical_loss": 3.6842828034148916,
      "tokens_seen": 905163776
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.0003665496489468405,
      "loss": 2.7145,
      "theoretical_loss": 3.684256949175275,
      "tokens_seen": 905229312
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.0003665396188565697,
      "loss": 2.7347,
      "theoretical_loss": 3.6842310973314163,
      "tokens_seen": 905294848
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.0003665295887662989,
      "loss": 2.7359,
      "theoretical_loss": 3.684205247882921,
      "tokens_seen": 905360384
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.0003665195586760281,
      "loss": 2.8178,
      "theoretical_loss": 3.684179400829392,
      "tokens_seen": 905425920
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.00036650952858575725,
      "loss": 2.8469,
      "theoretical_loss": 3.6841535561704353,
      "tokens_seen": 905491456
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.0003664994984954865,
      "loss": 2.7027,
      "theoretical_loss": 3.6841277139056556,
      "tokens_seen": 905556992
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.0003664894684052156,
      "loss": 2.7502,
      "theoretical_loss": 3.6841018740346576,
      "tokens_seen": 905622528
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.00036647943831494484,
      "loss": 2.7623,
      "theoretical_loss": 3.6840760365570464,
      "tokens_seen": 905688064
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.000366469408224674,
      "loss": 2.6832,
      "theoretical_loss": 3.684050201472428,
      "tokens_seen": 905753600
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.0003664593781344032,
      "loss": 2.9027,
      "theoretical_loss": 3.684024368780406,
      "tokens_seen": 905819136
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.0003664493480441324,
      "loss": 2.8143,
      "theoretical_loss": 3.683998538480587,
      "tokens_seen": 905884672
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.00036643931795386157,
      "loss": 2.7001,
      "theoretical_loss": 3.6839727105725766,
      "tokens_seen": 905950208
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.00036642928786359075,
      "loss": 2.7919,
      "theoretical_loss": 3.683946885055979,
      "tokens_seen": 906015744
    },
    {
      "epoch": 3.0,
      "objective/train/docs_used": 1071650,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7729568481445312,
      "objective/train/theoretical_loss": 3.683940429050442,
      "objective/train/tokens_used": 926492128,
      "theoretical_loss": 3.683940429050442,
      "tokens_seen": 906032128
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.00036641925777332,
      "loss": 2.7335,
      "theoretical_loss": 3.683921061930401,
      "tokens_seen": 906081280
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.0003664092276830491,
      "loss": 2.7948,
      "theoretical_loss": 3.6838952411954473,
      "tokens_seen": 906146816
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.00036639919759277835,
      "loss": 2.8415,
      "theoretical_loss": 3.6838694228507256,
      "tokens_seen": 906212352
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.00036638916750250753,
      "loss": 2.8463,
      "theoretical_loss": 3.68384360689584,
      "tokens_seen": 906277888
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.0003663791374122367,
      "loss": 2.9069,
      "theoretical_loss": 3.6838177933303964,
      "tokens_seen": 906343424
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.0003663691073219659,
      "loss": 2.8556,
      "theoretical_loss": 3.683791982154002,
      "tokens_seen": 906408960
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.00036635907723169507,
      "loss": 2.725,
      "theoretical_loss": 3.683766173366263,
      "tokens_seen": 906474496
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.0003663490471414243,
      "loss": 2.5566,
      "theoretical_loss": 3.6837403669667843,
      "tokens_seen": 906540032
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.0003663390170511535,
      "loss": 2.7766,
      "theoretical_loss": 3.683714562955174,
      "tokens_seen": 906605568
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.00036632898696088267,
      "loss": 2.9063,
      "theoretical_loss": 3.6836887613310374,
      "tokens_seen": 906671104
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.00036631895687061185,
      "loss": 2.8221,
      "theoretical_loss": 3.6836629620939814,
      "tokens_seen": 906736640
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036630892678034103,
      "loss": 2.9086,
      "theoretical_loss": 3.6836371652436126,
      "tokens_seen": 906802176
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003662988966900702,
      "loss": 2.8452,
      "theoretical_loss": 3.6836113707795377,
      "tokens_seen": 906867712
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036628886659979945,
      "loss": 2.7909,
      "theoretical_loss": 3.6835855787013636,
      "tokens_seen": 906933248
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003662788365095286,
      "loss": 2.8312,
      "theoretical_loss": 3.6835597890086973,
      "tokens_seen": 906998784
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003662688064192578,
      "loss": 2.5495,
      "theoretical_loss": 3.6835340017011458,
      "tokens_seen": 907064320
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036625877632898694,
      "loss": 2.7974,
      "theoretical_loss": 3.683508216778316,
      "tokens_seen": 907129856
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003662487462387162,
      "loss": 2.7805,
      "theoretical_loss": 3.6834824342398154,
      "tokens_seen": 907195392
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036623871614844535,
      "loss": 2.8686,
      "theoretical_loss": 3.6834566540852514,
      "tokens_seen": 907260928
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036622868605817454,
      "loss": 2.9814,
      "theoretical_loss": 3.683430876314231,
      "tokens_seen": 907326464
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003662186559679037,
      "loss": 2.874,
      "theoretical_loss": 3.683405100926362,
      "tokens_seen": 907392000
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036620862587763295,
      "loss": 2.522,
      "theoretical_loss": 3.6833793279212514,
      "tokens_seen": 907457536
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003661985957873621,
      "loss": 2.6894,
      "theoretical_loss": 3.683353557298507,
      "tokens_seen": 907523072
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003661885656970913,
      "loss": 2.8295,
      "theoretical_loss": 3.683327789057738,
      "tokens_seen": 907588608
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036617853560682044,
      "loss": 2.799,
      "theoretical_loss": 3.683302023198551,
      "tokens_seen": 907654144
    },
    {
      "epoch": 3.01,
      "objective/train/docs_used": 1076755,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.107651472091675,
      "objective/train/theoretical_loss": 3.68329558210583,
      "objective/train/tokens_used": 928130528,
      "theoretical_loss": 3.68329558210583,
      "tokens_seen": 907670528
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003661685055165497,
      "loss": 2.9,
      "theoretical_loss": 3.6832762597205533,
      "tokens_seen": 907719680
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036615847542627886,
      "loss": 2.9029,
      "theoretical_loss": 3.6832504986233543,
      "tokens_seen": 907785216
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036614844533600804,
      "loss": 2.8958,
      "theoretical_loss": 3.683224739906561,
      "tokens_seen": 907850752
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003661384152457372,
      "loss": 2.709,
      "theoretical_loss": 3.683198983569783,
      "tokens_seen": 907916288
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003661283851554664,
      "loss": 2.6734,
      "theoretical_loss": 3.6831732296126276,
      "tokens_seen": 907981824
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003661183550651956,
      "loss": 2.746,
      "theoretical_loss": 3.683147478034703,
      "tokens_seen": 908047360
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003661083249749248,
      "loss": 2.7898,
      "theoretical_loss": 3.6831217288356184,
      "tokens_seen": 908112896
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036609829488465394,
      "loss": 2.7533,
      "theoretical_loss": 3.683095982014982,
      "tokens_seen": 908178432
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003660882647943832,
      "loss": 2.5579,
      "theoretical_loss": 3.6830702375724025,
      "tokens_seen": 908243968
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036607823470411236,
      "loss": 2.7278,
      "theoretical_loss": 3.6830444955074895,
      "tokens_seen": 908309504
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036606820461384154,
      "loss": 2.7775,
      "theoretical_loss": 3.683018755819851,
      "tokens_seen": 908375040
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003660581745235707,
      "loss": 2.8285,
      "theoretical_loss": 3.6829930185090958,
      "tokens_seen": 908440576
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003660481444332999,
      "loss": 2.7422,
      "theoretical_loss": 3.6829672835748335,
      "tokens_seen": 908506112
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003660381143430291,
      "loss": 2.8331,
      "theoretical_loss": 3.682941551016673,
      "tokens_seen": 908571648
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003660280842527583,
      "loss": 2.7105,
      "theoretical_loss": 3.682915820834224,
      "tokens_seen": 908637184
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036601805416248745,
      "loss": 2.7307,
      "theoretical_loss": 3.682890093027095,
      "tokens_seen": 908702720
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003660080240722167,
      "loss": 2.8438,
      "theoretical_loss": 3.682864367594896,
      "tokens_seen": 908768256
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003659979939819458,
      "loss": 2.9085,
      "theoretical_loss": 3.682838644537237,
      "tokens_seen": 908833792
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036598796389167504,
      "loss": 2.607,
      "theoretical_loss": 3.6828129238537266,
      "tokens_seen": 908899328
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003659779338014042,
      "loss": 2.7233,
      "theoretical_loss": 3.682787205543975,
      "tokens_seen": 908964864
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003659679037111334,
      "loss": 2.5585,
      "theoretical_loss": 3.682761489607592,
      "tokens_seen": 909030400
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003659578736208626,
      "loss": 2.6669,
      "theoretical_loss": 3.6827357760441877,
      "tokens_seen": 909095936
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036594784353059177,
      "loss": 2.7123,
      "theoretical_loss": 3.6827100648533717,
      "tokens_seen": 909161472
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036593781344032095,
      "loss": 2.8216,
      "theoretical_loss": 3.682684356034754,
      "tokens_seen": 909227008
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003659277833500502,
      "loss": 2.5725,
      "theoretical_loss": 3.682658649587945,
      "tokens_seen": 909292544
    },
    {
      "epoch": 3.01,
      "objective/train/docs_used": 1081876,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.961531162261963,
      "objective/train/theoretical_loss": 3.6826522233467927,
      "objective/train/tokens_used": 929768928,
      "theoretical_loss": 3.6826522233467927,
      "tokens_seen": 909308928
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003659177532597793,
      "loss": 2.6922,
      "theoretical_loss": 3.6826329455125553,
      "tokens_seen": 909358080
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036590772316950855,
      "loss": 2.8059,
      "theoretical_loss": 3.682607243808195,
      "tokens_seen": 909423616
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036589769307923773,
      "loss": 2.6864,
      "theoretical_loss": 3.6825815444744743,
      "tokens_seen": 909489152
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003658876629889669,
      "loss": 2.6581,
      "theoretical_loss": 3.682555847511004,
      "tokens_seen": 909554688
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003658776328986961,
      "loss": 2.6658,
      "theoretical_loss": 3.6825301529173946,
      "tokens_seen": 909620224
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036586760280842527,
      "loss": 2.9523,
      "theoretical_loss": 3.6825044606932567,
      "tokens_seen": 909685760
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036585757271815445,
      "loss": 2.8693,
      "theoretical_loss": 3.682478770838202,
      "tokens_seen": 909751296
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003658475426278837,
      "loss": 2.8812,
      "theoretical_loss": 3.6824530833518394,
      "tokens_seen": 909816832
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003658375125376128,
      "loss": 2.7843,
      "theoretical_loss": 3.682427398233782,
      "tokens_seen": 909882368
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036582748244734205,
      "loss": 2.7056,
      "theoretical_loss": 3.68240171548364,
      "tokens_seen": 909947904
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003658174523570712,
      "loss": 2.8099,
      "theoretical_loss": 3.682376035101025,
      "tokens_seen": 910013440
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003658074222668004,
      "loss": 2.5475,
      "theoretical_loss": 3.6823503570855474,
      "tokens_seen": 910078976
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003657973921765296,
      "loss": 2.8597,
      "theoretical_loss": 3.6823246814368193,
      "tokens_seen": 910144512
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003657873620862588,
      "loss": 2.8058,
      "theoretical_loss": 3.6822990081544518,
      "tokens_seen": 910210048
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036577733199598796,
      "loss": 2.8276,
      "theoretical_loss": 3.6822733372380574,
      "tokens_seen": 910275584
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036576730190571714,
      "loss": 2.5292,
      "theoretical_loss": 3.6822476686872463,
      "tokens_seen": 910341120
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003657572718154463,
      "loss": 2.7811,
      "theoretical_loss": 3.682222002501631,
      "tokens_seen": 910406656
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036574724172517555,
      "loss": 2.9896,
      "theoretical_loss": 3.682196338680823,
      "tokens_seen": 910472192
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003657372116349047,
      "loss": 2.6515,
      "theoretical_loss": 3.682170677224435,
      "tokens_seen": 910537728
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003657271815446339,
      "loss": 2.842,
      "theoretical_loss": 3.682145018132078,
      "tokens_seen": 910603264
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003657171514543631,
      "loss": 2.5961,
      "theoretical_loss": 3.6821193614033643,
      "tokens_seen": 910668800
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003657071213640923,
      "loss": 2.8136,
      "theoretical_loss": 3.682093707037907,
      "tokens_seen": 910734336
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036569709127382146,
      "loss": 2.8072,
      "theoretical_loss": 3.6820680550353178,
      "tokens_seen": 910799872
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036568706118355064,
      "loss": 2.8705,
      "theoretical_loss": 3.682042405395208,
      "tokens_seen": 910865408
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003656770310932798,
      "loss": 2.6743,
      "theoretical_loss": 3.682016758117192,
      "tokens_seen": 910930944
    },
    {
      "epoch": 3.01,
      "objective/train/docs_used": 1086888,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.870638132095337,
      "objective/train/theoretical_loss": 3.6820103466667193,
      "objective/train/tokens_used": 931407328,
      "theoretical_loss": 3.6820103466667193,
      "tokens_seen": 910947328
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036566700100300906,
      "loss": 2.7901,
      "theoretical_loss": 3.681991113200881,
      "tokens_seen": 910996480
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003656569709127382,
      "loss": 2.7127,
      "theoretical_loss": 3.6819654706458884,
      "tokens_seen": 911062016
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003656469408224674,
      "loss": 2.7062,
      "theoretical_loss": 3.6819398304518267,
      "tokens_seen": 911127552
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036563691073219655,
      "loss": 2.7506,
      "theoretical_loss": 3.6819141926183083,
      "tokens_seen": 911193088
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003656268806419258,
      "loss": 2.6625,
      "theoretical_loss": 3.681888557144947,
      "tokens_seen": 911258624
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036561685055165496,
      "loss": 2.6967,
      "theoretical_loss": 3.6818629240313543,
      "tokens_seen": 911324160
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036560682046138414,
      "loss": 2.8149,
      "theoretical_loss": 3.6818372932771455,
      "tokens_seen": 911389696
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003655967903711134,
      "loss": 2.6254,
      "theoretical_loss": 3.681811664881932,
      "tokens_seen": 911455232
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036558676028084256,
      "loss": 2.7563,
      "theoretical_loss": 3.6817860388453276,
      "tokens_seen": 911520768
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036557673019057174,
      "loss": 2.7822,
      "theoretical_loss": 3.681760415166946,
      "tokens_seen": 911586304
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003655667001003009,
      "loss": 2.8973,
      "theoretical_loss": 3.6817347938464002,
      "tokens_seen": 911651840
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003655566700100301,
      "loss": 2.607,
      "theoretical_loss": 3.681709174883304,
      "tokens_seen": 911717376
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003655466399197593,
      "loss": 2.6586,
      "theoretical_loss": 3.681683558277271,
      "tokens_seen": 911782912
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003655366098294885,
      "loss": 2.852,
      "theoretical_loss": 3.6816579440279154,
      "tokens_seen": 911848448
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036552657973921765,
      "loss": 2.7226,
      "theoretical_loss": 3.6816323321348507,
      "tokens_seen": 911913984
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003655165496489469,
      "loss": 2.6382,
      "theoretical_loss": 3.68160672259769,
      "tokens_seen": 911979520
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.000365506519558676,
      "loss": 2.8484,
      "theoretical_loss": 3.6815811154160487,
      "tokens_seen": 912045056
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036549648946840525,
      "loss": 2.6745,
      "theoretical_loss": 3.68155551058954,
      "tokens_seen": 912110592
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003654864593781344,
      "loss": 2.8237,
      "theoretical_loss": 3.6815299081177786,
      "tokens_seen": 912176128
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003654764292878636,
      "loss": 3.1232,
      "theoretical_loss": 3.6815043080003784,
      "tokens_seen": 912241664
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003654663991975928,
      "loss": 2.6531,
      "theoretical_loss": 3.6814787102369544,
      "tokens_seen": 912307200
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036545636910732197,
      "loss": 2.7191,
      "theoretical_loss": 3.6814531148271206,
      "tokens_seen": 912372736
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036544633901705115,
      "loss": 2.5767,
      "theoretical_loss": 3.681427521770491,
      "tokens_seen": 912438272
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003654363089267804,
      "loss": 2.8236,
      "theoretical_loss": 3.6814019310666812,
      "tokens_seen": 912503808
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003654262788365095,
      "loss": 2.7541,
      "theoretical_loss": 3.6813763427153052,
      "tokens_seen": 912569344
    },
    {
      "epoch": 3.01,
      "objective/train/docs_used": 1091883,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6327216625213623,
      "objective/train/theoretical_loss": 3.6813699459949842,
      "objective/train/tokens_used": 933045728,
      "theoretical_loss": 3.6813699459949842,
      "tokens_seen": 912585728
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036541624874623875,
      "loss": 2.6855,
      "theoretical_loss": 3.681350756715979,
      "tokens_seen": 912634880
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036540621865596793,
      "loss": 2.6951,
      "theoretical_loss": 3.6813251730683163,
      "tokens_seen": 912700416
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003653961885656971,
      "loss": 2.8262,
      "theoretical_loss": 3.681299591771933,
      "tokens_seen": 912765952
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003653861584754263,
      "loss": 2.7102,
      "theoretical_loss": 3.681274012826443,
      "tokens_seen": 912831488
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036537612838515547,
      "loss": 2.6939,
      "theoretical_loss": 3.6812484362314626,
      "tokens_seen": 912897024
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036536609829488465,
      "loss": 2.8567,
      "theoretical_loss": 3.681222861986607,
      "tokens_seen": 912962560
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003653560682046139,
      "loss": 2.7321,
      "theoretical_loss": 3.6811972900914913,
      "tokens_seen": 913028096
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.000365346038114343,
      "loss": 2.6525,
      "theoretical_loss": 3.6811717205457306,
      "tokens_seen": 913093632
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036533600802407225,
      "loss": 2.6161,
      "theoretical_loss": 3.6811461533489416,
      "tokens_seen": 913159168
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003653259779338014,
      "loss": 2.7166,
      "theoretical_loss": 3.6811205885007388,
      "tokens_seen": 913224704
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003653159478435306,
      "loss": 2.5017,
      "theoretical_loss": 3.6810950260007385,
      "tokens_seen": 913290240
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003653059177532598,
      "loss": 2.8044,
      "theoretical_loss": 3.681069465848556,
      "tokens_seen": 913355776
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.000365295887662989,
      "loss": 2.7958,
      "theoretical_loss": 3.681043908043808,
      "tokens_seen": 913421312
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036528585757271816,
      "loss": 2.7558,
      "theoretical_loss": 3.6810183525861095,
      "tokens_seen": 913486848
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036527582748244734,
      "loss": 2.6713,
      "theoretical_loss": 3.680992799475078,
      "tokens_seen": 913552384
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003652657973921765,
      "loss": 2.7104,
      "theoretical_loss": 3.6809672487103287,
      "tokens_seen": 913617920
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036525576730190575,
      "loss": 2.8065,
      "theoretical_loss": 3.680941700291478,
      "tokens_seen": 913683456
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003652457372116349,
      "loss": 2.6175,
      "theoretical_loss": 3.680916154218143,
      "tokens_seen": 913748992
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003652357071213641,
      "loss": 2.7811,
      "theoretical_loss": 3.680890610489939,
      "tokens_seen": 913814528
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003652256770310933,
      "loss": 2.7338,
      "theoretical_loss": 3.680865069106483,
      "tokens_seen": 913880064
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003652156469408225,
      "loss": 2.7676,
      "theoretical_loss": 3.680839530067392,
      "tokens_seen": 913945600
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036520561685055166,
      "loss": 2.6906,
      "theoretical_loss": 3.680813993372282,
      "tokens_seen": 914011136
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036519558676028084,
      "loss": 2.9039,
      "theoretical_loss": 3.680788459020771,
      "tokens_seen": 914076672
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036518555667001,
      "loss": 2.7388,
      "theoretical_loss": 3.680762927012475,
      "tokens_seen": 914142208
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036517552657973926,
      "loss": 2.8223,
      "theoretical_loss": 3.680737397347011,
      "tokens_seen": 914207744
    },
    {
      "epoch": 3.01,
      "objective/train/docs_used": 1093340,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.954491138458252,
      "objective/train/theoretical_loss": 3.6807310152966677,
      "objective/train/tokens_used": 934684128,
      "theoretical_loss": 3.6807310152966677,
      "tokens_seen": 914224128
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003651654964894684,
      "loss": 3.0444,
      "theoretical_loss": 3.6807118700239965,
      "tokens_seen": 914273280
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003651554663991976,
      "loss": 2.813,
      "theoretical_loss": 3.6806863450430485,
      "tokens_seen": 914338816
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036514543630892675,
      "loss": 2.7931,
      "theoretical_loss": 3.680660822403784,
      "tokens_seen": 914404352
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.000365135406218656,
      "loss": 2.7916,
      "theoretical_loss": 3.680635302105821,
      "tokens_seen": 914469888
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036512537612838516,
      "loss": 2.5788,
      "theoretical_loss": 3.6806097841487766,
      "tokens_seen": 914535424
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036511534603811434,
      "loss": 2.7142,
      "theoretical_loss": 3.680584268532268,
      "tokens_seen": 914600960
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003651053159478435,
      "loss": 2.8717,
      "theoretical_loss": 3.680558755255914,
      "tokens_seen": 914666496
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036509528585757276,
      "loss": 2.8554,
      "theoretical_loss": 3.6805332443193306,
      "tokens_seen": 914732032
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003650852557673019,
      "loss": 2.6752,
      "theoretical_loss": 3.680507735722137,
      "tokens_seen": 914797568
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003650752256770311,
      "loss": 2.6707,
      "theoretical_loss": 3.6804822294639505,
      "tokens_seen": 914863104
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036506519558676025,
      "loss": 2.7038,
      "theoretical_loss": 3.6804567255443894,
      "tokens_seen": 914928640
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003650551654964895,
      "loss": 2.8116,
      "theoretical_loss": 3.680431223963071,
      "tokens_seen": 914994176
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036504513540621867,
      "loss": 2.8997,
      "theoretical_loss": 3.6804057247196145,
      "tokens_seen": 915059712
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036503510531594785,
      "loss": 2.6417,
      "theoretical_loss": 3.6803802278136377,
      "tokens_seen": 915125248
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036502507522567703,
      "loss": 2.8395,
      "theoretical_loss": 3.6803547332447595,
      "tokens_seen": 915190784
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003650150451354062,
      "loss": 2.854,
      "theoretical_loss": 3.680329241012597,
      "tokens_seen": 915256320
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003650050150451354,
      "loss": 2.9282,
      "theoretical_loss": 3.68030375111677,
      "tokens_seen": 915321856
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003649949849548646,
      "loss": 2.7012,
      "theoretical_loss": 3.6802782635568967,
      "tokens_seen": 915387392
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036498495486459375,
      "loss": 2.7663,
      "theoretical_loss": 3.6802527783325956,
      "tokens_seen": 915452928
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.000364974924774323,
      "loss": 2.8494,
      "theoretical_loss": 3.680227295443486,
      "tokens_seen": 915518464
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003649648946840521,
      "loss": 2.595,
      "theoretical_loss": 3.680201814889186,
      "tokens_seen": 915584000
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036495486459378135,
      "loss": 2.8819,
      "theoretical_loss": 3.6801763366693154,
      "tokens_seen": 915649536
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036494483450351053,
      "loss": 2.5959,
      "theoretical_loss": 3.6801508607834927,
      "tokens_seen": 915715072
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003649348044132397,
      "loss": 2.773,
      "theoretical_loss": 3.680125387231338,
      "tokens_seen": 915780608
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003649247743229689,
      "loss": 2.7749,
      "theoretical_loss": 3.680099916012469,
      "tokens_seen": 915846144
    },
    {
      "epoch": 3.01,
      "objective/train/docs_used": 1093676,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.2436115741729736,
      "objective/train/theoretical_loss": 3.6800935485722834,
      "objective/train/tokens_used": 936322528,
      "theoretical_loss": 3.6800935485722834,
      "tokens_seen": 915862528
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036491474423269813,
      "loss": 2.8122,
      "theoretical_loss": 3.6800744471265063,
      "tokens_seen": 915911680
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036490471414242726,
      "loss": 2.7161,
      "theoretical_loss": 3.680048980573069,
      "tokens_seen": 915977216
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003648946840521565,
      "loss": 2.7632,
      "theoretical_loss": 3.6800235163517767,
      "tokens_seen": 916042752
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003648846539618856,
      "loss": 2.7306,
      "theoretical_loss": 3.6799980544622484,
      "tokens_seen": 916108288
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036487462387161485,
      "loss": 2.8455,
      "theoretical_loss": 3.679972594904104,
      "tokens_seen": 916173824
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003648645937813441,
      "loss": 2.7776,
      "theoretical_loss": 3.679947137676964,
      "tokens_seen": 916239360
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003648545636910732,
      "loss": 2.7945,
      "theoretical_loss": 3.6799216827804475,
      "tokens_seen": 916304896
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036484453360080245,
      "loss": 2.7072,
      "theoretical_loss": 3.679896230214175,
      "tokens_seen": 916370432
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003648345035105316,
      "loss": 2.6346,
      "theoretical_loss": 3.679870779977766,
      "tokens_seen": 916435968
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003648244734202608,
      "loss": 2.9591,
      "theoretical_loss": 3.6798453320708413,
      "tokens_seen": 916501504
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036481444332999,
      "loss": 2.687,
      "theoretical_loss": 3.6798198864930205,
      "tokens_seen": 916567040
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003648044132397192,
      "loss": 2.5467,
      "theoretical_loss": 3.679794443243924,
      "tokens_seen": 916632576
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036479438314944836,
      "loss": 2.6591,
      "theoretical_loss": 3.6797690023231726,
      "tokens_seen": 916698112
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036478435305917754,
      "loss": 2.8308,
      "theoretical_loss": 3.6797435637303866,
      "tokens_seen": 916763648
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003647743229689067,
      "loss": 2.8794,
      "theoretical_loss": 3.6797181274651867,
      "tokens_seen": 916829184
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036476429287863595,
      "loss": 2.9907,
      "theoretical_loss": 3.679692693527193,
      "tokens_seen": 916894720
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003647542627883651,
      "loss": 2.7011,
      "theoretical_loss": 3.679667261916027,
      "tokens_seen": 916960256
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003647442326980943,
      "loss": 2.816,
      "theoretical_loss": 3.6796418326313094,
      "tokens_seen": 917025792
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003647342026078235,
      "loss": 2.8047,
      "theoretical_loss": 3.679616405672661,
      "tokens_seen": 917091328
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003647241725175527,
      "loss": 2.5946,
      "theoretical_loss": 3.679590981039703,
      "tokens_seen": 917156864
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036471414242728186,
      "loss": 2.7901,
      "theoretical_loss": 3.6795655587320555,
      "tokens_seen": 917222400
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036470411233701104,
      "loss": 2.6762,
      "theoretical_loss": 3.679540138749341,
      "tokens_seen": 917287936
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003646940822467402,
      "loss": 2.7547,
      "theoretical_loss": 3.6795147210911807,
      "tokens_seen": 917353472
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036468405215646946,
      "loss": 2.8656,
      "theoretical_loss": 3.679489305757196,
      "tokens_seen": 917419008
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003646740220661986,
      "loss": 2.8932,
      "theoretical_loss": 3.679463892747007,
      "tokens_seen": 917484544
    },
    {
      "epoch": 3.01,
      "objective/train/docs_used": 1095045,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7306671142578125,
      "objective/train/theoretical_loss": 3.6794575398575087,
      "objective/train/tokens_used": 937960928,
      "theoretical_loss": 3.6794575398575087,
      "tokens_seen": 917500928
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003646639919759278,
      "loss": 2.831,
      "theoretical_loss": 3.6794384820602364,
      "tokens_seen": 917550080
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036465396188565695,
      "loss": 2.8848,
      "theoretical_loss": 3.679413073696506,
      "tokens_seen": 917615616
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003646439317953862,
      "loss": 2.7306,
      "theoretical_loss": 3.679387667655438,
      "tokens_seen": 917681152
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036463390170511536,
      "loss": 2.708,
      "theoretical_loss": 3.6793622639366523,
      "tokens_seen": 917746688
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036462387161484454,
      "loss": 2.6047,
      "theoretical_loss": 3.679336862539772,
      "tokens_seen": 917812224
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003646138415245737,
      "loss": 2.7643,
      "theoretical_loss": 3.67931146346442,
      "tokens_seen": 917877760
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036460381143430296,
      "loss": 2.715,
      "theoretical_loss": 3.679286066710217,
      "tokens_seen": 917943296
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003645937813440321,
      "loss": 2.6738,
      "theoretical_loss": 3.679260672276786,
      "tokens_seen": 918008832
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003645837512537613,
      "loss": 2.8271,
      "theoretical_loss": 3.679235280163749,
      "tokens_seen": 918074368
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036457372116349045,
      "loss": 2.6745,
      "theoretical_loss": 3.679209890370728,
      "tokens_seen": 918139904
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003645636910732197,
      "loss": 2.9066,
      "theoretical_loss": 3.6791845028973453,
      "tokens_seen": 918205440
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036455366098294887,
      "loss": 2.7727,
      "theoretical_loss": 3.6791591177432252,
      "tokens_seen": 918270976
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036454363089267805,
      "loss": 2.7415,
      "theoretical_loss": 3.679133734907988,
      "tokens_seen": 918336512
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036453360080240723,
      "loss": 2.9084,
      "theoretical_loss": 3.679108354391258,
      "tokens_seen": 918402048
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003645235707121364,
      "loss": 2.7322,
      "theoretical_loss": 3.6790829761926567,
      "tokens_seen": 918467584
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003645135406218656,
      "loss": 2.6849,
      "theoretical_loss": 3.6790576003118085,
      "tokens_seen": 918533120
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003645035105315948,
      "loss": 2.623,
      "theoretical_loss": 3.679032226748335,
      "tokens_seen": 918598656
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036449348044132395,
      "loss": 2.8732,
      "theoretical_loss": 3.6790068555018607,
      "tokens_seen": 918664192
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003644834503510532,
      "loss": 2.6817,
      "theoretical_loss": 3.678981486572007,
      "tokens_seen": 918729728
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003644734202607823,
      "loss": 2.8337,
      "theoretical_loss": 3.6789561199583987,
      "tokens_seen": 918795264
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036446339017051155,
      "loss": 2.8237,
      "theoretical_loss": 3.6789307556606583,
      "tokens_seen": 918860800
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036445336008024073,
      "loss": 2.7562,
      "theoretical_loss": 3.6789053936784093,
      "tokens_seen": 918926336
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003644433299899699,
      "loss": 2.6789,
      "theoretical_loss": 3.6788800340112755,
      "tokens_seen": 918991872
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003644332998996991,
      "loss": 2.8857,
      "theoretical_loss": 3.6788546766588803,
      "tokens_seen": 919057408
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036442326980942833,
      "loss": 2.9261,
      "theoretical_loss": 3.6788293216208476,
      "tokens_seen": 919122944
    },
    {
      "epoch": 3.01,
      "objective/train/docs_used": 1095626,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8860342502593994,
      "objective/train/theoretical_loss": 3.678822983222914,
      "objective/train/tokens_used": 939599328,
      "theoretical_loss": 3.678822983222914,
      "tokens_seen": 919139328
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036441323971915746,
      "loss": 2.7617,
      "theoretical_loss": 3.6788039688968004,
      "tokens_seen": 919188480
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003644032096288867,
      "loss": 2.6231,
      "theoretical_loss": 3.678778618486364,
      "tokens_seen": 919254016
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003643931795386158,
      "loss": 2.918,
      "theoretical_loss": 3.678753270389161,
      "tokens_seen": 919319552
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036438314944834505,
      "loss": 2.6964,
      "theoretical_loss": 3.6787279246048157,
      "tokens_seen": 919385088
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036437311935807424,
      "loss": 2.5647,
      "theoretical_loss": 3.6787025811329523,
      "tokens_seen": 919450624
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003643630892678034,
      "loss": 2.8001,
      "theoretical_loss": 3.6786772399731955,
      "tokens_seen": 919516160
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003643530591775326,
      "loss": 2.8182,
      "theoretical_loss": 3.6786519011251695,
      "tokens_seen": 919581696
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003643430290872618,
      "loss": 2.8766,
      "theoretical_loss": 3.678626564588498,
      "tokens_seen": 919647232
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036433299899699096,
      "loss": 2.6051,
      "theoretical_loss": 3.678601230362806,
      "tokens_seen": 919712768
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003643229689067202,
      "loss": 2.6735,
      "theoretical_loss": 3.678575898447718,
      "tokens_seen": 919778304
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003643129388164493,
      "loss": 2.7497,
      "theoretical_loss": 3.6785505688428586,
      "tokens_seen": 919843840
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036430290872617856,
      "loss": 2.7237,
      "theoretical_loss": 3.6785252415478524,
      "tokens_seen": 919909376
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003642928786359077,
      "loss": 2.676,
      "theoretical_loss": 3.678499916562325,
      "tokens_seen": 919974912
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003642828485456369,
      "loss": 2.7326,
      "theoretical_loss": 3.6784745938859,
      "tokens_seen": 920040448
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003642728184553661,
      "loss": 2.7578,
      "theoretical_loss": 3.6784492735182033,
      "tokens_seen": 920105984
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003642627883650953,
      "loss": 2.8639,
      "theoretical_loss": 3.67842395545886,
      "tokens_seen": 920171520
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036425275827482446,
      "loss": 2.8007,
      "theoretical_loss": 3.678398639707495,
      "tokens_seen": 920237056
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003642427281845537,
      "loss": 2.8858,
      "theoretical_loss": 3.6783733262637335,
      "tokens_seen": 920302592
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003642326980942828,
      "loss": 2.7203,
      "theoretical_loss": 3.678348015127201,
      "tokens_seen": 920368128
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036422266800401206,
      "loss": 2.795,
      "theoretical_loss": 3.678322706297523,
      "tokens_seen": 920433664
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003642126379137412,
      "loss": 2.7759,
      "theoretical_loss": 3.678297399774325,
      "tokens_seen": 920499200
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003642026078234704,
      "loss": 2.5123,
      "theoretical_loss": 3.678272095557232,
      "tokens_seen": 920564736
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003641925777331996,
      "loss": 2.8862,
      "theoretical_loss": 3.6782467936458705,
      "tokens_seen": 920630272
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003641825476429288,
      "loss": 2.7438,
      "theoretical_loss": 3.6782214940398665,
      "tokens_seen": 920695808
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036417251755265797,
      "loss": 2.6299,
      "theoretical_loss": 3.6781961967388446,
      "tokens_seen": 920761344
    },
    {
      "epoch": 3.01,
      "objective/train/docs_used": 1097042,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9992458820343018,
      "objective/train/theoretical_loss": 3.6781898727736992,
      "objective/train/tokens_used": 941237728,
      "theoretical_loss": 3.6781898727736992,
      "tokens_seen": 920777728
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036416248746238715,
      "loss": 2.9299,
      "theoretical_loss": 3.6781709017424324,
      "tokens_seen": 920826880
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036415245737211633,
      "loss": 2.8001,
      "theoretical_loss": 3.6781456090502544,
      "tokens_seen": 920892416
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036414242728184556,
      "loss": 2.7548,
      "theoretical_loss": 3.678120318661938,
      "tokens_seen": 920957952
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003641323971915747,
      "loss": 2.8482,
      "theoretical_loss": 3.6780950305771087,
      "tokens_seen": 921023488
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003641223671013039,
      "loss": 2.9371,
      "theoretical_loss": 3.6780697447953927,
      "tokens_seen": 921089024
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036411233701103316,
      "loss": 2.6601,
      "theoretical_loss": 3.678044461316417,
      "tokens_seen": 921154560
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003641023069207623,
      "loss": 2.464,
      "theoretical_loss": 3.678019180139808,
      "tokens_seen": 921220096
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003640922768304915,
      "loss": 2.8123,
      "theoretical_loss": 3.677993901265191,
      "tokens_seen": 921285632
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036408224674022065,
      "loss": 2.6211,
      "theoretical_loss": 3.6779686246921948,
      "tokens_seen": 921351168
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003640722166499499,
      "loss": 2.7441,
      "theoretical_loss": 3.6779433504204446,
      "tokens_seen": 921416704
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036406218655967907,
      "loss": 2.6296,
      "theoretical_loss": 3.6779180784495678,
      "tokens_seen": 921482240
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036405215646940825,
      "loss": 2.7529,
      "theoretical_loss": 3.677892808779191,
      "tokens_seen": 921547776
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036404212637913743,
      "loss": 2.6744,
      "theoretical_loss": 3.6778675414089417,
      "tokens_seen": 921613312
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003640320962888666,
      "loss": 2.568,
      "theoretical_loss": 3.6778422763384464,
      "tokens_seen": 921678848
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003640220661985958,
      "loss": 2.6847,
      "theoretical_loss": 3.677817013567333,
      "tokens_seen": 921744384
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.000364012036108325,
      "loss": 2.672,
      "theoretical_loss": 3.677791753095228,
      "tokens_seen": 921809920
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036400200601805415,
      "loss": 2.6619,
      "theoretical_loss": 3.677766494921759,
      "tokens_seen": 921875456
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003639919759277834,
      "loss": 2.6684,
      "theoretical_loss": 3.6777412390465543,
      "tokens_seen": 921940992
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003639819458375125,
      "loss": 2.7847,
      "theoretical_loss": 3.6777159854692396,
      "tokens_seen": 922006528
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036397191574724175,
      "loss": 2.6895,
      "theoretical_loss": 3.6776907341894445,
      "tokens_seen": 922072064
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036396188565697093,
      "loss": 2.7031,
      "theoretical_loss": 3.677665485206796,
      "tokens_seen": 922137600
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003639518555667001,
      "loss": 2.5351,
      "theoretical_loss": 3.6776402385209206,
      "tokens_seen": 922203136
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003639418254764293,
      "loss": 2.9291,
      "theoretical_loss": 3.6776149941314484,
      "tokens_seen": 922268672
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036393179538615853,
      "loss": 2.5709,
      "theoretical_loss": 3.6775897520380054,
      "tokens_seen": 922334208
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036392176529588766,
      "loss": 2.4427,
      "theoretical_loss": 3.6775645122402207,
      "tokens_seen": 922399744
    },
    {
      "epoch": 3.01,
      "objective/train/docs_used": 1097866,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.2819089889526367,
      "objective/train/theoretical_loss": 3.6775582026494273,
      "objective/train/tokens_used": 942876128,
      "theoretical_loss": 3.6775582026494273,
      "tokens_seen": 922416128
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003639117352056169,
      "loss": 2.811,
      "theoretical_loss": 3.6775392747377227,
      "tokens_seen": 922465280
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.000363901705115346,
      "loss": 2.6732,
      "theoretical_loss": 3.6775140395301387,
      "tokens_seen": 922530816
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036389167502507525,
      "loss": 2.7929,
      "theoretical_loss": 3.677488806617097,
      "tokens_seen": 922596352
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036388164493480444,
      "loss": 2.6178,
      "theoretical_loss": 3.6774635759982273,
      "tokens_seen": 922661888
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003638716148445336,
      "loss": 2.6879,
      "theoretical_loss": 3.677438347673157,
      "tokens_seen": 922727424
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003638615847542628,
      "loss": 2.5878,
      "theoretical_loss": 3.677413121641515,
      "tokens_seen": 922792960
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.000363851554663992,
      "loss": 2.6521,
      "theoretical_loss": 3.6773878979029293,
      "tokens_seen": 922858496
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036384152457372116,
      "loss": 2.6997,
      "theoretical_loss": 3.67736267645703,
      "tokens_seen": 922924032
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003638314944834504,
      "loss": 2.7684,
      "theoretical_loss": 3.6773374573034445,
      "tokens_seen": 922989568
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003638214643931795,
      "loss": 2.5416,
      "theoretical_loss": 3.6773122404418026,
      "tokens_seen": 923055104
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036381143430290876,
      "loss": 2.6587,
      "theoretical_loss": 3.677287025871733,
      "tokens_seen": 923120640
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003638014042126379,
      "loss": 2.8315,
      "theoretical_loss": 3.6772618135928647,
      "tokens_seen": 923186176
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003637913741223671,
      "loss": 2.598,
      "theoretical_loss": 3.6772366036048276,
      "tokens_seen": 923251712
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003637813440320963,
      "loss": 2.8329,
      "theoretical_loss": 3.67721139590725,
      "tokens_seen": 923317248
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003637713139418255,
      "loss": 2.7133,
      "theoretical_loss": 3.6771861904997616,
      "tokens_seen": 923382784
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036376128385155466,
      "loss": 2.5512,
      "theoretical_loss": 3.677160987381992,
      "tokens_seen": 923448320
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003637512537612839,
      "loss": 2.7946,
      "theoretical_loss": 3.6771357865535705,
      "tokens_seen": 923513856
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.000363741223671013,
      "loss": 2.6753,
      "theoretical_loss": 3.6771105880141266,
      "tokens_seen": 923579392
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036373119358074226,
      "loss": 2.7061,
      "theoretical_loss": 3.67708539176329,
      "tokens_seen": 923644928
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003637211634904714,
      "loss": 2.7782,
      "theoretical_loss": 3.6770601978006914,
      "tokens_seen": 923710464
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003637111334002006,
      "loss": 2.6796,
      "theoretical_loss": 3.6770350061259593,
      "tokens_seen": 923776000
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003637011033099298,
      "loss": 2.9432,
      "theoretical_loss": 3.6770098167387246,
      "tokens_seen": 923841536
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.000363691073219659,
      "loss": 2.8094,
      "theoretical_loss": 3.676984629638617,
      "tokens_seen": 923907072
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036368104312938817,
      "loss": 2.6513,
      "theoretical_loss": 3.6769594448252665,
      "tokens_seen": 923972608
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036367101303911735,
      "loss": 2.7836,
      "theoretical_loss": 3.6769342622983032,
      "tokens_seen": 924038144
    },
    {
      "epoch": 3.01,
      "objective/train/docs_used": 1099373,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.109006881713867,
      "objective/train/theoretical_loss": 3.676927967023768,
      "objective/train/tokens_used": 944514528,
      "theoretical_loss": 3.676927967023768,
      "tokens_seen": 924054528
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036366098294884653,
      "loss": 2.7884,
      "theoretical_loss": 3.6769090820573584,
      "tokens_seen": 924103680
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036365095285857576,
      "loss": 2.8461,
      "theoretical_loss": 3.6768839041020613,
      "tokens_seen": 924169216
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003636409227683049,
      "loss": 2.6621,
      "theoretical_loss": 3.676858728432043,
      "tokens_seen": 924234752
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003636308926780341,
      "loss": 2.8803,
      "theoretical_loss": 3.676833555046933,
      "tokens_seen": 924300288
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036362086258776325,
      "loss": 2.8337,
      "theoretical_loss": 3.6768083839463634,
      "tokens_seen": 924365824
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003636108324974925,
      "loss": 2.8779,
      "theoretical_loss": 3.6767832151299644,
      "tokens_seen": 924431360
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036360080240722167,
      "loss": 2.5029,
      "theoretical_loss": 3.6767580485973665,
      "tokens_seen": 924496896
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036359077231695085,
      "loss": 2.586,
      "theoretical_loss": 3.6767328843482012,
      "tokens_seen": 924562432
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036358074222668003,
      "loss": 2.6846,
      "theoretical_loss": 3.6767077223820985,
      "tokens_seen": 924627968
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036357071213640927,
      "loss": 2.8209,
      "theoretical_loss": 3.6766825626986908,
      "tokens_seen": 924693504
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003635606820461384,
      "loss": 2.9991,
      "theoretical_loss": 3.6766574052976075,
      "tokens_seen": 924759040
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036355065195586763,
      "loss": 2.7583,
      "theoretical_loss": 3.676632250178482,
      "tokens_seen": 924824576
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036354062186559676,
      "loss": 2.6988,
      "theoretical_loss": 3.676607097340944,
      "tokens_seen": 924890112
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.000363530591775326,
      "loss": 2.7356,
      "theoretical_loss": 3.676581946784625,
      "tokens_seen": 924955648
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036352056168505517,
      "loss": 2.8004,
      "theoretical_loss": 3.676556798509157,
      "tokens_seen": 925021184
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036351053159478435,
      "loss": 2.6135,
      "theoretical_loss": 3.676531652514172,
      "tokens_seen": 925086720
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036350050150451353,
      "loss": 2.7654,
      "theoretical_loss": 3.6765065087993007,
      "tokens_seen": 925152256
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003634904714142427,
      "loss": 2.7685,
      "theoretical_loss": 3.6764813673641754,
      "tokens_seen": 925217792
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003634804413239719,
      "loss": 2.6842,
      "theoretical_loss": 3.676456228208428,
      "tokens_seen": 925283328
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036347041123370113,
      "loss": 2.8569,
      "theoretical_loss": 3.67643109133169,
      "tokens_seen": 925348864
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036346038114343026,
      "loss": 2.7765,
      "theoretical_loss": 3.6764059567335936,
      "tokens_seen": 925414400
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003634503510531595,
      "loss": 2.703,
      "theoretical_loss": 3.6763808244137715,
      "tokens_seen": 925479936
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003634403209628886,
      "loss": 2.7041,
      "theoretical_loss": 3.676355694371855,
      "tokens_seen": 925545472
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036343029087261786,
      "loss": 2.7526,
      "theoretical_loss": 3.6763305666074766,
      "tokens_seen": 925611008
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036342026078234704,
      "loss": 2.7518,
      "theoretical_loss": 3.676305441120269,
      "tokens_seen": 925676544
    },
    {
      "epoch": 3.01,
      "objective/train/docs_used": 1100200,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.052302360534668,
      "objective/train/theoretical_loss": 3.6762991601042323,
      "objective/train/tokens_used": 946152928,
      "theoretical_loss": 3.6762991601042323,
      "tokens_seen": 925692928
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003634102306920762,
      "loss": 2.7231,
      "theoretical_loss": 3.6762803179098644,
      "tokens_seen": 925742080
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003634002006018054,
      "loss": 2.6954,
      "theoretical_loss": 3.6762551969758954,
      "tokens_seen": 925807616
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036339017051153464,
      "loss": 2.6857,
      "theoretical_loss": 3.676230078317994,
      "tokens_seen": 925873152
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036338014042126376,
      "loss": 2.7367,
      "theoretical_loss": 3.676204961935794,
      "tokens_seen": 925938688
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.000363370110330993,
      "loss": 2.6934,
      "theoretical_loss": 3.6761798478289274,
      "tokens_seen": 926004224
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003633600802407222,
      "loss": 2.6842,
      "theoretical_loss": 3.6761547359970272,
      "tokens_seen": 926069760
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036335005015045136,
      "loss": 2.7722,
      "theoretical_loss": 3.6761296264397263,
      "tokens_seen": 926135296
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003633400200601806,
      "loss": 2.6782,
      "theoretical_loss": 3.6761045191566586,
      "tokens_seen": 926200832
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003633299899699097,
      "loss": 2.9231,
      "theoretical_loss": 3.6760794141474555,
      "tokens_seen": 926266368
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036331995987963896,
      "loss": 2.7677,
      "theoretical_loss": 3.6760543114117517,
      "tokens_seen": 926331904
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003633099297893681,
      "loss": 2.7595,
      "theoretical_loss": 3.6760292109491797,
      "tokens_seen": 926397440
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003632998996990973,
      "loss": 2.8369,
      "theoretical_loss": 3.676004112759373,
      "tokens_seen": 926462976
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003632898696088265,
      "loss": 2.7427,
      "theoretical_loss": 3.675979016841966,
      "tokens_seen": 926528512
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003632798395185557,
      "loss": 2.7186,
      "theoretical_loss": 3.675953923196591,
      "tokens_seen": 926594048
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036326980942828486,
      "loss": 2.7138,
      "theoretical_loss": 3.6759288318228815,
      "tokens_seen": 926659584
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003632597793380141,
      "loss": 2.7673,
      "theoretical_loss": 3.6759037427204726,
      "tokens_seen": 926725120
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003632497492477432,
      "loss": 2.6312,
      "theoretical_loss": 3.675878655888997,
      "tokens_seen": 926790656
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036323971915747246,
      "loss": 2.6802,
      "theoretical_loss": 3.675853571328089,
      "tokens_seen": 926856192
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003632296890672016,
      "loss": 2.9694,
      "theoretical_loss": 3.6758284890373822,
      "tokens_seen": 926921728
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003632196589769308,
      "loss": 2.7553,
      "theoretical_loss": 3.675803409016511,
      "tokens_seen": 926987264
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036320962888666,
      "loss": 2.7427,
      "theoretical_loss": 3.6757783312651093,
      "tokens_seen": 927052800
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003631995987963892,
      "loss": 2.7938,
      "theoretical_loss": 3.6757532557828116,
      "tokens_seen": 927118336
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036318956870611837,
      "loss": 2.8564,
      "theoretical_loss": 3.6757281825692525,
      "tokens_seen": 927183872
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036317953861584755,
      "loss": 2.817,
      "theoretical_loss": 3.675703111624065,
      "tokens_seen": 927249408
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036316950852557673,
      "loss": 2.6925,
      "theoretical_loss": 3.6756780429468856,
      "tokens_seen": 927314944
    },
    {
      "epoch": 3.01,
      "objective/train/docs_used": 1101371,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7425971031188965,
      "objective/train/theoretical_loss": 3.6756717761319235,
      "objective/train/tokens_used": 947791328,
      "theoretical_loss": 3.6756717761319235,
      "tokens_seen": 927331328
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036315947843530596,
      "loss": 2.738,
      "theoretical_loss": 3.6756529765373473,
      "tokens_seen": 927380480
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003631494483450351,
      "loss": 2.9168,
      "theoretical_loss": 3.6756279123950852,
      "tokens_seen": 927446016
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003631394182547643,
      "loss": 2.8124,
      "theoretical_loss": 3.675602850519734,
      "tokens_seen": 927511552
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036312938816449345,
      "loss": 2.9051,
      "theoretical_loss": 3.6755777909109284,
      "tokens_seen": 927577088
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003631193580742227,
      "loss": 2.7232,
      "theoretical_loss": 3.675552733568304,
      "tokens_seen": 927642624
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036310932798395187,
      "loss": 2.5916,
      "theoretical_loss": 3.675527678491495,
      "tokens_seen": 927708160
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036309929789368105,
      "loss": 2.794,
      "theoretical_loss": 3.675502625680137,
      "tokens_seen": 927773696
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036308926780341023,
      "loss": 2.7731,
      "theoretical_loss": 3.675477575133865,
      "tokens_seen": 927839232
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036307923771313947,
      "loss": 2.7715,
      "theoretical_loss": 3.6754525268523137,
      "tokens_seen": 927904768
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003630692076228686,
      "loss": 2.779,
      "theoretical_loss": 3.6754274808351193,
      "tokens_seen": 927970304
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036305917753259783,
      "loss": 2.7027,
      "theoretical_loss": 3.6754024370819165,
      "tokens_seen": 928035840
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036304914744232696,
      "loss": 2.8426,
      "theoretical_loss": 3.6753773955923412,
      "tokens_seen": 928101376
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003630391173520562,
      "loss": 2.7696,
      "theoretical_loss": 3.6753523563660293,
      "tokens_seen": 928166912
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036302908726178537,
      "loss": 2.7511,
      "theoretical_loss": 3.6753273194026157,
      "tokens_seen": 928232448
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036301905717151455,
      "loss": 2.5918,
      "theoretical_loss": 3.6753022847017363,
      "tokens_seen": 928297984
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036300902708124374,
      "loss": 2.5532,
      "theoretical_loss": 3.675277252263027,
      "tokens_seen": 928363520
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003629989969909729,
      "loss": 2.7777,
      "theoretical_loss": 3.675252222086124,
      "tokens_seen": 928429056
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003629889669007021,
      "loss": 2.6777,
      "theoretical_loss": 3.675227194170663,
      "tokens_seen": 928494592
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036297893681043133,
      "loss": 2.8747,
      "theoretical_loss": 3.6752021685162806,
      "tokens_seen": 928560128
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036296890672016046,
      "loss": 2.7622,
      "theoretical_loss": 3.6751771451226123,
      "tokens_seen": 928625664
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003629588766298897,
      "loss": 2.5471,
      "theoretical_loss": 3.6751521239892946,
      "tokens_seen": 928691200
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003629488465396188,
      "loss": 2.5757,
      "theoretical_loss": 3.6751271051159637,
      "tokens_seen": 928756736
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036293881644934806,
      "loss": 2.7508,
      "theoretical_loss": 3.675102088502257,
      "tokens_seen": 928822272
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036292878635907724,
      "loss": 2.8222,
      "theoretical_loss": 3.675077074147809,
      "tokens_seen": 928887808
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003629187562688064,
      "loss": 2.7448,
      "theoretical_loss": 3.6750520620522584,
      "tokens_seen": 928953344
    },
    {
      "epoch": 3.01,
      "objective/train/docs_used": 1102157,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.89856219291687,
      "objective/train/theoretical_loss": 3.67504580938128,
      "objective/train/tokens_used": 949429728,
      "theoretical_loss": 3.67504580938128,
      "tokens_seen": 928969728
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003629087261785356,
      "loss": 2.7577,
      "theoretical_loss": 3.67502705221524,
      "tokens_seen": 929018880
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036289869608826484,
      "loss": 2.764,
      "theoretical_loss": 3.6750020446363925,
      "tokens_seen": 929084416
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036288866599799396,
      "loss": 2.5949,
      "theoretical_loss": 3.674977039315351,
      "tokens_seen": 929149952
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003628786359077232,
      "loss": 2.7679,
      "theoretical_loss": 3.674952036251754,
      "tokens_seen": 929215488
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003628686058174523,
      "loss": 2.8058,
      "theoretical_loss": 3.6749270354452372,
      "tokens_seen": 929281024
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036285857572718156,
      "loss": 2.5555,
      "theoretical_loss": 3.674902036895438,
      "tokens_seen": 929346560
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036284854563691074,
      "loss": 2.6937,
      "theoretical_loss": 3.6748770406019937,
      "tokens_seen": 929412096
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003628385155466399,
      "loss": 2.813,
      "theoretical_loss": 3.674852046564542,
      "tokens_seen": 929477632
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003628284854563691,
      "loss": 2.573,
      "theoretical_loss": 3.6748270547827198,
      "tokens_seen": 929543168
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003628184553660983,
      "loss": 2.8186,
      "theoretical_loss": 3.6748020652561646,
      "tokens_seen": 929608704
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036280842527582747,
      "loss": 2.6764,
      "theoretical_loss": 3.6747770779845137,
      "tokens_seen": 929674240
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003627983951855567,
      "loss": 2.6099,
      "theoretical_loss": 3.674752092967405,
      "tokens_seen": 929739776
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036278836509528583,
      "loss": 2.979,
      "theoretical_loss": 3.6747271102044756,
      "tokens_seen": 929805312
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036277833500501506,
      "loss": 2.8215,
      "theoretical_loss": 3.6747021296953646,
      "tokens_seen": 929870848
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036276830491474424,
      "loss": 2.8194,
      "theoretical_loss": 3.6746771514397083,
      "tokens_seen": 929936384
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003627582748244734,
      "loss": 2.6955,
      "theoretical_loss": 3.6746521754371457,
      "tokens_seen": 930001920
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003627482447342026,
      "loss": 2.886,
      "theoretical_loss": 3.674627201687314,
      "tokens_seen": 930067456
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003627382146439318,
      "loss": 2.6595,
      "theoretical_loss": 3.674602230189852,
      "tokens_seen": 930132992
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036272818455366097,
      "loss": 2.8569,
      "theoretical_loss": 3.674577260944398,
      "tokens_seen": 930198528
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003627181544633902,
      "loss": 2.7004,
      "theoretical_loss": 3.6745522939505895,
      "tokens_seen": 930264064
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036270812437311933,
      "loss": 2.537,
      "theoretical_loss": 3.674527329208065,
      "tokens_seen": 930329600
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036269809428284857,
      "loss": 2.6057,
      "theoretical_loss": 3.674502366716464,
      "tokens_seen": 930395136
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003626880641925777,
      "loss": 2.6501,
      "theoretical_loss": 3.6744774064754235,
      "tokens_seen": 930460672
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036267803410230693,
      "loss": 2.7078,
      "theoretical_loss": 3.6744524484845824,
      "tokens_seen": 930526208
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003626680040120361,
      "loss": 2.6438,
      "theoretical_loss": 3.6744274927435807,
      "tokens_seen": 930591744
    },
    {
      "epoch": 3.01,
      "objective/train/docs_used": 1103642,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.0793814659118652,
      "objective/train/theoretical_loss": 3.6744212541598245,
      "objective/train/tokens_used": 951068128,
      "theoretical_loss": 3.6744212541598245,
      "tokens_seen": 930608128
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003626579739217653,
      "loss": 2.5507,
      "theoretical_loss": 3.6744025392520556,
      "tokens_seen": 930657280
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036264794383149447,
      "loss": 2.7309,
      "theoretical_loss": 3.674377588009647,
      "tokens_seen": 930722816
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036263791374122365,
      "loss": 2.6374,
      "theoretical_loss": 3.674352639015993,
      "tokens_seen": 930788352
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036262788365095283,
      "loss": 2.8798,
      "theoretical_loss": 3.674327692270733,
      "tokens_seen": 930853888
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036261785356068207,
      "loss": 2.7256,
      "theoretical_loss": 3.674302747773506,
      "tokens_seen": 930919424
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036260782347041125,
      "loss": 2.8085,
      "theoretical_loss": 3.674277805523952,
      "tokens_seen": 930984960
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036259779338014043,
      "loss": 2.7442,
      "theoretical_loss": 3.6742528655217095,
      "tokens_seen": 931050496
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036258776328986967,
      "loss": 2.6867,
      "theoretical_loss": 3.6742279277664176,
      "tokens_seen": 931116032
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003625777331995988,
      "loss": 2.8555,
      "theoretical_loss": 3.6742029922577157,
      "tokens_seen": 931181568
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036256770310932803,
      "loss": 2.6521,
      "theoretical_loss": 3.674178058995244,
      "tokens_seen": 931247104
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036255767301905716,
      "loss": 2.7883,
      "theoretical_loss": 3.6741531279786424,
      "tokens_seen": 931312640
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003625476429287864,
      "loss": 2.8195,
      "theoretical_loss": 3.674128199207549,
      "tokens_seen": 931378176
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003625376128385156,
      "loss": 2.6329,
      "theoretical_loss": 3.674103272681605,
      "tokens_seen": 931443712
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036252758274824475,
      "loss": 2.8075,
      "theoretical_loss": 3.67407834840045,
      "tokens_seen": 931509248
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036251755265797394,
      "loss": 2.5847,
      "theoretical_loss": 3.6740534263637237,
      "tokens_seen": 931574784
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003625075225677031,
      "loss": 2.7421,
      "theoretical_loss": 3.674028506571066,
      "tokens_seen": 931640320
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003624974924774323,
      "loss": 2.9755,
      "theoretical_loss": 3.6740035890221168,
      "tokens_seen": 931705856
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036248746238716153,
      "loss": 2.8285,
      "theoretical_loss": 3.673978673716517,
      "tokens_seen": 931771392
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036247743229689066,
      "loss": 2.9064,
      "theoretical_loss": 3.673953760653906,
      "tokens_seen": 931836928
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003624674022066199,
      "loss": 2.7024,
      "theoretical_loss": 3.673928849833925,
      "tokens_seen": 931902464
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.000362457372116349,
      "loss": 2.83,
      "theoretical_loss": 3.6739039412562144,
      "tokens_seen": 931968000
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036244734202607826,
      "loss": 2.6702,
      "theoretical_loss": 3.6738790349204136,
      "tokens_seen": 932033536
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036243731193580744,
      "loss": 2.7042,
      "theoretical_loss": 3.6738541308261645,
      "tokens_seen": 932099072
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003624272818455366,
      "loss": 2.8335,
      "theoretical_loss": 3.673829228973107,
      "tokens_seen": 932164608
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003624172517552658,
      "loss": 2.7473,
      "theoretical_loss": 3.673804329360882,
      "tokens_seen": 932230144
    },
    {
      "epoch": 3.01,
      "objective/train/docs_used": 1104463,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.2996299266815186,
      "objective/train/theoretical_loss": 3.673798104807914,
      "objective/train/tokens_used": 952706528,
      "theoretical_loss": 3.673798104807914,
      "tokens_seen": 932246528
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036240722166499504,
      "loss": 2.9647,
      "theoretical_loss": 3.6737794319891313,
      "tokens_seen": 932295680
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036239719157472416,
      "loss": 2.7751,
      "theoretical_loss": 3.673754536857494,
      "tokens_seen": 932361216
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003623871614844534,
      "loss": 2.8086,
      "theoretical_loss": 3.673729643965612,
      "tokens_seen": 932426752
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003623771313941825,
      "loss": 2.5253,
      "theoretical_loss": 3.6737047533131273,
      "tokens_seen": 932492288
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036236710130391176,
      "loss": 2.8928,
      "theoretical_loss": 3.67367986489968,
      "tokens_seen": 932557824
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036235707121364094,
      "loss": 2.7164,
      "theoretical_loss": 3.6736549787249118,
      "tokens_seen": 932623360
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003623470411233701,
      "loss": 2.8051,
      "theoretical_loss": 3.6736300947884635,
      "tokens_seen": 932688896
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003623370110330993,
      "loss": 2.6297,
      "theoretical_loss": 3.673605213089977,
      "tokens_seen": 932754432
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003623269809428285,
      "loss": 2.6382,
      "theoretical_loss": 3.6735803336290935,
      "tokens_seen": 932819968
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036231695085255767,
      "loss": 2.8193,
      "theoretical_loss": 3.673555456405455,
      "tokens_seen": 932885504
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003623069207622869,
      "loss": 2.6228,
      "theoretical_loss": 3.673530581418703,
      "tokens_seen": 932951040
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036229689067201603,
      "loss": 2.6079,
      "theoretical_loss": 3.673505708668479,
      "tokens_seen": 933016576
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036228686058174526,
      "loss": 2.7311,
      "theoretical_loss": 3.6734808381544255,
      "tokens_seen": 933082112
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036227683049147444,
      "loss": 2.6853,
      "theoretical_loss": 3.6734559698761835,
      "tokens_seen": 933147648
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003622668004012036,
      "loss": 2.6714,
      "theoretical_loss": 3.673431103833396,
      "tokens_seen": 933213184
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003622567703109328,
      "loss": 2.8825,
      "theoretical_loss": 3.673406240025704,
      "tokens_seen": 933278720
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.000362246740220662,
      "loss": 2.5792,
      "theoretical_loss": 3.6733813784527505,
      "tokens_seen": 933344256
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036223671013039117,
      "loss": 2.8156,
      "theoretical_loss": 3.6733565191141775,
      "tokens_seen": 933409792
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003622266800401204,
      "loss": 2.9049,
      "theoretical_loss": 3.673331662009627,
      "tokens_seen": 933475328
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036221664994984953,
      "loss": 2.9021,
      "theoretical_loss": 3.673306807138742,
      "tokens_seen": 933540864
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036220661985957877,
      "loss": 2.8239,
      "theoretical_loss": 3.6732819545011646,
      "tokens_seen": 933606400
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003621965897693079,
      "loss": 2.6474,
      "theoretical_loss": 3.6732571040965376,
      "tokens_seen": 933671936
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036218655967903713,
      "loss": 2.8546,
      "theoretical_loss": 3.6732322559245034,
      "tokens_seen": 933737472
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003621765295887663,
      "loss": 2.6471,
      "theoretical_loss": 3.6732074099847054,
      "tokens_seen": 933803008
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003621664994984955,
      "loss": 2.714,
      "theoretical_loss": 3.673182566276785,
      "tokens_seen": 933868544
    },
    {
      "epoch": 3.01,
      "objective/train/docs_used": 1105263,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4603049755096436,
      "objective/train/theoretical_loss": 3.6731763556984944,
      "objective/train/tokens_used": 954344928,
      "theoretical_loss": 3.6731763556984944,
      "tokens_seen": 933884928
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036215646940822467,
      "loss": 2.6152,
      "theoretical_loss": 3.6731577248003866,
      "tokens_seen": 933934080
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036214643931795385,
      "loss": 2.7085,
      "theoretical_loss": 3.6731328855551526,
      "tokens_seen": 933999616
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036213640922768303,
      "loss": 2.6834,
      "theoretical_loss": 3.6731080485407266,
      "tokens_seen": 934065152
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036212637913741227,
      "loss": 2.7269,
      "theoretical_loss": 3.6730832137567506,
      "tokens_seen": 934130688
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003621163490471414,
      "loss": 2.749,
      "theoretical_loss": 3.6730583812028685,
      "tokens_seen": 934196224
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036210631895687063,
      "loss": 2.7608,
      "theoretical_loss": 3.673033550878724,
      "tokens_seen": 934261760
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003620962888665998,
      "loss": 2.9078,
      "theoretical_loss": 3.67300872278396,
      "tokens_seen": 934327296
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.000362086258776329,
      "loss": 2.8538,
      "theoretical_loss": 3.67298389691822,
      "tokens_seen": 934392832
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003620762286860582,
      "loss": 2.6955,
      "theoretical_loss": 3.6729590732811475,
      "tokens_seen": 934458368
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036206619859578736,
      "loss": 2.7588,
      "theoretical_loss": 3.6729342518723866,
      "tokens_seen": 934523904
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036205616850551654,
      "loss": 2.8656,
      "theoretical_loss": 3.67290943269158,
      "tokens_seen": 934589440
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003620461384152458,
      "loss": 2.8229,
      "theoretical_loss": 3.672884615738373,
      "tokens_seen": 934654976
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003620361083249749,
      "loss": 2.8307,
      "theoretical_loss": 3.6728598010124083,
      "tokens_seen": 934720512
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036202607823470414,
      "loss": 2.7492,
      "theoretical_loss": 3.6728349885133307,
      "tokens_seen": 934786048
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036201604814443326,
      "loss": 2.6536,
      "theoretical_loss": 3.672810178240784,
      "tokens_seen": 934851584
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003620060180541625,
      "loss": 2.8426,
      "theoretical_loss": 3.6727853701944118,
      "tokens_seen": 934917120
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003619959879638917,
      "loss": 2.745,
      "theoretical_loss": 3.672760564373859,
      "tokens_seen": 934982656
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036198595787362086,
      "loss": 2.6526,
      "theoretical_loss": 3.672735760778769,
      "tokens_seen": 935048192
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036197592778335004,
      "loss": 2.4051,
      "theoretical_loss": 3.6727109594087874,
      "tokens_seen": 935113728
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003619658976930792,
      "loss": 2.4586,
      "theoretical_loss": 3.672686160263558,
      "tokens_seen": 935179264
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003619558676028084,
      "loss": 2.8034,
      "theoretical_loss": 3.6726613633427254,
      "tokens_seen": 935244800
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036194583751253764,
      "loss": 2.8574,
      "theoretical_loss": 3.6726365686459346,
      "tokens_seen": 935310336
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036193580742226677,
      "loss": 3.0104,
      "theoretical_loss": 3.6726117761728294,
      "tokens_seen": 935375872
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.000361925777331996,
      "loss": 2.7326,
      "theoretical_loss": 3.672586985923055,
      "tokens_seen": 935441408
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003619157472417252,
      "loss": 2.5144,
      "theoretical_loss": 3.672562197896257,
      "tokens_seen": 935506944
    },
    {
      "epoch": 3.01,
      "objective/train/docs_used": 1106660,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.0614864826202393,
      "objective/train/theoretical_loss": 3.6725560012368557,
      "objective/train/tokens_used": 955983328,
      "theoretical_loss": 3.6725560012368557,
      "tokens_seen": 935523328
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036190571715145436,
      "loss": 2.7475,
      "theoretical_loss": 3.6725374120920797,
      "tokens_seen": 935572480
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036189568706118354,
      "loss": 2.5965,
      "theoretical_loss": 3.672512628510168,
      "tokens_seen": 935638016
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003618856569709127,
      "loss": 2.6269,
      "theoretical_loss": 3.672487847150167,
      "tokens_seen": 935703552
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003618756268806419,
      "loss": 2.6792,
      "theoretical_loss": 3.672463068011723,
      "tokens_seen": 935769088
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036186559679037114,
      "loss": 2.7316,
      "theoretical_loss": 3.6724382910944797,
      "tokens_seen": 935834624
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003618555667001003,
      "loss": 2.7486,
      "theoretical_loss": 3.672413516398083,
      "tokens_seen": 935900160
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003618455366098295,
      "loss": 2.6199,
      "theoretical_loss": 3.672388743922179,
      "tokens_seen": 935965696
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003618355065195587,
      "loss": 2.6766,
      "theoretical_loss": 3.672363973666413,
      "tokens_seen": 936031232
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036182547642928787,
      "loss": 2.7611,
      "theoretical_loss": 3.67233920563043,
      "tokens_seen": 936096768
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003618154463390171,
      "loss": 2.4793,
      "theoretical_loss": 3.672314439813876,
      "tokens_seen": 936162304
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036180541624874623,
      "loss": 2.5873,
      "theoretical_loss": 3.672289676216397,
      "tokens_seen": 936227840
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036179538615847546,
      "loss": 2.6239,
      "theoretical_loss": 3.6722649148376383,
      "tokens_seen": 936293376
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036178535606820464,
      "loss": 2.7087,
      "theoretical_loss": 3.6722401556772466,
      "tokens_seen": 936358912
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003617753259779338,
      "loss": 2.7876,
      "theoretical_loss": 3.672215398734867,
      "tokens_seen": 936424448
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.000361765295887663,
      "loss": 2.6702,
      "theoretical_loss": 3.672190644010147,
      "tokens_seen": 936489984
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003617552657973922,
      "loss": 2.788,
      "theoretical_loss": 3.672165891502731,
      "tokens_seen": 936555520
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036174523570712137,
      "loss": 2.7138,
      "theoretical_loss": 3.6721411412122666,
      "tokens_seen": 936621056
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003617352056168506,
      "loss": 2.7349,
      "theoretical_loss": 3.6721163931383995,
      "tokens_seen": 936686592
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036172517552657973,
      "loss": 2.7485,
      "theoretical_loss": 3.6720916472807765,
      "tokens_seen": 936752128
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036171514543630897,
      "loss": 2.8504,
      "theoretical_loss": 3.6720669036390436,
      "tokens_seen": 936817664
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003617051153460381,
      "loss": 2.7481,
      "theoretical_loss": 3.672042162212848,
      "tokens_seen": 936883200
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036169508525576733,
      "loss": 2.9067,
      "theoretical_loss": 3.672017423001836,
      "tokens_seen": 936948736
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003616850551654965,
      "loss": 2.4963,
      "theoretical_loss": 3.6719926860056544,
      "tokens_seen": 937014272
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003616750250752257,
      "loss": 2.6033,
      "theoretical_loss": 3.67196795122395,
      "tokens_seen": 937079808
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036166499498495487,
      "loss": 2.7483,
      "theoretical_loss": 3.6719432186563696,
      "tokens_seen": 937145344
    },
    {
      "epoch": 3.01,
      "objective/train/docs_used": 1107254,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6422669887542725,
      "objective/train/theoretical_loss": 3.67193703586039,
      "objective/train/tokens_used": 957621728,
      "theoretical_loss": 3.67193703586039,
      "tokens_seen": 937161728
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036165496489468405,
      "loss": 2.6973,
      "theoretical_loss": 3.67191848830256,
      "tokens_seen": 937210880
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036164493480441323,
      "loss": 2.6796,
      "theoretical_loss": 3.6718937601621695,
      "tokens_seen": 937276416
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036163490471414247,
      "loss": 2.6766,
      "theoretical_loss": 3.6718690342348435,
      "tokens_seen": 937341952
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003616248746238716,
      "loss": 2.8178,
      "theoretical_loss": 3.6718443105202305,
      "tokens_seen": 937407488
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036161484453360083,
      "loss": 2.6341,
      "theoretical_loss": 3.6718195890179772,
      "tokens_seen": 937473024
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036160481444333,
      "loss": 2.8366,
      "theoretical_loss": 3.671794869727731,
      "tokens_seen": 937538560
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003615947843530592,
      "loss": 2.6977,
      "theoretical_loss": 3.67177015264914,
      "tokens_seen": 937604096
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003615847542627884,
      "loss": 2.6461,
      "theoretical_loss": 3.6717454377818513,
      "tokens_seen": 937669632
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036157472417251756,
      "loss": 2.6455,
      "theoretical_loss": 3.671720725125512,
      "tokens_seen": 937735168
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036156469408224674,
      "loss": 2.7254,
      "theoretical_loss": 3.671696014679771,
      "tokens_seen": 937800704
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.000361554663991976,
      "loss": 2.6941,
      "theoretical_loss": 3.6716713064442743,
      "tokens_seen": 937866240
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003615446339017051,
      "loss": 2.7385,
      "theoretical_loss": 3.671646600418672,
      "tokens_seen": 937931776
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036153460381143434,
      "loss": 2.5716,
      "theoretical_loss": 3.671621896602611,
      "tokens_seen": 937997312
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036152457372116346,
      "loss": 2.5538,
      "theoretical_loss": 3.671597194995739,
      "tokens_seen": 938062848
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003615145436308927,
      "loss": 2.7646,
      "theoretical_loss": 3.671572495597704,
      "tokens_seen": 938128384
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003615045135406219,
      "loss": 2.5636,
      "theoretical_loss": 3.671547798408155,
      "tokens_seen": 938193920
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036149448345035106,
      "loss": 2.6606,
      "theoretical_loss": 3.6715231034267397,
      "tokens_seen": 938259456
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036148445336008024,
      "loss": 2.7052,
      "theoretical_loss": 3.671498410653107,
      "tokens_seen": 938324992
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003614744232698094,
      "loss": 2.7681,
      "theoretical_loss": 3.6714737200869045,
      "tokens_seen": 938390528
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003614643931795386,
      "loss": 2.3929,
      "theoretical_loss": 3.6714490317277817,
      "tokens_seen": 938456064
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036145436308926784,
      "loss": 2.6983,
      "theoretical_loss": 3.6714243455753865,
      "tokens_seen": 938521600
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036144433299899697,
      "loss": 2.6932,
      "theoretical_loss": 3.6713996616293683,
      "tokens_seen": 938587136
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003614343029087262,
      "loss": 2.5358,
      "theoretical_loss": 3.671374979889375,
      "tokens_seen": 938652672
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003614242728184554,
      "loss": 2.583,
      "theoretical_loss": 3.671350300355056,
      "tokens_seen": 938718208
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036141424272818456,
      "loss": 2.7081,
      "theoretical_loss": 3.6713256230260596,
      "tokens_seen": 938783744
    },
    {
      "epoch": 3.01,
      "objective/train/docs_used": 1108649,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.576571464538574,
      "objective/train/theoretical_loss": 3.671319454038351,
      "objective/train/tokens_used": 959260128,
      "theoretical_loss": 3.671319454038351,
      "tokens_seen": 938800128
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036140421263791374,
      "loss": 2.7355,
      "theoretical_loss": 3.6713009479020355,
      "tokens_seen": 938849280
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003613941825476429,
      "loss": 2.7932,
      "theoretical_loss": 3.6712762749826324,
      "tokens_seen": 938914816
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003613841524573721,
      "loss": 2.659,
      "theoretical_loss": 3.6712516042675,
      "tokens_seen": 938980352
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036137412236710134,
      "loss": 2.4067,
      "theoretical_loss": 3.6712269357562866,
      "tokens_seen": 939045888
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036136409227683047,
      "loss": 2.6325,
      "theoretical_loss": 3.671202269448642,
      "tokens_seen": 939111424
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003613540621865597,
      "loss": 2.7483,
      "theoretical_loss": 3.671177605344216,
      "tokens_seen": 939176960
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036134403209628883,
      "loss": 2.6851,
      "theoretical_loss": 3.671152943442658,
      "tokens_seen": 939242496
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036133400200601807,
      "loss": 2.7459,
      "theoretical_loss": 3.6711282837436174,
      "tokens_seen": 939308032
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036132397191574725,
      "loss": 2.6561,
      "theoretical_loss": 3.671103626246743,
      "tokens_seen": 939373568
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036131394182547643,
      "loss": 2.8424,
      "theoretical_loss": 3.671078970951686,
      "tokens_seen": 939439104
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003613039117352056,
      "loss": 2.7681,
      "theoretical_loss": 3.671054317858095,
      "tokens_seen": 939504640
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036129388164493484,
      "loss": 2.7278,
      "theoretical_loss": 3.6710296669656204,
      "tokens_seen": 939570176
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036128385155466397,
      "loss": 2.9318,
      "theoretical_loss": 3.6710050182739127,
      "tokens_seen": 939635712
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0003612738214643932,
      "loss": 2.843,
      "theoretical_loss": 3.6709803717826213,
      "tokens_seen": 939701248
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.00036126379137412233,
      "loss": 2.7645,
      "theoretical_loss": 3.670955727491396,
      "tokens_seen": 939766784
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00036125376128385157,
      "loss": 2.7355,
      "theoretical_loss": 3.6709310853998876,
      "tokens_seen": 939832320
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00036124373119358075,
      "loss": 2.8614,
      "theoretical_loss": 3.6709064455077467,
      "tokens_seen": 939897856
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00036123370110330993,
      "loss": 2.78,
      "theoretical_loss": 3.670881807814623,
      "tokens_seen": 939963392
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003612236710130391,
      "loss": 2.6633,
      "theoretical_loss": 3.6708571723201664,
      "tokens_seen": 940028928
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003612136409227683,
      "loss": 2.7481,
      "theoretical_loss": 3.670832539024029,
      "tokens_seen": 940094464
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003612036108324975,
      "loss": 2.7512,
      "theoretical_loss": 3.6708079079258598,
      "tokens_seen": 940160000
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003611935807422267,
      "loss": 2.5802,
      "theoretical_loss": 3.6707832790253105,
      "tokens_seen": 940225536
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00036118355065195584,
      "loss": 2.8009,
      "theoretical_loss": 3.670758652322032,
      "tokens_seen": 940291072
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00036117352056168507,
      "loss": 2.8288,
      "theoretical_loss": 3.6707340278156746,
      "tokens_seen": 940356608
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003611634904714142,
      "loss": 2.9308,
      "theoretical_loss": 3.6707094055058898,
      "tokens_seen": 940422144
    },
    {
      "epoch": 3.02,
      "objective/train/docs_used": 1109497,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9072909355163574,
      "objective/train/theoretical_loss": 3.670703250271617,
      "objective/train/tokens_used": 960898528,
      "theoretical_loss": 3.670703250271617,
      "tokens_seen": 940438528
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00036115346038114343,
      "loss": 2.8859,
      "theoretical_loss": 3.6706847853923277,
      "tokens_seen": 940487680
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003611434302908726,
      "loss": 2.8009,
      "theoretical_loss": 3.6706601674746397,
      "tokens_seen": 940553216
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003611334002006018,
      "loss": 2.7237,
      "theoretical_loss": 3.6706355517524774,
      "tokens_seen": 940618752
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.000361123370110331,
      "loss": 2.5759,
      "theoretical_loss": 3.670610938225492,
      "tokens_seen": 940684288
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003611133400200602,
      "loss": 2.6999,
      "theoretical_loss": 3.6705863268933347,
      "tokens_seen": 940749824
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003611033099297894,
      "loss": 2.6335,
      "theoretical_loss": 3.6705617177556564,
      "tokens_seen": 940815360
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003610932798395186,
      "loss": 2.9326,
      "theoretical_loss": 3.670537110812109,
      "tokens_seen": 940880896
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00036108324974924776,
      "loss": 2.7079,
      "theoretical_loss": 3.6705125060623445,
      "tokens_seen": 940946432
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00036107321965897694,
      "loss": 2.8174,
      "theoretical_loss": 3.670487903506014,
      "tokens_seen": 941011968
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003610631895687062,
      "loss": 2.6992,
      "theoretical_loss": 3.6704633031427694,
      "tokens_seen": 941077504
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003610531594784353,
      "loss": 2.5846,
      "theoretical_loss": 3.6704387049722618,
      "tokens_seen": 941143040
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00036104312938816454,
      "loss": 2.5115,
      "theoretical_loss": 3.6704141089941444,
      "tokens_seen": 941208576
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00036103309929789366,
      "loss": 2.8338,
      "theoretical_loss": 3.6703895152080683,
      "tokens_seen": 941274112
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003610230692076229,
      "loss": 2.7403,
      "theoretical_loss": 3.670364923613686,
      "tokens_seen": 941339648
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003610130391173521,
      "loss": 2.7208,
      "theoretical_loss": 3.670340334210649,
      "tokens_seen": 941405184
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00036100300902708126,
      "loss": 2.8817,
      "theoretical_loss": 3.67031574699861,
      "tokens_seen": 941470720
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00036099297893681044,
      "loss": 2.4179,
      "theoretical_loss": 3.6702911619772207,
      "tokens_seen": 941536256
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003609829488465396,
      "loss": 2.5508,
      "theoretical_loss": 3.6702665791461344,
      "tokens_seen": 941601792
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003609729187562688,
      "loss": 2.8291,
      "theoretical_loss": 3.670241998505003,
      "tokens_seen": 941667328
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00036096288866599804,
      "loss": 2.8138,
      "theoretical_loss": 3.670217420053479,
      "tokens_seen": 941732864
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00036095285857572717,
      "loss": 2.5917,
      "theoretical_loss": 3.670192843791215,
      "tokens_seen": 941798400
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003609428284854564,
      "loss": 2.6289,
      "theoretical_loss": 3.6701682697178635,
      "tokens_seen": 941863936
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003609327983951856,
      "loss": 2.6654,
      "theoretical_loss": 3.670143697833078,
      "tokens_seen": 941929472
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00036092276830491476,
      "loss": 2.7178,
      "theoretical_loss": 3.67011912813651,
      "tokens_seen": 941995008
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00036091273821464394,
      "loss": 2.6515,
      "theoretical_loss": 3.670094560627814,
      "tokens_seen": 942060544
    },
    {
      "epoch": 3.02,
      "objective/train/docs_used": 1110990,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.0038723945617676,
      "objective/train/theoretical_loss": 3.6700884190924543,
      "objective/train/tokens_used": 962536928,
      "theoretical_loss": 3.6700884190924543,
      "tokens_seen": 942076928
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003609027081243731,
      "loss": 2.5359,
      "theoretical_loss": 3.6700699953066422,
      "tokens_seen": 942126080
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003608926780341023,
      "loss": 2.7436,
      "theoretical_loss": 3.670045432172647,
      "tokens_seen": 942191616
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00036088264794383154,
      "loss": 2.8468,
      "theoretical_loss": 3.670020871225483,
      "tokens_seen": 942257152
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00036087261785356067,
      "loss": 2.7068,
      "theoretical_loss": 3.669996312464802,
      "tokens_seen": 942322688
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003608625877632899,
      "loss": 2.7132,
      "theoretical_loss": 3.6699717558902583,
      "tokens_seen": 942388224
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00036085255767301903,
      "loss": 2.9056,
      "theoretical_loss": 3.669947201501505,
      "tokens_seen": 942453760
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00036084252758274827,
      "loss": 2.6841,
      "theoretical_loss": 3.6699226492981953,
      "tokens_seen": 942519296
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00036083249749247745,
      "loss": 2.6888,
      "theoretical_loss": 3.6698980992799832,
      "tokens_seen": 942584832
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00036082246740220663,
      "loss": 2.6369,
      "theoretical_loss": 3.669873551446522,
      "tokens_seen": 942650368
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003608124373119358,
      "loss": 2.7913,
      "theoretical_loss": 3.6698490057974658,
      "tokens_seen": 942715904
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00036080240722166505,
      "loss": 2.7292,
      "theoretical_loss": 3.6698244623324676,
      "tokens_seen": 942781440
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00036079237713139417,
      "loss": 2.8,
      "theoretical_loss": 3.6697999210511822,
      "tokens_seen": 942846976
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003607823470411234,
      "loss": 2.492,
      "theoretical_loss": 3.6697753819532624,
      "tokens_seen": 942912512
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00036077231695085253,
      "loss": 2.8689,
      "theoretical_loss": 3.6697508450383634,
      "tokens_seen": 942978048
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00036076228686058177,
      "loss": 2.7812,
      "theoretical_loss": 3.6697263103061393,
      "tokens_seen": 943043584
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00036075225677031095,
      "loss": 2.5538,
      "theoretical_loss": 3.6697017777562433,
      "tokens_seen": 943109120
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00036074222668004013,
      "loss": 2.7327,
      "theoretical_loss": 3.6696772473883303,
      "tokens_seen": 943174656
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003607321965897693,
      "loss": 2.7159,
      "theoretical_loss": 3.6696527192020545,
      "tokens_seen": 943240192
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003607221664994985,
      "loss": 2.5015,
      "theoretical_loss": 3.66962819319707,
      "tokens_seen": 943305728
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003607121364092277,
      "loss": 2.6532,
      "theoretical_loss": 3.6696036693730316,
      "tokens_seen": 943371264
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003607021063189569,
      "loss": 2.7995,
      "theoretical_loss": 3.669579147729594,
      "tokens_seen": 943436800
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00036069207622868604,
      "loss": 2.6887,
      "theoretical_loss": 3.6695546282664115,
      "tokens_seen": 943502336
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00036068204613841527,
      "loss": 2.9105,
      "theoretical_loss": 3.669530110983139,
      "tokens_seen": 943567872
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003606720160481444,
      "loss": 2.4251,
      "theoretical_loss": 3.669505595879432,
      "tokens_seen": 943633408
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00036066198595787364,
      "loss": 2.7481,
      "theoretical_loss": 3.6694810829549445,
      "tokens_seen": 943698944
    },
    {
      "epoch": 3.02,
      "objective/train/docs_used": 1111689,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.0604324340820312,
      "objective/train/theoretical_loss": 3.669474955064285,
      "objective/train/tokens_used": 964175328,
      "theoretical_loss": 3.669474955064285,
      "tokens_seen": 943715328
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003606519558676028,
      "loss": 2.7923,
      "theoretical_loss": 3.6694565722093317,
      "tokens_seen": 943764480
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.000360641925777332,
      "loss": 2.8104,
      "theoretical_loss": 3.6694320636422484,
      "tokens_seen": 943830016
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003606318956870612,
      "loss": 2.9466,
      "theoretical_loss": 3.66940755725335,
      "tokens_seen": 943895552
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003606218655967904,
      "loss": 2.6342,
      "theoretical_loss": 3.669383053042292,
      "tokens_seen": 943961088
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00036061183550651954,
      "loss": 2.73,
      "theoretical_loss": 3.6693585510087288,
      "tokens_seen": 944026624
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003606018054162488,
      "loss": 2.8563,
      "theoretical_loss": 3.6693340511523167,
      "tokens_seen": 944092160
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003605917753259779,
      "loss": 2.6699,
      "theoretical_loss": 3.6693095534727105,
      "tokens_seen": 944157696
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00036058174523570714,
      "loss": 2.7349,
      "theoretical_loss": 3.6692850579695664,
      "tokens_seen": 944223232
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003605717151454363,
      "loss": 2.8572,
      "theoretical_loss": 3.6692605646425394,
      "tokens_seen": 944288768
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003605616850551655,
      "loss": 2.6415,
      "theoretical_loss": 3.6692360734912848,
      "tokens_seen": 944354304
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003605516549648947,
      "loss": 2.7184,
      "theoretical_loss": 3.6692115845154594,
      "tokens_seen": 944419840
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00036054162487462386,
      "loss": 2.7668,
      "theoretical_loss": 3.669187097714718,
      "tokens_seen": 944485376
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00036053159478435304,
      "loss": 2.4856,
      "theoretical_loss": 3.6691626130887176,
      "tokens_seen": 944550912
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003605215646940823,
      "loss": 2.7285,
      "theoretical_loss": 3.669138130637114,
      "tokens_seen": 944616448
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003605115346038114,
      "loss": 2.8301,
      "theoretical_loss": 3.669113650359562,
      "tokens_seen": 944681984
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00036050150451354064,
      "loss": 2.9229,
      "theoretical_loss": 3.6690891722557186,
      "tokens_seen": 944747520
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00036049147442326977,
      "loss": 2.8787,
      "theoretical_loss": 3.6690646963252402,
      "tokens_seen": 944813056
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.000360481444332999,
      "loss": 2.8984,
      "theoretical_loss": 3.6690402225677827,
      "tokens_seen": 944878592
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003604714142427282,
      "loss": 2.8173,
      "theoretical_loss": 3.6690157509830033,
      "tokens_seen": 944944128
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00036046138415245737,
      "loss": 2.3796,
      "theoretical_loss": 3.668991281570557,
      "tokens_seen": 945009664
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00036045135406218655,
      "loss": 2.8952,
      "theoretical_loss": 3.668966814330102,
      "tokens_seen": 945075200
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003604413239719158,
      "loss": 2.8054,
      "theoretical_loss": 3.668942349261294,
      "tokens_seen": 945140736
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003604312938816449,
      "loss": 2.7728,
      "theoretical_loss": 3.6689178863637895,
      "tokens_seen": 945206272
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00036042126379137414,
      "loss": 2.8421,
      "theoretical_loss": 3.668893425637245,
      "tokens_seen": 945271808
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00036041123370110327,
      "loss": 2.74,
      "theoretical_loss": 3.668868967081319,
      "tokens_seen": 945337344
    },
    {
      "epoch": 3.02,
      "objective/train/docs_used": 1112870,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6617534160614014,
      "objective/train/theoretical_loss": 3.668862852781456,
      "objective/train/tokens_used": 965813728,
      "theoretical_loss": 3.668862852781456,
      "tokens_seen": 945353728
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003604012036108325,
      "loss": 2.8314,
      "theoretical_loss": 3.6688445106956666,
      "tokens_seen": 945402880
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003603911735205617,
      "loss": 2.5809,
      "theoretical_loss": 3.668820056479946,
      "tokens_seen": 945468416
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00036038114343029087,
      "loss": 2.5361,
      "theoretical_loss": 3.6687956044338135,
      "tokens_seen": 945533952
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00036037111334002005,
      "loss": 2.7613,
      "theoretical_loss": 3.6687711545569264,
      "tokens_seen": 945599488
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00036036108324974923,
      "loss": 2.7731,
      "theoretical_loss": 3.668746706848942,
      "tokens_seen": 945665024
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00036035105315947847,
      "loss": 2.8019,
      "theoretical_loss": 3.668722261309518,
      "tokens_seen": 945730560
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00036034102306920765,
      "loss": 2.8439,
      "theoretical_loss": 3.6686978179383116,
      "tokens_seen": 945796096
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00036033099297893683,
      "loss": 2.8839,
      "theoretical_loss": 3.6686733767349797,
      "tokens_seen": 945861632
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.000360320962888666,
      "loss": 2.5422,
      "theoretical_loss": 3.6686489376991807,
      "tokens_seen": 945927168
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00036031093279839525,
      "loss": 2.8942,
      "theoretical_loss": 3.668624500830571,
      "tokens_seen": 945992704
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00036030090270812437,
      "loss": 2.5538,
      "theoretical_loss": 3.6686000661288105,
      "tokens_seen": 946058240
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003602908726178536,
      "loss": 2.7451,
      "theoretical_loss": 3.6685756335935547,
      "tokens_seen": 946123776
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00036028084252758273,
      "loss": 2.7525,
      "theoretical_loss": 3.668551203224463,
      "tokens_seen": 946189312
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00036027081243731197,
      "loss": 2.7951,
      "theoretical_loss": 3.668526775021192,
      "tokens_seen": 946254848
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00036026078234704115,
      "loss": 2.6365,
      "theoretical_loss": 3.6685023489834006,
      "tokens_seen": 946320384
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00036025075225677033,
      "loss": 2.812,
      "theoretical_loss": 3.6684779251107464,
      "tokens_seen": 946385920
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003602407221664995,
      "loss": 2.5312,
      "theoretical_loss": 3.6684535034028887,
      "tokens_seen": 946451456
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003602306920762287,
      "loss": 2.7091,
      "theoretical_loss": 3.668429083859484,
      "tokens_seen": 946516992
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003602206619859579,
      "loss": 2.9199,
      "theoretical_loss": 3.668404666480192,
      "tokens_seen": 946582528
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003602106318956871,
      "loss": 2.807,
      "theoretical_loss": 3.6683802512646704,
      "tokens_seen": 946648064
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00036020060180541624,
      "loss": 2.5709,
      "theoretical_loss": 3.6683558382125785,
      "tokens_seen": 946713600
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003601905717151455,
      "loss": 2.6406,
      "theoretical_loss": 3.6683314273235736,
      "tokens_seen": 946779136
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003601805416248746,
      "loss": 2.895,
      "theoretical_loss": 3.6683070185973152,
      "tokens_seen": 946844672
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00036017051153460384,
      "loss": 2.7009,
      "theoretical_loss": 3.668282612033462,
      "tokens_seen": 946910208
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.000360160481444333,
      "loss": 2.796,
      "theoretical_loss": 3.6682582076316725,
      "tokens_seen": 946975744
    },
    {
      "epoch": 3.02,
      "objective/train/docs_used": 1113638,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.09346079826355,
      "objective/train/theoretical_loss": 3.6682521068690077,
      "objective/train/tokens_used": 967452128,
      "theoretical_loss": 3.6682521068690077,
      "tokens_seen": 946992128
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003601504513540622,
      "loss": 2.9157,
      "theoretical_loss": 3.6682338053916057,
      "tokens_seen": 947041280
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003601404212637914,
      "loss": 2.5278,
      "theoretical_loss": 3.6682094053129206,
      "tokens_seen": 947106816
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003601303911735206,
      "loss": 2.7553,
      "theoretical_loss": 3.6681850073952758,
      "tokens_seen": 947172352
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00036012036108324974,
      "loss": 2.7387,
      "theoretical_loss": 3.6681606116383314,
      "tokens_seen": 947237888
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.000360110330992979,
      "loss": 2.6852,
      "theoretical_loss": 3.668136218041745,
      "tokens_seen": 947303424
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003601003009027081,
      "loss": 2.4123,
      "theoretical_loss": 3.6681118266051778,
      "tokens_seen": 947368960
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00036009027081243734,
      "loss": 2.7635,
      "theoretical_loss": 3.6680874373282872,
      "tokens_seen": 947434496
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003600802407221665,
      "loss": 2.7806,
      "theoretical_loss": 3.6680630502107334,
      "tokens_seen": 947500032
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003600702106318957,
      "loss": 2.7379,
      "theoretical_loss": 3.668038665252177,
      "tokens_seen": 947565568
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003600601805416249,
      "loss": 2.7847,
      "theoretical_loss": 3.668014282452276,
      "tokens_seen": 947631104
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00036005015045135406,
      "loss": 2.6898,
      "theoretical_loss": 3.6679899018106905,
      "tokens_seen": 947696640
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00036004012036108324,
      "loss": 2.7477,
      "theoretical_loss": 3.66796552332708,
      "tokens_seen": 947762176
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003600300902708125,
      "loss": 2.9667,
      "theoretical_loss": 3.667941147001105,
      "tokens_seen": 947827712
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003600200601805416,
      "loss": 2.6944,
      "theoretical_loss": 3.6679167728324247,
      "tokens_seen": 947893248
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00036001003009027084,
      "loss": 2.8867,
      "theoretical_loss": 3.6678924008206995,
      "tokens_seen": 947958784
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035999999999999997,
      "loss": 2.6801,
      "theoretical_loss": 3.6678680309655896,
      "tokens_seen": 948024320
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003599899699097292,
      "loss": 2.6097,
      "theoretical_loss": 3.667843663266754,
      "tokens_seen": 948089856
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003599799398194584,
      "loss": 2.7772,
      "theoretical_loss": 3.667819297723854,
      "tokens_seen": 948155392
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035996990972918757,
      "loss": 2.722,
      "theoretical_loss": 3.6677949343365492,
      "tokens_seen": 948220928
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035995987963891675,
      "loss": 2.6455,
      "theoretical_loss": 3.6677705731045007,
      "tokens_seen": 948286464
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.000359949849548646,
      "loss": 2.7656,
      "theoretical_loss": 3.6677462140273684,
      "tokens_seen": 948352000
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003599398194583751,
      "loss": 2.8047,
      "theoretical_loss": 3.6677218571048122,
      "tokens_seen": 948417536
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035992978936810434,
      "loss": 2.7955,
      "theoretical_loss": 3.6676975023364937,
      "tokens_seen": 948483072
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035991975927783347,
      "loss": 2.789,
      "theoretical_loss": 3.667673149722073,
      "tokens_seen": 948548608
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003599097291875627,
      "loss": 2.9306,
      "theoretical_loss": 3.667648799261211,
      "tokens_seen": 948614144
    },
    {
      "epoch": 3.02,
      "objective/train/docs_used": 1115063,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7750394344329834,
      "objective/train/theoretical_loss": 3.6676427119824493,
      "objective/train/tokens_used": 969090528,
      "theoretical_loss": 3.6676427119824493,
      "tokens_seen": 948630528
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003598996990972919,
      "loss": 2.7399,
      "theoretical_loss": 3.6676244509535687,
      "tokens_seen": 948679680
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035988966900702107,
      "loss": 2.7975,
      "theoretical_loss": 3.6676001047988063,
      "tokens_seen": 948745216
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035987963891675025,
      "loss": 2.7424,
      "theoretical_loss": 3.667575760796585,
      "tokens_seen": 948810752
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035986960882647943,
      "loss": 2.9323,
      "theoretical_loss": 3.667551418946566,
      "tokens_seen": 948876288
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003598595787362086,
      "loss": 2.8068,
      "theoretical_loss": 3.6675270792484107,
      "tokens_seen": 948941824
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035984954864593785,
      "loss": 2.9573,
      "theoretical_loss": 3.66750274170178,
      "tokens_seen": 949007360
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.000359839518555667,
      "loss": 2.6737,
      "theoretical_loss": 3.6674784063063344,
      "tokens_seen": 949072896
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003598294884653962,
      "loss": 2.7313,
      "theoretical_loss": 3.6674540730617364,
      "tokens_seen": 949138432
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035981945837512534,
      "loss": 2.6642,
      "theoretical_loss": 3.6674297419676476,
      "tokens_seen": 949203968
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035980942828485457,
      "loss": 2.8812,
      "theoretical_loss": 3.6674054130237277,
      "tokens_seen": 949269504
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035979939819458375,
      "loss": 2.6248,
      "theoretical_loss": 3.66738108622964,
      "tokens_seen": 949335040
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035978936810431293,
      "loss": 2.7558,
      "theoretical_loss": 3.6673567615850455,
      "tokens_seen": 949400576
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003597793380140421,
      "loss": 2.6226,
      "theoretical_loss": 3.6673324390896056,
      "tokens_seen": 949466112
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035976930792377135,
      "loss": 2.5855,
      "theoretical_loss": 3.667308118742983,
      "tokens_seen": 949531648
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003597592778335005,
      "loss": 2.7705,
      "theoretical_loss": 3.6672838005448387,
      "tokens_seen": 949597184
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003597492477432297,
      "loss": 2.692,
      "theoretical_loss": 3.667259484494835,
      "tokens_seen": 949662720
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035973921765295884,
      "loss": 2.8313,
      "theoretical_loss": 3.667235170592634,
      "tokens_seen": 949728256
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003597291875626881,
      "loss": 2.8125,
      "theoretical_loss": 3.667210858837897,
      "tokens_seen": 949793792
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035971915747241726,
      "loss": 2.6625,
      "theoretical_loss": 3.6671865492302875,
      "tokens_seen": 949859328
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035970912738214644,
      "loss": 2.6021,
      "theoretical_loss": 3.667162241769467,
      "tokens_seen": 949924864
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003596990972918756,
      "loss": 2.829,
      "theoretical_loss": 3.667137936455098,
      "tokens_seen": 949990400
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003596890672016048,
      "loss": 2.7101,
      "theoretical_loss": 3.667113633286842,
      "tokens_seen": 950055936
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.000359679037111334,
      "loss": 2.6948,
      "theoretical_loss": 3.6670893322643634,
      "tokens_seen": 950121472
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003596690070210632,
      "loss": 2.716,
      "theoretical_loss": 3.667065033387323,
      "tokens_seen": 950187008
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035965897693079234,
      "loss": 2.7286,
      "theoretical_loss": 3.6670407366553848,
      "tokens_seen": 950252544
    },
    {
      "epoch": 3.02,
      "objective/train/docs_used": 1115859,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.0115370750427246,
      "objective/train/theoretical_loss": 3.6670346628075325,
      "objective/train/tokens_used": 970728928,
      "theoretical_loss": 3.6670346628075325,
      "tokens_seen": 950268928
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003596489468405216,
      "loss": 2.8904,
      "theoretical_loss": 3.66701644206821,
      "tokens_seen": 950318080
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035963891675025076,
      "loss": 2.7355,
      "theoretical_loss": 3.6669921496254627,
      "tokens_seen": 950383616
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035962888665997994,
      "loss": 2.6775,
      "theoretical_loss": 3.666967859326805,
      "tokens_seen": 950449152
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003596188565697091,
      "loss": 2.7278,
      "theoretical_loss": 3.6669435711719,
      "tokens_seen": 950514688
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003596088264794383,
      "loss": 2.8868,
      "theoretical_loss": 3.6669192851604113,
      "tokens_seen": 950580224
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035959879638916754,
      "loss": 2.7949,
      "theoretical_loss": 3.666895001292001,
      "tokens_seen": 950645760
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003595887662988967,
      "loss": 2.8372,
      "theoretical_loss": 3.666870719566333,
      "tokens_seen": 950711296
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003595787362086259,
      "loss": 2.6296,
      "theoretical_loss": 3.66684643998307,
      "tokens_seen": 950776832
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003595687061183551,
      "loss": 2.5486,
      "theoretical_loss": 3.6668221625418758,
      "tokens_seen": 950842368
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035955867602808426,
      "loss": 2.7116,
      "theoretical_loss": 3.666797887242414,
      "tokens_seen": 950907904
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035954864593781344,
      "loss": 2.6305,
      "theoretical_loss": 3.6667736140843474,
      "tokens_seen": 950973440
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003595386158475427,
      "loss": 2.7846,
      "theoretical_loss": 3.6667493430673397,
      "tokens_seen": 951038976
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003595285857572718,
      "loss": 2.7952,
      "theoretical_loss": 3.6667250741910546,
      "tokens_seen": 951104512
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035951855566700104,
      "loss": 2.8643,
      "theoretical_loss": 3.666700807455156,
      "tokens_seen": 951170048
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035950852557673017,
      "loss": 2.6962,
      "theoretical_loss": 3.6666765428593076,
      "tokens_seen": 951235584
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003594984954864594,
      "loss": 2.7865,
      "theoretical_loss": 3.6666522804031736,
      "tokens_seen": 951301120
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003594884653961886,
      "loss": 2.5407,
      "theoretical_loss": 3.666628020086417,
      "tokens_seen": 951366656
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035947843530591777,
      "loss": 2.3828,
      "theoretical_loss": 3.6666037619087026,
      "tokens_seen": 951432192
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035946840521564695,
      "loss": 2.8197,
      "theoretical_loss": 3.6665795058696937,
      "tokens_seen": 951497728
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003594583751253762,
      "loss": 2.7182,
      "theoretical_loss": 3.666555251969055,
      "tokens_seen": 951563264
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003594483450351053,
      "loss": 2.729,
      "theoretical_loss": 3.6665310002064513,
      "tokens_seen": 951628800
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035943831494483454,
      "loss": 2.6999,
      "theoretical_loss": 3.6665067505815454,
      "tokens_seen": 951694336
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035942828485456367,
      "loss": 2.7045,
      "theoretical_loss": 3.6664825030940036,
      "tokens_seen": 951759872
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003594182547642929,
      "loss": 2.5441,
      "theoretical_loss": 3.6664582577434888,
      "tokens_seen": 951825408
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003594082246740221,
      "loss": 2.7172,
      "theoretical_loss": 3.666434014529666,
      "tokens_seen": 951890944
    },
    {
      "epoch": 3.02,
      "objective/train/docs_used": 1116983,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5958456993103027,
      "objective/train/theoretical_loss": 3.666427954060029,
      "objective/train/tokens_used": 972367328,
      "theoretical_loss": 3.666427954060029,
      "tokens_seen": 951907328
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035939819458375127,
      "loss": 2.7405,
      "theoretical_loss": 3.6664097734522,
      "tokens_seen": 951956480
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035938816449348045,
      "loss": 2.8947,
      "theoretical_loss": 3.666385534510755,
      "tokens_seen": 952022016
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035937813440320963,
      "loss": 2.7593,
      "theoretical_loss": 3.666361297704997,
      "tokens_seen": 952087552
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003593681043129388,
      "loss": 2.8236,
      "theoretical_loss": 3.666337063034589,
      "tokens_seen": 952153088
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035935807422266805,
      "loss": 2.4366,
      "theoretical_loss": 3.666312830499197,
      "tokens_seen": 952218624
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003593480441323972,
      "loss": 2.8637,
      "theoretical_loss": 3.6662886000984862,
      "tokens_seen": 952284160
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003593380140421264,
      "loss": 2.8358,
      "theoretical_loss": 3.6662643718321215,
      "tokens_seen": 952349696
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035932798395185554,
      "loss": 2.7572,
      "theoretical_loss": 3.6662401456997675,
      "tokens_seen": 952415232
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035931795386158477,
      "loss": 2.8309,
      "theoretical_loss": 3.66621592170109,
      "tokens_seen": 952480768
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035930792377131395,
      "loss": 3.0672,
      "theoretical_loss": 3.6661916998357538,
      "tokens_seen": 952546304
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035929789368104313,
      "loss": 2.6583,
      "theoretical_loss": 3.6661674801034243,
      "tokens_seen": 952611840
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003592878635907723,
      "loss": 2.6992,
      "theoretical_loss": 3.666143262503768,
      "tokens_seen": 952677376
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035927783350050155,
      "loss": 2.7778,
      "theoretical_loss": 3.666119047036449,
      "tokens_seen": 952742912
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003592678034102307,
      "loss": 2.7916,
      "theoretical_loss": 3.6660948337011336,
      "tokens_seen": 952808448
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003592577733199599,
      "loss": 2.6527,
      "theoretical_loss": 3.6660706224974877,
      "tokens_seen": 952873984
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035924774322968904,
      "loss": 2.7234,
      "theoretical_loss": 3.666046413425176,
      "tokens_seen": 952939520
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003592377131394183,
      "loss": 2.7744,
      "theoretical_loss": 3.6660222064838655,
      "tokens_seen": 953005056
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035922768304914746,
      "loss": 2.608,
      "theoretical_loss": 3.665998001673221,
      "tokens_seen": 953070592
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035921765295887664,
      "loss": 2.6749,
      "theoretical_loss": 3.66597379899291,
      "tokens_seen": 953136128
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003592076228686058,
      "loss": 2.5654,
      "theoretical_loss": 3.665949598442597,
      "tokens_seen": 953201664
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.000359197592778335,
      "loss": 2.7041,
      "theoretical_loss": 3.665925400021949,
      "tokens_seen": 953267200
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003591875626880642,
      "loss": 2.906,
      "theoretical_loss": 3.6659012037306318,
      "tokens_seen": 953332736
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003591775325977934,
      "loss": 2.7529,
      "theoretical_loss": 3.665877009568312,
      "tokens_seen": 953398272
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035916750250752254,
      "loss": 2.8305,
      "theoretical_loss": 3.665852817534655,
      "tokens_seen": 953463808
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003591574724172518,
      "loss": 2.9454,
      "theoretical_loss": 3.6658286276293293,
      "tokens_seen": 953529344
    },
    {
      "epoch": 3.02,
      "objective/train/docs_used": 1117727,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.588460683822632,
      "objective/train/theoretical_loss": 3.66582258048551,
      "objective/train/tokens_used": 974005728,
      "theoretical_loss": 3.66582258048551,
      "tokens_seen": 953545728
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035914744232698096,
      "loss": 2.6301,
      "theoretical_loss": 3.6658044398519993,
      "tokens_seen": 953594880
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035913741223671014,
      "loss": 2.8647,
      "theoretical_loss": 3.6657802542023323,
      "tokens_seen": 953660416
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003591273821464393,
      "loss": 2.7181,
      "theoretical_loss": 3.665756070679995,
      "tokens_seen": 953725952
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003591173520561685,
      "loss": 2.856,
      "theoretical_loss": 3.6657318892846544,
      "tokens_seen": 953791488
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003591073219658977,
      "loss": 2.7305,
      "theoretical_loss": 3.665707710015977,
      "tokens_seen": 953857024
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003590972918756269,
      "loss": 2.9282,
      "theoretical_loss": 3.6656835328736292,
      "tokens_seen": 953922560
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035908726178535605,
      "loss": 2.6648,
      "theoretical_loss": 3.6656593578572787,
      "tokens_seen": 953988096
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003590772316950853,
      "loss": 2.7987,
      "theoretical_loss": 3.6656351849665922,
      "tokens_seen": 954053632
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003590672016048144,
      "loss": 2.4934,
      "theoretical_loss": 3.665611014201237,
      "tokens_seen": 954119168
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035905717151454364,
      "loss": 2.6069,
      "theoretical_loss": 3.66558684556088,
      "tokens_seen": 954184704
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003590471414242728,
      "loss": 2.8056,
      "theoretical_loss": 3.665562679045189,
      "tokens_seen": 954250240
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.000359037111334002,
      "loss": 2.9904,
      "theoretical_loss": 3.6655385146538304,
      "tokens_seen": 954315776
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003590270812437312,
      "loss": 2.6449,
      "theoretical_loss": 3.665514352386472,
      "tokens_seen": 954381312
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035901705115346037,
      "loss": 2.7939,
      "theoretical_loss": 3.665490192242782,
      "tokens_seen": 954446848
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035900702106318955,
      "loss": 2.7474,
      "theoretical_loss": 3.6654660342224266,
      "tokens_seen": 954512384
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003589969909729188,
      "loss": 2.6304,
      "theoretical_loss": 3.6654418783250744,
      "tokens_seen": 954577920
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003589869608826479,
      "loss": 2.6741,
      "theoretical_loss": 3.6654177245503927,
      "tokens_seen": 954643456
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035897693079237715,
      "loss": 2.6727,
      "theoretical_loss": 3.66539357289805,
      "tokens_seen": 954708992
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035896690070210633,
      "loss": 2.6506,
      "theoretical_loss": 3.6653694233677125,
      "tokens_seen": 954774528
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003589568706118355,
      "loss": 2.7298,
      "theoretical_loss": 3.6653452759590497,
      "tokens_seen": 954840064
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003589468405215647,
      "loss": 2.7313,
      "theoretical_loss": 3.665321130671729,
      "tokens_seen": 954905600
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035893681043129387,
      "loss": 2.9473,
      "theoretical_loss": 3.665296987505419,
      "tokens_seen": 954971136
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035892678034102305,
      "loss": 2.7933,
      "theoretical_loss": 3.6652728464597866,
      "tokens_seen": 955036672
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003589167502507523,
      "loss": 2.6645,
      "theoretical_loss": 3.6652487075345013,
      "tokens_seen": 955102208
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003589067201604814,
      "loss": 2.605,
      "theoretical_loss": 3.6652245707292304,
      "tokens_seen": 955167744
    },
    {
      "epoch": 3.02,
      "objective/train/docs_used": 1119307,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7554943561553955,
      "objective/train/theoretical_loss": 3.6652185368591264,
      "objective/train/tokens_used": 975644128,
      "theoretical_loss": 3.6652185368591264,
      "tokens_seen": 955184128
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035889669007021065,
      "loss": 2.7885,
      "theoretical_loss": 3.6652004360436434,
      "tokens_seen": 955233280
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003588866599799398,
      "loss": 2.7021,
      "theoretical_loss": 3.6651763034774074,
      "tokens_seen": 955298816
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.000358876629889669,
      "loss": 2.7756,
      "theoretical_loss": 3.665152173030192,
      "tokens_seen": 955364352
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003588665997993982,
      "loss": 2.6781,
      "theoretical_loss": 3.6651280447016656,
      "tokens_seen": 955429888
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003588565697091274,
      "loss": 2.8244,
      "theoretical_loss": 3.6651039184914964,
      "tokens_seen": 955495424
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003588465396188566,
      "loss": 2.6043,
      "theoretical_loss": 3.665079794399354,
      "tokens_seen": 955560960
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035883650952858574,
      "loss": 2.5664,
      "theoretical_loss": 3.6650556724249057,
      "tokens_seen": 955626496
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035882647943831497,
      "loss": 2.6075,
      "theoretical_loss": 3.665031552567822,
      "tokens_seen": 955692032
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035881644934804415,
      "loss": 2.5505,
      "theoretical_loss": 3.665007434827771,
      "tokens_seen": 955757568
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035880641925777333,
      "loss": 2.5502,
      "theoretical_loss": 3.664983319204422,
      "tokens_seen": 955823104
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003587963891675025,
      "loss": 2.5772,
      "theoretical_loss": 3.664959205697444,
      "tokens_seen": 955888640
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035878635907723175,
      "loss": 2.7139,
      "theoretical_loss": 3.664935094306506,
      "tokens_seen": 955954176
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003587763289869609,
      "loss": 2.8381,
      "theoretical_loss": 3.6649109850312778,
      "tokens_seen": 956019712
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003587662988966901,
      "loss": 2.5263,
      "theoretical_loss": 3.664886877871428,
      "tokens_seen": 956085248
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035875626880641924,
      "loss": 2.7676,
      "theoretical_loss": 3.664862772826627,
      "tokens_seen": 956150784
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003587462387161485,
      "loss": 2.7778,
      "theoretical_loss": 3.6648386698965436,
      "tokens_seen": 956216320
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035873620862587766,
      "loss": 2.637,
      "theoretical_loss": 3.664814569080847,
      "tokens_seen": 956281856
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035872617853560684,
      "loss": 2.7548,
      "theoretical_loss": 3.664790470379208,
      "tokens_seen": 956347392
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.000358716148445336,
      "loss": 2.8223,
      "theoretical_loss": 3.6647663737912954,
      "tokens_seen": 956412928
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003587061183550652,
      "loss": 2.8319,
      "theoretical_loss": 3.664742279316779,
      "tokens_seen": 956478464
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003586960882647944,
      "loss": 2.6764,
      "theoretical_loss": 3.664718186955329,
      "tokens_seen": 956544000
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003586860581745236,
      "loss": 2.7721,
      "theoretical_loss": 3.6646940967066155,
      "tokens_seen": 956609536
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035867602808425274,
      "loss": 2.5923,
      "theoretical_loss": 3.664670008570308,
      "tokens_seen": 956675072
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.000358665997993982,
      "loss": 2.7841,
      "theoretical_loss": 3.664645922546077,
      "tokens_seen": 956740608
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035865596790371116,
      "loss": 2.5776,
      "theoretical_loss": 3.6646218386335923,
      "tokens_seen": 956806144
    },
    {
      "epoch": 3.02,
      "objective/train/docs_used": 1120060,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.80765438079834,
      "objective/train/theoretical_loss": 3.6646158179853927,
      "objective/train/tokens_used": 977282528,
      "theoretical_loss": 3.6646158179853927,
      "tokens_seen": 956822528
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035864593781344034,
      "loss": 2.6522,
      "theoretical_loss": 3.6645977568325243,
      "tokens_seen": 956871680
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003586359077231695,
      "loss": 2.8831,
      "theoretical_loss": 3.664573677142543,
      "tokens_seen": 956937216
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003586258776328987,
      "loss": 2.4891,
      "theoretical_loss": 3.664549599563319,
      "tokens_seen": 957002752
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003586158475426279,
      "loss": 2.6942,
      "theoretical_loss": 3.6645255240945236,
      "tokens_seen": 957068288
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003586058174523571,
      "loss": 2.6859,
      "theoretical_loss": 3.664501450735826,
      "tokens_seen": 957133824
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035859578736208625,
      "loss": 2.7707,
      "theoretical_loss": 3.6644773794868972,
      "tokens_seen": 957199360
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003585857572718155,
      "loss": 2.6198,
      "theoretical_loss": 3.664453310347408,
      "tokens_seen": 957264896
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003585757271815446,
      "loss": 2.8187,
      "theoretical_loss": 3.6644292433170293,
      "tokens_seen": 957330432
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035856569709127384,
      "loss": 2.5686,
      "theoretical_loss": 3.664405178395432,
      "tokens_seen": 957395968
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.000358555667001003,
      "loss": 2.922,
      "theoretical_loss": 3.6643811155822865,
      "tokens_seen": 957461504
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003585456369107322,
      "loss": 2.66,
      "theoretical_loss": 3.6643570548772644,
      "tokens_seen": 957527040
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003585356068204614,
      "loss": 2.8352,
      "theoretical_loss": 3.6643329962800353,
      "tokens_seen": 957592576
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035852557673019057,
      "loss": 2.6895,
      "theoretical_loss": 3.6643089397902724,
      "tokens_seen": 957658112
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035851554663991975,
      "loss": 2.9951,
      "theoretical_loss": 3.6642848854076453,
      "tokens_seen": 957723648
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.000358505516549649,
      "loss": 2.7052,
      "theoretical_loss": 3.6642608331318263,
      "tokens_seen": 957789184
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003584954864593781,
      "loss": 2.7671,
      "theoretical_loss": 3.6642367829624862,
      "tokens_seen": 957854720
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035848545636910735,
      "loss": 2.7657,
      "theoretical_loss": 3.664212734899296,
      "tokens_seen": 957920256
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035847542627883653,
      "loss": 2.7118,
      "theoretical_loss": 3.6641886889419286,
      "tokens_seen": 957985792
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003584653961885657,
      "loss": 2.7238,
      "theoretical_loss": 3.6641646450900542,
      "tokens_seen": 958051328
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003584553660982949,
      "loss": 2.8321,
      "theoretical_loss": 3.664140603343345,
      "tokens_seen": 958116864
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035844533600802407,
      "loss": 2.7182,
      "theoretical_loss": 3.664116563701472,
      "tokens_seen": 958182400
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035843530591775325,
      "loss": 2.5963,
      "theoretical_loss": 3.664092526164108,
      "tokens_seen": 958247936
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003584252758274825,
      "loss": 2.9829,
      "theoretical_loss": 3.6640684907309247,
      "tokens_seen": 958313472
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003584152457372116,
      "loss": 2.8543,
      "theoretical_loss": 3.6640444574015927,
      "tokens_seen": 958379008
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035840521564694085,
      "loss": 2.6679,
      "theoretical_loss": 3.664020426175786,
      "tokens_seen": 958444544
    },
    {
      "epoch": 3.02,
      "objective/train/docs_used": 1121284,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.857239246368408,
      "objective/train/theoretical_loss": 3.6640144186979713,
      "objective/train/tokens_used": 978920928,
      "theoretical_loss": 3.6640144186979713,
      "tokens_seen": 958460928
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035839518555667,
      "loss": 2.6887,
      "theoretical_loss": 3.6639963970531753,
      "tokens_seen": 958510080
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003583851554663992,
      "loss": 2.8986,
      "theoretical_loss": 3.6639723700334335,
      "tokens_seen": 958575616
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003583751253761284,
      "loss": 2.7256,
      "theoretical_loss": 3.663948345116232,
      "tokens_seen": 958641152
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003583650952858576,
      "loss": 2.7297,
      "theoretical_loss": 3.663924322301244,
      "tokens_seen": 958706688
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035835506519558676,
      "loss": 2.6622,
      "theoretical_loss": 3.663900301588141,
      "tokens_seen": 958772224
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035834503510531594,
      "loss": 2.5843,
      "theoretical_loss": 3.663876282976596,
      "tokens_seen": 958837760
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003583350050150451,
      "loss": 2.6251,
      "theoretical_loss": 3.6638522664662814,
      "tokens_seen": 958903296
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035832497492477435,
      "loss": 2.8493,
      "theoretical_loss": 3.66382825205687,
      "tokens_seen": 958968832
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003583149448345035,
      "loss": 2.9119,
      "theoretical_loss": 3.6638042397480337,
      "tokens_seen": 959034368
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003583049147442327,
      "loss": 2.6118,
      "theoretical_loss": 3.663780229539446,
      "tokens_seen": 959099904
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003582948846539619,
      "loss": 2.6434,
      "theoretical_loss": 3.6637562214307797,
      "tokens_seen": 959165440
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003582848545636911,
      "loss": 2.779,
      "theoretical_loss": 3.6637322154217076,
      "tokens_seen": 959230976
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035827482447342026,
      "loss": 2.773,
      "theoretical_loss": 3.6637082115119024,
      "tokens_seen": 959296512
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035826479438314944,
      "loss": 2.7536,
      "theoretical_loss": 3.663684209701037,
      "tokens_seen": 959362048
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003582547642928786,
      "loss": 2.7197,
      "theoretical_loss": 3.6636602099887847,
      "tokens_seen": 959427584
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035824473420260786,
      "loss": 2.8215,
      "theoretical_loss": 3.6636362123748194,
      "tokens_seen": 959493120
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.000358234704112337,
      "loss": 2.921,
      "theoretical_loss": 3.663612216858813,
      "tokens_seen": 959558656
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003582246740220662,
      "loss": 2.8807,
      "theoretical_loss": 3.6635882234404393,
      "tokens_seen": 959624192
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035821464393179535,
      "loss": 2.7939,
      "theoretical_loss": 3.6635642321193727,
      "tokens_seen": 959689728
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003582046138415246,
      "loss": 2.8857,
      "theoretical_loss": 3.6635402428952855,
      "tokens_seen": 959755264
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035819458375125376,
      "loss": 2.5746,
      "theoretical_loss": 3.6635162557678513,
      "tokens_seen": 959820800
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035818455366098294,
      "loss": 2.6412,
      "theoretical_loss": 3.6634922707367443,
      "tokens_seen": 959886336
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003581745235707121,
      "loss": 2.4484,
      "theoretical_loss": 3.6634682878016376,
      "tokens_seen": 959951872
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035816449348044136,
      "loss": 2.7905,
      "theoretical_loss": 3.6634443069622056,
      "tokens_seen": 960017408
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003581544633901705,
      "loss": 2.6442,
      "theoretical_loss": 3.6634203282181215,
      "tokens_seen": 960082944
    },
    {
      "epoch": 3.02,
      "objective/train/docs_used": 1121988,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.3468472957611084,
      "objective/train/theoretical_loss": 3.6634143338594605,
      "objective/train/tokens_used": 980559328,
      "theoretical_loss": 3.6634143338594605,
      "tokens_seen": 960099328
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003581444332998997,
      "loss": 2.4599,
      "theoretical_loss": 3.6633963515690597,
      "tokens_seen": 960148480
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035813440320962885,
      "loss": 2.7676,
      "theoretical_loss": 3.663372377014694,
      "tokens_seen": 960214016
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003581243731193581,
      "loss": 2.9168,
      "theoretical_loss": 3.6633484045546982,
      "tokens_seen": 960279552
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035811434302908727,
      "loss": 2.7659,
      "theoretical_loss": 3.6633244341887465,
      "tokens_seen": 960345088
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035810431293881645,
      "loss": 2.8983,
      "theoretical_loss": 3.663300465916514,
      "tokens_seen": 960410624
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003580942828485457,
      "loss": 2.7638,
      "theoretical_loss": 3.663276499737673,
      "tokens_seen": 960476160
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003580842527582748,
      "loss": 2.673,
      "theoretical_loss": 3.6632525356518997,
      "tokens_seen": 960541696
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035807422266800404,
      "loss": 2.8277,
      "theoretical_loss": 3.663228573658868,
      "tokens_seen": 960607232
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003580641925777332,
      "loss": 2.9438,
      "theoretical_loss": 3.6632046137582526,
      "tokens_seen": 960672768
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003580541624874624,
      "loss": 2.6973,
      "theoretical_loss": 3.663180655949727,
      "tokens_seen": 960738304
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003580441323971916,
      "loss": 2.8067,
      "theoretical_loss": 3.663156700232967,
      "tokens_seen": 960803840
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035803410230692077,
      "loss": 2.8689,
      "theoretical_loss": 3.663132746607647,
      "tokens_seen": 960869376
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035802407221664995,
      "loss": 2.9861,
      "theoretical_loss": 3.663108795073441,
      "tokens_seen": 960934912
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003580140421263792,
      "loss": 2.9207,
      "theoretical_loss": 3.663084845630025,
      "tokens_seen": 961000448
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003580040120361083,
      "loss": 2.8269,
      "theoretical_loss": 3.663060898277073,
      "tokens_seen": 961065984
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035799398194583755,
      "loss": 2.7846,
      "theoretical_loss": 3.663036953014261,
      "tokens_seen": 961131520
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035798395185556673,
      "loss": 2.8141,
      "theoretical_loss": 3.663013009841263,
      "tokens_seen": 961197056
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003579739217652959,
      "loss": 2.8821,
      "theoretical_loss": 3.6629890687577547,
      "tokens_seen": 961262592
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003579638916750251,
      "loss": 2.8052,
      "theoretical_loss": 3.662965129763412,
      "tokens_seen": 961328128
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035795386158475427,
      "loss": 2.7982,
      "theoretical_loss": 3.6629411928579083,
      "tokens_seen": 961393664
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035794383149448345,
      "loss": 2.7195,
      "theoretical_loss": 3.662917258040921,
      "tokens_seen": 961459200
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003579338014042127,
      "loss": 2.8697,
      "theoretical_loss": 3.6628933253121243,
      "tokens_seen": 961524736
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003579237713139418,
      "loss": 2.8098,
      "theoretical_loss": 3.6628693946711937,
      "tokens_seen": 961590272
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035791374122367105,
      "loss": 2.8547,
      "theoretical_loss": 3.6628454661178056,
      "tokens_seen": 961655808
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003579037111334002,
      "loss": 2.541,
      "theoretical_loss": 3.662821539651635,
      "tokens_seen": 961721344
    },
    {
      "epoch": 3.02,
      "objective/train/docs_used": 1123320,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.837873697280884,
      "objective/train/theoretical_loss": 3.6628155583611823,
      "objective/train/tokens_used": 982197728,
      "theoretical_loss": 3.6628155583611823,
      "tokens_seen": 961737728
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003578936810431294,
      "loss": 2.8232,
      "theoretical_loss": 3.662797615272358,
      "tokens_seen": 961786880
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003578836509528586,
      "loss": 2.8893,
      "theoretical_loss": 3.6627736929796497,
      "tokens_seen": 961852416
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003578736208625878,
      "loss": 2.537,
      "theoretical_loss": 3.6627497727731866,
      "tokens_seen": 961917952
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035786359077231696,
      "loss": 2.8157,
      "theoretical_loss": 3.6627258546526447,
      "tokens_seen": 961983488
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035785356068204614,
      "loss": 3.0576,
      "theoretical_loss": 3.6627019386177,
      "tokens_seen": 962049024
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003578435305917753,
      "loss": 2.7024,
      "theoretical_loss": 3.662678024668028,
      "tokens_seen": 962114560
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035783350050150455,
      "loss": 2.7436,
      "theoretical_loss": 3.6626541128033057,
      "tokens_seen": 962180096
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003578234704112337,
      "loss": 2.741,
      "theoretical_loss": 3.662630203023209,
      "tokens_seen": 962245632
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003578134403209629,
      "loss": 2.8045,
      "theoretical_loss": 3.6626062953274134,
      "tokens_seen": 962311168
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003578034102306921,
      "loss": 2.6184,
      "theoretical_loss": 3.6625823897155962,
      "tokens_seen": 962376704
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003577933801404213,
      "loss": 2.9925,
      "theoretical_loss": 3.662558486187434,
      "tokens_seen": 962442240
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035778335005015046,
      "loss": 3.0443,
      "theoretical_loss": 3.6625345847426027,
      "tokens_seen": 962507776
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035777331995987964,
      "loss": 2.714,
      "theoretical_loss": 3.662510685380779,
      "tokens_seen": 962573312
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003577632898696088,
      "loss": 2.6968,
      "theoretical_loss": 3.6624867881016403,
      "tokens_seen": 962638848
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035775325977933806,
      "loss": 2.931,
      "theoretical_loss": 3.662462892904862,
      "tokens_seen": 962704384
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003577432296890672,
      "loss": 2.6591,
      "theoretical_loss": 3.662438999790122,
      "tokens_seen": 962769920
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003577331995987964,
      "loss": 3.067,
      "theoretical_loss": 3.662415108757097,
      "tokens_seen": 962835456
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035772316950852555,
      "loss": 2.8951,
      "theoretical_loss": 3.6623912198054636,
      "tokens_seen": 962900992
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003577131394182548,
      "loss": 2.7869,
      "theoretical_loss": 3.662367332934899,
      "tokens_seen": 962966528
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035770310932798396,
      "loss": 3.0066,
      "theoretical_loss": 3.66234344814508,
      "tokens_seen": 963032064
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035769307923771314,
      "loss": 2.7544,
      "theoretical_loss": 3.6623195654356846,
      "tokens_seen": 963097600
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003576830491474423,
      "loss": 2.7924,
      "theoretical_loss": 3.6622956848063897,
      "tokens_seen": 963163136
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035767301905717156,
      "loss": 2.6071,
      "theoretical_loss": 3.6622718062568715,
      "tokens_seen": 963228672
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003576629889669007,
      "loss": 2.8556,
      "theoretical_loss": 3.662247929786809,
      "tokens_seen": 963294208
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003576529588766299,
      "loss": 2.9046,
      "theoretical_loss": 3.662224055395879,
      "tokens_seen": 963359744
    },
    {
      "epoch": 3.02,
      "objective/train/docs_used": 1123898,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7481706142425537,
      "objective/train/theoretical_loss": 3.662218087122973,
      "objective/train/tokens_used": 983836128,
      "theoretical_loss": 3.662218087122973,
      "tokens_seen": 963376128
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035764292878635905,
      "loss": 2.9226,
      "theoretical_loss": 3.662200183083759,
      "tokens_seen": 963425280
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003576328986960883,
      "loss": 2.8547,
      "theoretical_loss": 3.6621763128501263,
      "tokens_seen": 963490816
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035762286860581747,
      "loss": 2.7839,
      "theoretical_loss": 3.6621524446946587,
      "tokens_seen": 963556352
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035761283851554665,
      "loss": 2.7891,
      "theoretical_loss": 3.662128578617035,
      "tokens_seen": 963621888
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035760280842527583,
      "loss": 2.7659,
      "theoretical_loss": 3.662104714616931,
      "tokens_seen": 963687424
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.000357592778335005,
      "loss": 2.864,
      "theoretical_loss": 3.662080852694027,
      "tokens_seen": 963752960
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003575827482447342,
      "loss": 2.6925,
      "theoretical_loss": 3.662056992847999,
      "tokens_seen": 963818496
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003575727181544634,
      "loss": 2.9939,
      "theoretical_loss": 3.6620331350785253,
      "tokens_seen": 963884032
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035756268806419255,
      "loss": 2.5997,
      "theoretical_loss": 3.662009279385285,
      "tokens_seen": 963949568
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003575526579739218,
      "loss": 2.5903,
      "theoretical_loss": 3.661985425767956,
      "tokens_seen": 964015104
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003575426278836509,
      "loss": 2.9253,
      "theoretical_loss": 3.6619615742262157,
      "tokens_seen": 964080640
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035753259779338015,
      "loss": 3.0226,
      "theoretical_loss": 3.661937724759743,
      "tokens_seen": 964146176
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035752256770310933,
      "loss": 2.7521,
      "theoretical_loss": 3.6619138773682165,
      "tokens_seen": 964211712
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003575125376128385,
      "loss": 2.8833,
      "theoretical_loss": 3.6618900320513146,
      "tokens_seen": 964277248
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003575025075225677,
      "loss": 2.9383,
      "theoretical_loss": 3.661866188808715,
      "tokens_seen": 964342784
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035749247743229693,
      "loss": 2.7874,
      "theoretical_loss": 3.661842347640098,
      "tokens_seen": 964408320
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035748244734202606,
      "loss": 2.7917,
      "theoretical_loss": 3.6618185085451405,
      "tokens_seen": 964473856
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003574724172517553,
      "loss": 2.7022,
      "theoretical_loss": 3.661794671523522,
      "tokens_seen": 964539392
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003574623871614844,
      "loss": 2.6121,
      "theoretical_loss": 3.661770836574922,
      "tokens_seen": 964604928
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035745235707121365,
      "loss": 2.8655,
      "theoretical_loss": 3.661747003699018,
      "tokens_seen": 964670464
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035744232698094283,
      "loss": 2.7242,
      "theoretical_loss": 3.66172317289549,
      "tokens_seen": 964736000
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.000357432296890672,
      "loss": 2.9418,
      "theoretical_loss": 3.661699344164017,
      "tokens_seen": 964801536
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003574222668004012,
      "loss": 2.9072,
      "theoretical_loss": 3.661675517504277,
      "tokens_seen": 964867072
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003574122367101304,
      "loss": 2.6928,
      "theoretical_loss": 3.6616516929159504,
      "tokens_seen": 964932608
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035740220661985956,
      "loss": 2.5438,
      "theoretical_loss": 3.661627870398716,
      "tokens_seen": 964998144
    },
    {
      "epoch": 3.02,
      "objective/train/docs_used": 1124893,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8355164527893066,
      "objective/train/theoretical_loss": 3.6616219150929785,
      "objective/train/tokens_used": 985474528,
      "theoretical_loss": 3.6616219150929785,
      "tokens_seen": 965014528
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003573921765295888,
      "loss": 2.7413,
      "theoretical_loss": 3.6616040499522535,
      "tokens_seen": 965063680
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003573821464393179,
      "loss": 2.8473,
      "theoretical_loss": 3.661580231576242,
      "tokens_seen": 965129216
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035737211634904716,
      "loss": 2.5775,
      "theoretical_loss": 3.6615564152703604,
      "tokens_seen": 965194752
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003573620862587763,
      "loss": 2.9372,
      "theoretical_loss": 3.6615326010342892,
      "tokens_seen": 965260288
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003573520561685055,
      "loss": 2.6619,
      "theoretical_loss": 3.6615087888677076,
      "tokens_seen": 965325824
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035734202607823475,
      "loss": 2.8003,
      "theoretical_loss": 3.661484978770295,
      "tokens_seen": 965391360
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003573319959879639,
      "loss": 2.6679,
      "theoretical_loss": 3.661461170741732,
      "tokens_seen": 965456896
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003573219658976931,
      "loss": 2.7185,
      "theoretical_loss": 3.6614373647816976,
      "tokens_seen": 965522432
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003573119358074223,
      "loss": 2.7959,
      "theoretical_loss": 3.661413560889871,
      "tokens_seen": 965587968
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003573019057171515,
      "loss": 2.6025,
      "theoretical_loss": 3.6613897590659343,
      "tokens_seen": 965653504
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035729187562688066,
      "loss": 2.8268,
      "theoretical_loss": 3.661365959309566,
      "tokens_seen": 965719040
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035728184553660984,
      "loss": 2.8639,
      "theoretical_loss": 3.6613421616204462,
      "tokens_seen": 965784576
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.000357271815446339,
      "loss": 2.7539,
      "theoretical_loss": 3.6613183659982553,
      "tokens_seen": 965850112
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035726178535606826,
      "loss": 2.7141,
      "theoretical_loss": 3.6612945724426744,
      "tokens_seen": 965915648
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003572517552657974,
      "loss": 2.6672,
      "theoretical_loss": 3.661270780953383,
      "tokens_seen": 965981184
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003572417251755266,
      "loss": 2.8184,
      "theoretical_loss": 3.661246991530061,
      "tokens_seen": 966046720
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035723169508525575,
      "loss": 2.7297,
      "theoretical_loss": 3.66122320417239,
      "tokens_seen": 966112256
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.000357221664994985,
      "loss": 2.7294,
      "theoretical_loss": 3.661199418880049,
      "tokens_seen": 966177792
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035721163490471416,
      "loss": 2.9766,
      "theoretical_loss": 3.6611756356527203,
      "tokens_seen": 966243328
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035720160481444334,
      "loss": 2.7843,
      "theoretical_loss": 3.661151854490084,
      "tokens_seen": 966308864
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003571915747241725,
      "loss": 2.684,
      "theoretical_loss": 3.6611280753918205,
      "tokens_seen": 966374400
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035718154463390176,
      "loss": 2.5815,
      "theoretical_loss": 3.6611042983576105,
      "tokens_seen": 966439936
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003571715145436309,
      "loss": 2.6649,
      "theoretical_loss": 3.6610805233871355,
      "tokens_seen": 966505472
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003571614844533601,
      "loss": 2.8501,
      "theoretical_loss": 3.6610567504800766,
      "tokens_seen": 966571008
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035715145436308925,
      "loss": 2.8742,
      "theoretical_loss": 3.6610329796361136,
      "tokens_seen": 966636544
    },
    {
      "epoch": 3.02,
      "objective/train/docs_used": 1125576,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.660078525543213,
      "objective/train/theoretical_loss": 3.6610270372474445,
      "objective/train/tokens_used": 987112928,
      "theoretical_loss": 3.6610270372474445,
      "tokens_seen": 966652928
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003571414242728185,
      "loss": 2.6385,
      "theoretical_loss": 3.661009210854929,
      "tokens_seen": 966702080
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035713139418254767,
      "loss": 2.7278,
      "theoretical_loss": 3.660985444136203,
      "tokens_seen": 966767616
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035712136409227685,
      "loss": 2.8441,
      "theoretical_loss": 3.660961679479617,
      "tokens_seen": 966833152
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035711133400200603,
      "loss": 2.8036,
      "theoretical_loss": 3.660937916884853,
      "tokens_seen": 966898688
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003571013039117352,
      "loss": 2.7275,
      "theoretical_loss": 3.660914156351592,
      "tokens_seen": 966964224
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003570912738214644,
      "loss": 2.7555,
      "theoretical_loss": 3.6608903978795153,
      "tokens_seen": 967029760
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003570812437311936,
      "loss": 3.0237,
      "theoretical_loss": 3.6608666414683038,
      "tokens_seen": 967095296
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035707121364092275,
      "loss": 2.7074,
      "theoretical_loss": 3.6608428871176404,
      "tokens_seen": 967160832
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.000357061183550652,
      "loss": 2.7985,
      "theoretical_loss": 3.6608191348272063,
      "tokens_seen": 967226368
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003570511534603811,
      "loss": 2.5992,
      "theoretical_loss": 3.6607953845966827,
      "tokens_seen": 967291904
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035704112337011035,
      "loss": 2.8545,
      "theoretical_loss": 3.6607716364257525,
      "tokens_seen": 967357440
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035703109327983953,
      "loss": 2.8628,
      "theoretical_loss": 3.660747890314096,
      "tokens_seen": 967422976
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003570210631895687,
      "loss": 2.6079,
      "theoretical_loss": 3.6607241462613964,
      "tokens_seen": 967488512
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003570110330992979,
      "loss": 2.7346,
      "theoretical_loss": 3.6607004042673355,
      "tokens_seen": 967554048
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035700100300902713,
      "loss": 2.6818,
      "theoretical_loss": 3.660676664331595,
      "tokens_seen": 967619584
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035699097291875626,
      "loss": 2.7036,
      "theoretical_loss": 3.6606529264538574,
      "tokens_seen": 967685120
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003569809428284855,
      "loss": 2.7349,
      "theoretical_loss": 3.660629190633805,
      "tokens_seen": 967750656
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003569709127382146,
      "loss": 3.0038,
      "theoretical_loss": 3.66060545687112,
      "tokens_seen": 967816192
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035696088264794385,
      "loss": 2.7256,
      "theoretical_loss": 3.6605817251654846,
      "tokens_seen": 967881728
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035695085255767303,
      "loss": 2.7116,
      "theoretical_loss": 3.6605579955165815,
      "tokens_seen": 967947264
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003569408224674022,
      "loss": 2.833,
      "theoretical_loss": 3.660534267924093,
      "tokens_seen": 968012800
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003569307923771314,
      "loss": 2.8843,
      "theoretical_loss": 3.6605105423877013,
      "tokens_seen": 968078336
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003569207622868606,
      "loss": 2.5918,
      "theoretical_loss": 3.6604868189070903,
      "tokens_seen": 968143872
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035691073219658976,
      "loss": 2.6069,
      "theoretical_loss": 3.6604630974819417,
      "tokens_seen": 968209408
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.000356900702106319,
      "loss": 3.0027,
      "theoretical_loss": 3.6604393781119384,
      "tokens_seen": 968274944
    },
    {
      "epoch": 3.02,
      "objective/train/docs_used": 1126260,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6405274868011475,
      "objective/train/theoretical_loss": 3.660433448590517,
      "objective/train/tokens_used": 988751328,
      "theoretical_loss": 3.660433448590517,
      "tokens_seen": 968291328
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003568906720160481,
      "loss": 2.7458,
      "theoretical_loss": 3.6604156607967635,
      "tokens_seen": 968340480
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035688064192577736,
      "loss": 2.8098,
      "theoretical_loss": 3.6603919455361,
      "tokens_seen": 968406016
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003568706118355065,
      "loss": 2.9405,
      "theoretical_loss": 3.6603682323296307,
      "tokens_seen": 968471552
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003568605817452357,
      "loss": 2.5768,
      "theoretical_loss": 3.6603445211770387,
      "tokens_seen": 968537088
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003568505516549649,
      "loss": 2.9667,
      "theoretical_loss": 3.660320812078007,
      "tokens_seen": 968602624
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003568405215646941,
      "loss": 2.7004,
      "theoretical_loss": 3.6602971050322197,
      "tokens_seen": 968668160
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035683049147442326,
      "loss": 2.6829,
      "theoretical_loss": 3.660273400039359,
      "tokens_seen": 968733696
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003568204613841525,
      "loss": 2.6353,
      "theoretical_loss": 3.6602496970991085,
      "tokens_seen": 968799232
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003568104312938816,
      "loss": 2.5665,
      "theoretical_loss": 3.660225996211152,
      "tokens_seen": 968864768
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035680040120361086,
      "loss": 2.9588,
      "theoretical_loss": 3.6602022973751724,
      "tokens_seen": 968930304
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035679037111334,
      "loss": 2.7872,
      "theoretical_loss": 3.6601786005908545,
      "tokens_seen": 968995840
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003567803410230692,
      "loss": 2.8798,
      "theoretical_loss": 3.6601549058578806,
      "tokens_seen": 969061376
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003567703109327984,
      "loss": 2.9378,
      "theoretical_loss": 3.660131213175935,
      "tokens_seen": 969126912
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003567602808425276,
      "loss": 2.8404,
      "theoretical_loss": 3.6601075225447017,
      "tokens_seen": 969192448
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035675025075225677,
      "loss": 2.7157,
      "theoretical_loss": 3.6600838339638635,
      "tokens_seen": 969257984
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035674022066198595,
      "loss": 2.8261,
      "theoretical_loss": 3.660060147433106,
      "tokens_seen": 969323520
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035673019057171513,
      "loss": 2.8776,
      "theoretical_loss": 3.660036462952112,
      "tokens_seen": 969389056
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035672016048144436,
      "loss": 2.857,
      "theoretical_loss": 3.6600127805205656,
      "tokens_seen": 969454592
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003567101303911735,
      "loss": 2.8433,
      "theoretical_loss": 3.659989100138151,
      "tokens_seen": 969520128
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003567001003009027,
      "loss": 2.7035,
      "theoretical_loss": 3.6599654218045528,
      "tokens_seen": 969585664
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035669007021063185,
      "loss": 2.7555,
      "theoretical_loss": 3.659941745519455,
      "tokens_seen": 969651200
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003566800401203611,
      "loss": 2.4502,
      "theoretical_loss": 3.659918071282542,
      "tokens_seen": 969716736
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035667001003009027,
      "loss": 2.8731,
      "theoretical_loss": 3.6598943990934987,
      "tokens_seen": 969782272
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035665997993981945,
      "loss": 2.4678,
      "theoretical_loss": 3.6598707289520087,
      "tokens_seen": 969847808
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035664994984954863,
      "loss": 2.9327,
      "theoretical_loss": 3.659847060857757,
      "tokens_seen": 969913344
    },
    {
      "epoch": 3.02,
      "objective/train/docs_used": 1127813,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4290738105773926,
      "objective/train/theoretical_loss": 3.659841144154038,
      "objective/train/tokens_used": 990389728,
      "theoretical_loss": 3.659841144154038,
      "tokens_seen": 969929728
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035663991975927787,
      "loss": 2.5046,
      "theoretical_loss": 3.6598233948104277,
      "tokens_seen": 969978880
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.000356629889669007,
      "loss": 2.6093,
      "theoretical_loss": 3.6597997308097066,
      "tokens_seen": 970044416
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035661985957873623,
      "loss": 2.8465,
      "theoretical_loss": 3.6597760688552774,
      "tokens_seen": 970109952
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035660982948846536,
      "loss": 2.9324,
      "theoretical_loss": 3.6597524089468254,
      "tokens_seen": 970175488
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003565997993981946,
      "loss": 2.6557,
      "theoretical_loss": 3.6597287510840353,
      "tokens_seen": 970241024
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003565897693079238,
      "loss": 2.8453,
      "theoretical_loss": 3.6597050952665926,
      "tokens_seen": 970306560
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035657973921765295,
      "loss": 2.8741,
      "theoretical_loss": 3.6596814414941816,
      "tokens_seen": 970372096
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003565697091273822,
      "loss": 2.7244,
      "theoretical_loss": 3.6596577897664884,
      "tokens_seen": 970437632
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003565596790371113,
      "loss": 2.7886,
      "theoretical_loss": 3.659634140083197,
      "tokens_seen": 970503168
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035654964894684055,
      "loss": 2.6121,
      "theoretical_loss": 3.659610492443993,
      "tokens_seen": 970568704
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035653961885656973,
      "loss": 2.6921,
      "theoretical_loss": 3.659586846848563,
      "tokens_seen": 970634240
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003565295887662989,
      "loss": 2.7066,
      "theoretical_loss": 3.6595632032965897,
      "tokens_seen": 970699776
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003565195586760281,
      "loss": 2.7141,
      "theoretical_loss": 3.6595395617877613,
      "tokens_seen": 970765312
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035650952858575733,
      "loss": 2.8058,
      "theoretical_loss": 3.659515922321762,
      "tokens_seen": 970830848
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035649949849548646,
      "loss": 2.8712,
      "theoretical_loss": 3.6594922848982776,
      "tokens_seen": 970896384
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003564894684052157,
      "loss": 2.8611,
      "theoretical_loss": 3.6594686495169935,
      "tokens_seen": 970961920
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003564794383149448,
      "loss": 3.0298,
      "theoretical_loss": 3.659445016177596,
      "tokens_seen": 971027456
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035646940822467405,
      "loss": 2.7334,
      "theoretical_loss": 3.6594213848797703,
      "tokens_seen": 971092992
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035645937813440323,
      "loss": 2.6687,
      "theoretical_loss": 3.6593977556232025,
      "tokens_seen": 971158528
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003564493480441324,
      "loss": 2.9812,
      "theoretical_loss": 3.6593741284075794,
      "tokens_seen": 971224064
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003564393179538616,
      "loss": 2.7254,
      "theoretical_loss": 3.6593505032325853,
      "tokens_seen": 971289600
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003564292878635908,
      "loss": 2.6469,
      "theoretical_loss": 3.659326880097908,
      "tokens_seen": 971355136
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035641925777331996,
      "loss": 2.77,
      "theoretical_loss": 3.6593032590032326,
      "tokens_seen": 971420672
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003564092276830492,
      "loss": 2.8566,
      "theoretical_loss": 3.6592796399482452,
      "tokens_seen": 971486208
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003563991975927783,
      "loss": 2.9147,
      "theoretical_loss": 3.659256022932633,
      "tokens_seen": 971551744
    },
    {
      "epoch": 3.02,
      "objective/train/docs_used": 1128556,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.751553773880005,
      "objective/train/theoretical_loss": 3.6592501189973454,
      "objective/train/tokens_used": 992028128,
      "theoretical_loss": 3.6592501189973454,
      "tokens_seen": 971568128
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035638916750250756,
      "loss": 3.0277,
      "theoretical_loss": 3.6592324079560816,
      "tokens_seen": 971617280
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003563791374122367,
      "loss": 2.8446,
      "theoretical_loss": 3.6592087950182783,
      "tokens_seen": 971682816
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003563691073219659,
      "loss": 2.7536,
      "theoretical_loss": 3.6591851841189085,
      "tokens_seen": 971748352
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003563590772316951,
      "loss": 2.8624,
      "theoretical_loss": 3.6591615752576594,
      "tokens_seen": 971813888
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003563490471414243,
      "loss": 2.779,
      "theoretical_loss": 3.659137968434217,
      "tokens_seen": 971879424
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035633901705115346,
      "loss": 2.836,
      "theoretical_loss": 3.6591143636482695,
      "tokens_seen": 971944960
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003563289869608827,
      "loss": 2.9205,
      "theoretical_loss": 3.659090760899502,
      "tokens_seen": 972010496
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003563189568706118,
      "loss": 2.655,
      "theoretical_loss": 3.659067160187602,
      "tokens_seen": 972076032
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035630892678034106,
      "loss": 2.7509,
      "theoretical_loss": 3.6590435615122567,
      "tokens_seen": 972141568
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003562988966900702,
      "loss": 2.7927,
      "theoretical_loss": 3.6590199648731527,
      "tokens_seen": 972207104
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003562888665997994,
      "loss": 2.8891,
      "theoretical_loss": 3.6589963702699775,
      "tokens_seen": 972272640
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003562788365095286,
      "loss": 2.9476,
      "theoretical_loss": 3.658972777702418,
      "tokens_seen": 972338176
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003562688064192578,
      "loss": 2.8627,
      "theoretical_loss": 3.658949187170161,
      "tokens_seen": 972403712
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035625877632898697,
      "loss": 2.7483,
      "theoretical_loss": 3.6589255986728944,
      "tokens_seen": 972469248
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035624874623871615,
      "loss": 2.5822,
      "theoretical_loss": 3.658902012210305,
      "tokens_seen": 972534784
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035623871614844533,
      "loss": 2.962,
      "theoretical_loss": 3.6588784277820805,
      "tokens_seen": 972600320
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.00035622868605817456,
      "loss": 2.7324,
      "theoretical_loss": 3.658854845387908,
      "tokens_seen": 972665856
    },
    {
      "epoch": 3.02,
      "learning_rate": 0.0003562186559679037,
      "loss": 2.7687,
      "theoretical_loss": 3.6588312650274757,
      "tokens_seen": 972731392
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003562086258776329,
      "loss": 2.4566,
      "theoretical_loss": 3.658807686700471,
      "tokens_seen": 972796928
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035619859578736205,
      "loss": 2.6661,
      "theoretical_loss": 3.6587841104065815,
      "tokens_seen": 972862464
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003561885656970913,
      "loss": 2.6573,
      "theoretical_loss": 3.6587605361454942,
      "tokens_seen": 972928000
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035617853560682047,
      "loss": 2.6695,
      "theoretical_loss": 3.658736963916898,
      "tokens_seen": 972993536
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035616850551654965,
      "loss": 2.8766,
      "theoretical_loss": 3.6587133937204808,
      "tokens_seen": 973059072
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035615847542627883,
      "loss": 2.6962,
      "theoretical_loss": 3.6586898255559293,
      "tokens_seen": 973124608
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035614844533600807,
      "loss": 2.7565,
      "theoretical_loss": 3.6586662594229327,
      "tokens_seen": 973190144
    },
    {
      "epoch": 3.03,
      "objective/train/docs_used": 1129663,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.72713303565979,
      "objective/train/theoretical_loss": 3.6586603682070775,
      "objective/train/tokens_used": 993666528,
      "theoretical_loss": 3.6586603682070775,
      "tokens_seen": 973206528
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003561384152457372,
      "loss": 2.8701,
      "theoretical_loss": 3.658642695321179,
      "tokens_seen": 973255680
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035612838515546643,
      "loss": 2.7955,
      "theoretical_loss": 3.6586191332503555,
      "tokens_seen": 973321216
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035611835506519556,
      "loss": 2.8356,
      "theoretical_loss": 3.6585955732101514,
      "tokens_seen": 973386752
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003561083249749248,
      "loss": 2.7941,
      "theoretical_loss": 3.6585720152002548,
      "tokens_seen": 973452288
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035609829488465397,
      "loss": 2.9232,
      "theoretical_loss": 3.658548459220354,
      "tokens_seen": 973517824
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035608826479438315,
      "loss": 2.6519,
      "theoretical_loss": 3.658524905270137,
      "tokens_seen": 973583360
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035607823470411233,
      "loss": 2.7978,
      "theoretical_loss": 3.6585013533492927,
      "tokens_seen": 973648896
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003560682046138415,
      "loss": 2.7026,
      "theoretical_loss": 3.65847780345751,
      "tokens_seen": 973714432
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003560581745235707,
      "loss": 2.689,
      "theoretical_loss": 3.6584542555944766,
      "tokens_seen": 973779968
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035604814443329993,
      "loss": 3.099,
      "theoretical_loss": 3.658430709759882,
      "tokens_seen": 973845504
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035603811434302906,
      "loss": 2.7942,
      "theoretical_loss": 3.658407165953415,
      "tokens_seen": 973911040
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003560280842527583,
      "loss": 2.629,
      "theoretical_loss": 3.6583836241747636,
      "tokens_seen": 973976576
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003560180541624874,
      "loss": 2.7733,
      "theoretical_loss": 3.6583600844236175,
      "tokens_seen": 974042112
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035600802407221666,
      "loss": 2.8486,
      "theoretical_loss": 3.658336546699666,
      "tokens_seen": 974107648
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035599799398194584,
      "loss": 2.648,
      "theoretical_loss": 3.658313011002597,
      "tokens_seen": 974173184
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.000355987963891675,
      "loss": 2.8866,
      "theoretical_loss": 3.658289477332101,
      "tokens_seen": 974238720
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003559779338014042,
      "loss": 2.6934,
      "theoretical_loss": 3.658265945687866,
      "tokens_seen": 974304256
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035596790371113344,
      "loss": 2.653,
      "theoretical_loss": 3.6582424160695814,
      "tokens_seen": 974369792
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035595787362086256,
      "loss": 2.7027,
      "theoretical_loss": 3.658218888476937,
      "tokens_seen": 974435328
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003559478435305918,
      "loss": 2.7906,
      "theoretical_loss": 3.6581953629096224,
      "tokens_seen": 974500864
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003559378134403209,
      "loss": 2.6374,
      "theoretical_loss": 3.658171839367327,
      "tokens_seen": 974566400
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035592778335005016,
      "loss": 2.64,
      "theoretical_loss": 3.6581483178497396,
      "tokens_seen": 974631936
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035591775325977934,
      "loss": 2.6977,
      "theoretical_loss": 3.6581247983565506,
      "tokens_seen": 974697472
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003559077231695085,
      "loss": 2.9227,
      "theoretical_loss": 3.6581012808874487,
      "tokens_seen": 974763008
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003558976930792377,
      "loss": 2.7544,
      "theoretical_loss": 3.6580777654421244,
      "tokens_seen": 974828544
    },
    {
      "epoch": 3.03,
      "objective/train/docs_used": 1130250,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.891390562057495,
      "objective/train/theoretical_loss": 3.6580718868969724,
      "objective/train/tokens_used": 995304928,
      "theoretical_loss": 3.6580718868969724,
      "tokens_seen": 974844928
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003558876629889669,
      "loss": 2.7537,
      "theoretical_loss": 3.6580542520202677,
      "tokens_seen": 974894080
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035587763289869607,
      "loss": 2.8645,
      "theoretical_loss": 3.658030740621568,
      "tokens_seen": 974959616
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003558676028084253,
      "loss": 2.6031,
      "theoretical_loss": 3.6580072312457155,
      "tokens_seen": 975025152
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035585757271815443,
      "loss": 2.9164,
      "theoretical_loss": 3.6579837238924,
      "tokens_seen": 975090688
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035584754262788366,
      "loss": 2.5411,
      "theoretical_loss": 3.657960218561312,
      "tokens_seen": 975156224
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003558375125376129,
      "loss": 2.7479,
      "theoretical_loss": 3.6579367152521414,
      "tokens_seen": 975221760
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.000355827482447342,
      "loss": 2.8431,
      "theoretical_loss": 3.6579132139645774,
      "tokens_seen": 975287296
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035581745235707126,
      "loss": 2.8175,
      "theoretical_loss": 3.657889714698312,
      "tokens_seen": 975352832
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003558074222668004,
      "loss": 2.8773,
      "theoretical_loss": 3.6578662174530345,
      "tokens_seen": 975418368
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003557973921765296,
      "loss": 2.6768,
      "theoretical_loss": 3.657842722228436,
      "tokens_seen": 975483904
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003557873620862588,
      "loss": 2.6195,
      "theoretical_loss": 3.6578192290242066,
      "tokens_seen": 975549440
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.000355777331995988,
      "loss": 2.7406,
      "theoretical_loss": 3.6577957378400368,
      "tokens_seen": 975614976
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035576730190571717,
      "loss": 2.7412,
      "theoretical_loss": 3.657772248675617,
      "tokens_seen": 975680512
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035575727181544635,
      "loss": 2.6126,
      "theoretical_loss": 3.6577487615306383,
      "tokens_seen": 975746048
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035574724172517553,
      "loss": 2.6626,
      "theoretical_loss": 3.6577252764047916,
      "tokens_seen": 975811584
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035573721163490476,
      "loss": 3.0388,
      "theoretical_loss": 3.6577017932977673,
      "tokens_seen": 975877120
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003557271815446339,
      "loss": 2.8096,
      "theoretical_loss": 3.6576783122092564,
      "tokens_seen": 975942656
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003557171514543631,
      "loss": 2.8943,
      "theoretical_loss": 3.6576548331389507,
      "tokens_seen": 976008192
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035570712136409225,
      "loss": 2.8692,
      "theoretical_loss": 3.6576313560865397,
      "tokens_seen": 976073728
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003556970912738215,
      "loss": 2.913,
      "theoretical_loss": 3.6576078810517156,
      "tokens_seen": 976139264
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035568706118355067,
      "loss": 2.7489,
      "theoretical_loss": 3.657584408034169,
      "tokens_seen": 976204800
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035567703109327985,
      "loss": 2.5959,
      "theoretical_loss": 3.6575609370335913,
      "tokens_seen": 976270336
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035566700100300903,
      "loss": 2.962,
      "theoretical_loss": 3.657537468049674,
      "tokens_seen": 976335872
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035565697091273827,
      "loss": 2.7383,
      "theoretical_loss": 3.657514001082109,
      "tokens_seen": 976401408
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003556469408224674,
      "loss": 3.0656,
      "theoretical_loss": 3.6574905361305867,
      "tokens_seen": 976466944
    },
    {
      "epoch": 3.03,
      "objective/train/docs_used": 1131717,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6520185470581055,
      "objective/train/theoretical_loss": 3.657484670207677,
      "objective/train/tokens_used": 996943328,
      "theoretical_loss": 3.657484670207677,
      "tokens_seen": 976483328
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035563691073219663,
      "loss": 2.8189,
      "theoretical_loss": 3.657467073194799,
      "tokens_seen": 976532480
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035562688064192576,
      "loss": 2.6232,
      "theoretical_loss": 3.657443612274438,
      "tokens_seen": 976598016
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.000355616850551655,
      "loss": 2.6579,
      "theoretical_loss": 3.657420153369194,
      "tokens_seen": 976663552
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035560682046138417,
      "loss": 2.7898,
      "theoretical_loss": 3.6573966964787603,
      "tokens_seen": 976729088
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035559679037111335,
      "loss": 2.707,
      "theoretical_loss": 3.6573732416028277,
      "tokens_seen": 976794624
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035558676028084253,
      "loss": 3.012,
      "theoretical_loss": 3.6573497887410884,
      "tokens_seen": 976860160
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003555767301905717,
      "loss": 2.7479,
      "theoretical_loss": 3.6573263378932346,
      "tokens_seen": 976925696
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003555667001003009,
      "loss": 2.8411,
      "theoretical_loss": 3.6573028890589576,
      "tokens_seen": 976991232
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035555667001003013,
      "loss": 2.8103,
      "theoretical_loss": 3.6572794422379493,
      "tokens_seen": 977056768
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035554663991975926,
      "loss": 2.8471,
      "theoretical_loss": 3.6572559974299033,
      "tokens_seen": 977122304
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003555366098294885,
      "loss": 2.6952,
      "theoretical_loss": 3.6572325546345104,
      "tokens_seen": 977187840
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003555265797392176,
      "loss": 2.4342,
      "theoretical_loss": 3.657209113851463,
      "tokens_seen": 977253376
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035551654964894686,
      "loss": 2.7549,
      "theoretical_loss": 3.657185675080454,
      "tokens_seen": 977318912
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035550651955867604,
      "loss": 2.7508,
      "theoretical_loss": 3.657162238321175,
      "tokens_seen": 977384448
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003554964894684052,
      "loss": 2.8389,
      "theoretical_loss": 3.6571388035733197,
      "tokens_seen": 977449984
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003554864593781344,
      "loss": 3.0033,
      "theoretical_loss": 3.657115370836579,
      "tokens_seen": 977515520
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035547642928786364,
      "loss": 2.782,
      "theoretical_loss": 3.657091940110647,
      "tokens_seen": 977581056
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035546639919759276,
      "loss": 2.7965,
      "theoretical_loss": 3.6570685113952153,
      "tokens_seen": 977646592
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.000355456369107322,
      "loss": 2.6375,
      "theoretical_loss": 3.6570450846899774,
      "tokens_seen": 977712128
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003554463390170511,
      "loss": 2.6154,
      "theoretical_loss": 3.6570216599946255,
      "tokens_seen": 977777664
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035543630892678036,
      "loss": 2.8124,
      "theoretical_loss": 3.656998237308853,
      "tokens_seen": 977843200
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035542627883650954,
      "loss": 2.8307,
      "theoretical_loss": 3.6569748166323524,
      "tokens_seen": 977908736
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003554162487462387,
      "loss": 2.6667,
      "theoretical_loss": 3.656951397964817,
      "tokens_seen": 977974272
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003554062186559679,
      "loss": 2.8426,
      "theoretical_loss": 3.6569279813059397,
      "tokens_seen": 978039808
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003553961885656971,
      "loss": 2.6248,
      "theoretical_loss": 3.656904566655413,
      "tokens_seen": 978105344
    },
    {
      "epoch": 3.03,
      "objective/train/docs_used": 1132115,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.2902779579162598,
      "objective/train/theoretical_loss": 3.6568987133065507,
      "objective/train/tokens_used": 998581728,
      "theoretical_loss": 3.6568987133065507,
      "tokens_seen": 978121728
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035538615847542627,
      "loss": 2.984,
      "theoretical_loss": 3.6568811540129316,
      "tokens_seen": 978170880
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003553761283851555,
      "loss": 2.8201,
      "theoretical_loss": 3.656857743378188,
      "tokens_seen": 978236416
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035536609829488463,
      "loss": 2.8302,
      "theoretical_loss": 3.656834334750875,
      "tokens_seen": 978301952
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035535606820461386,
      "loss": 2.6622,
      "theoretical_loss": 3.6568109281306866,
      "tokens_seen": 978367488
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035534603811434304,
      "loss": 2.8905,
      "theoretical_loss": 3.656787523517316,
      "tokens_seen": 978433024
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003553360080240722,
      "loss": 2.8052,
      "theoretical_loss": 3.656764120910457,
      "tokens_seen": 978498560
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003553259779338014,
      "loss": 2.9525,
      "theoretical_loss": 3.6567407203098035,
      "tokens_seen": 978564096
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003553159478435306,
      "loss": 2.6697,
      "theoretical_loss": 3.6567173217150484,
      "tokens_seen": 978629632
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035530591775325977,
      "loss": 2.8138,
      "theoretical_loss": 3.656693925125886,
      "tokens_seen": 978695168
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.000355295887662989,
      "loss": 2.7755,
      "theoretical_loss": 3.6566705305420095,
      "tokens_seen": 978760704
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035528585757271813,
      "loss": 3.0818,
      "theoretical_loss": 3.656647137963114,
      "tokens_seen": 978826240
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035527582748244737,
      "loss": 2.766,
      "theoretical_loss": 3.656623747388892,
      "tokens_seen": 978891776
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003552657973921765,
      "loss": 3.0058,
      "theoretical_loss": 3.6566003588190386,
      "tokens_seen": 978957312
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035525576730190573,
      "loss": 2.8212,
      "theoretical_loss": 3.656576972253247,
      "tokens_seen": 979022848
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003552457372116349,
      "loss": 2.9931,
      "theoretical_loss": 3.656553587691212,
      "tokens_seen": 979088384
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003552357071213641,
      "loss": 2.9403,
      "theoretical_loss": 3.6565302051326274,
      "tokens_seen": 979153920
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035522567703109327,
      "loss": 2.7488,
      "theoretical_loss": 3.656506824577188,
      "tokens_seen": 979219456
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035521564694082245,
      "loss": 2.9203,
      "theoretical_loss": 3.6564834460245876,
      "tokens_seen": 979284992
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035520561685055163,
      "loss": 2.9973,
      "theoretical_loss": 3.6564600694745213,
      "tokens_seen": 979350528
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035519558676028087,
      "loss": 2.8523,
      "theoretical_loss": 3.6564366949266827,
      "tokens_seen": 979416064
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035518555667001,
      "loss": 2.8387,
      "theoretical_loss": 3.6564133223807667,
      "tokens_seen": 979481600
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035517552657973923,
      "loss": 2.8515,
      "theoretical_loss": 3.656389951836468,
      "tokens_seen": 979547136
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003551654964894684,
      "loss": 2.7526,
      "theoretical_loss": 3.6563665832934813,
      "tokens_seen": 979612672
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003551554663991976,
      "loss": 2.8189,
      "theoretical_loss": 3.6563432167515018,
      "tokens_seen": 979678208
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003551454363089268,
      "loss": 2.7372,
      "theoretical_loss": 3.656319852210223,
      "tokens_seen": 979743744
    },
    {
      "epoch": 3.03,
      "objective/train/docs_used": 1133253,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4291090965270996,
      "objective/train/theoretical_loss": 3.6563140113874777,
      "objective/train/tokens_used": 1000220128,
      "theoretical_loss": 3.6563140113874777,
      "tokens_seen": 979760128
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035513540621865596,
      "loss": 2.6382,
      "theoretical_loss": 3.6562964896693417,
      "tokens_seen": 979809280
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035512537612838514,
      "loss": 2.7368,
      "theoretical_loss": 3.656273129128551,
      "tokens_seen": 979874816
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035511534603811437,
      "loss": 2.7358,
      "theoretical_loss": 3.6562497705875465,
      "tokens_seen": 979940352
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003551053159478435,
      "loss": 2.9211,
      "theoretical_loss": 3.6562264140460243,
      "tokens_seen": 980005888
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035509528585757273,
      "loss": 2.8374,
      "theoretical_loss": 3.656203059503678,
      "tokens_seen": 980071424
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003550852557673019,
      "loss": 2.5918,
      "theoretical_loss": 3.6561797069602036,
      "tokens_seen": 980136960
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003550752256770311,
      "loss": 2.7636,
      "theoretical_loss": 3.6561563564152966,
      "tokens_seen": 980202496
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035506519558676033,
      "loss": 2.8558,
      "theoretical_loss": 3.656133007868652,
      "tokens_seen": 980268032
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035505516549648946,
      "loss": 2.6737,
      "theoretical_loss": 3.6561096613199653,
      "tokens_seen": 980333568
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003550451354062187,
      "loss": 2.7461,
      "theoretical_loss": 3.656086316768932,
      "tokens_seen": 980399104
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003550351053159478,
      "loss": 2.9711,
      "theoretical_loss": 3.6560629742152475,
      "tokens_seen": 980464640
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035502507522567706,
      "loss": 2.765,
      "theoretical_loss": 3.656039633658607,
      "tokens_seen": 980530176
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035501504513540624,
      "loss": 2.7839,
      "theoretical_loss": 3.6560162950987083,
      "tokens_seen": 980595712
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003550050150451354,
      "loss": 2.7088,
      "theoretical_loss": 3.6559929585352444,
      "tokens_seen": 980661248
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003549949849548646,
      "loss": 2.9667,
      "theoretical_loss": 3.6559696239679127,
      "tokens_seen": 980726784
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035498495486459384,
      "loss": 2.9507,
      "theoretical_loss": 3.6559462913964085,
      "tokens_seen": 980792320
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035497492477432296,
      "loss": 2.906,
      "theoretical_loss": 3.655922960820428,
      "tokens_seen": 980857856
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003549648946840522,
      "loss": 2.8641,
      "theoretical_loss": 3.6558996322396666,
      "tokens_seen": 980923392
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003549548645937813,
      "loss": 2.6324,
      "theoretical_loss": 3.655876305653822,
      "tokens_seen": 980988928
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035494483450351056,
      "loss": 2.8169,
      "theoretical_loss": 3.655852981062589,
      "tokens_seen": 981054464
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035493480441323974,
      "loss": 2.8943,
      "theoretical_loss": 3.6558296584656635,
      "tokens_seen": 981120000
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003549247743229689,
      "loss": 2.7909,
      "theoretical_loss": 3.655806337862743,
      "tokens_seen": 981185536
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003549147442326981,
      "loss": 2.7996,
      "theoretical_loss": 3.655783019253523,
      "tokens_seen": 981251072
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003549047141424273,
      "loss": 2.625,
      "theoretical_loss": 3.6557597026377,
      "tokens_seen": 981316608
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035489468405215647,
      "loss": 2.9213,
      "theoretical_loss": 3.655736388014971,
      "tokens_seen": 981382144
    },
    {
      "epoch": 3.03,
      "objective/train/docs_used": 1133558,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7345311641693115,
      "objective/train/theoretical_loss": 3.655730559670674,
      "objective/train/tokens_used": 1001858528,
      "theoretical_loss": 3.655730559670674,
      "tokens_seen": 981398528
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003548846539618857,
      "loss": 2.7113,
      "theoretical_loss": 3.655713075385032,
      "tokens_seen": 981447680
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035487462387161483,
      "loss": 2.9233,
      "theoretical_loss": 3.6556897647475797,
      "tokens_seen": 981513216
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035486459378134406,
      "loss": 2.635,
      "theoretical_loss": 3.6556664561023116,
      "tokens_seen": 981578752
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035485456369107324,
      "loss": 2.9342,
      "theoretical_loss": 3.655643149448923,
      "tokens_seen": 981644288
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003548445336008024,
      "loss": 2.739,
      "theoretical_loss": 3.655619844787112,
      "tokens_seen": 981709824
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003548345035105316,
      "loss": 2.8912,
      "theoretical_loss": 3.6555965421165744,
      "tokens_seen": 981775360
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003548244734202608,
      "loss": 3.0486,
      "theoretical_loss": 3.6555732414370077,
      "tokens_seen": 981840896
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035481444332998997,
      "loss": 3.1694,
      "theoretical_loss": 3.6555499427481095,
      "tokens_seen": 981906432
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003548044132397192,
      "loss": 3.1176,
      "theoretical_loss": 3.655526646049576,
      "tokens_seen": 981971968
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035479438314944833,
      "loss": 2.8801,
      "theoretical_loss": 3.655503351341105,
      "tokens_seen": 982037504
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035478435305917757,
      "loss": 2.9109,
      "theoretical_loss": 3.655480058622393,
      "tokens_seen": 982103040
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003547743229689067,
      "loss": 2.9983,
      "theoretical_loss": 3.6554567678931376,
      "tokens_seen": 982168576
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035476429287863593,
      "loss": 2.7778,
      "theoretical_loss": 3.6554334791530363,
      "tokens_seen": 982234112
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003547542627883651,
      "loss": 2.7951,
      "theoretical_loss": 3.6554101924017863,
      "tokens_seen": 982299648
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003547442326980943,
      "loss": 2.9124,
      "theoretical_loss": 3.6553869076390857,
      "tokens_seen": 982365184
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035473420260782347,
      "loss": 2.6658,
      "theoretical_loss": 3.6553636248646306,
      "tokens_seen": 982430720
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035472417251755265,
      "loss": 3.0599,
      "theoretical_loss": 3.6553403440781205,
      "tokens_seen": 982496256
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035471414242728183,
      "loss": 2.8763,
      "theoretical_loss": 3.6553170652792515,
      "tokens_seen": 982561792
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035470411233701107,
      "loss": 2.8212,
      "theoretical_loss": 3.655293788467722,
      "tokens_seen": 982627328
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003546940822467402,
      "loss": 2.677,
      "theoretical_loss": 3.65527051364323,
      "tokens_seen": 982692864
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035468405215646943,
      "loss": 2.8286,
      "theoretical_loss": 3.655247240805473,
      "tokens_seen": 982758400
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003546740220661986,
      "loss": 2.8833,
      "theoretical_loss": 3.6552239699541493,
      "tokens_seen": 982823936
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003546639919759278,
      "loss": 2.8848,
      "theoretical_loss": 3.6552007010889564,
      "tokens_seen": 982889472
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.000354653961885657,
      "loss": 2.9692,
      "theoretical_loss": 3.6551774342095933,
      "tokens_seen": 982955008
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035464393179538616,
      "loss": 2.8621,
      "theoretical_loss": 3.6551541693157574,
      "tokens_seen": 983020544
    },
    {
      "debugging/Self-BLEU-5": 0.39763851797645977,
      "debugging/distinct-1-grams": 0.7825305380824156,
      "debugging/distinct-2-grams": 0.9587387596654564,
      "debugging/entropy-1-grams": 5.676356416295845,
      "debugging/entropy-2-grams": 6.320223384568614,
      "debugging/length": 540.7142857142857,
      "debugging/num_segments": 7,
      "debugging/score": 0.006025538542352743,
      "debugging/score_std": 0.005169474872870312,
      "epoch": 3.03,
      "objective/train/docs_used": 1133558,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.444734573364258,
      "objective/train/theoretical_loss": 3.6551483534025015,
      "objective/train/tokens_used": 1003496928,
      "theoretical_loss": 3.6551483534025015,
      "tokens_seen": 983036928
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035463390170511534,
      "loss": 2.7331,
      "theoretical_loss": 3.6551309064071464,
      "tokens_seen": 983086080
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035462387161484457,
      "loss": 2.7984,
      "theoretical_loss": 3.6551076454834597,
      "tokens_seen": 983151616
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003546138415245737,
      "loss": 2.6689,
      "theoretical_loss": 3.655084386544395,
      "tokens_seen": 983217152
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035460381143430293,
      "loss": 2.9601,
      "theoretical_loss": 3.655061129589651,
      "tokens_seen": 983282688
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035459378134403206,
      "loss": 2.8545,
      "theoretical_loss": 3.655037874618926,
      "tokens_seen": 983348224
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003545837512537613,
      "loss": 2.8236,
      "theoretical_loss": 3.655014621631918,
      "tokens_seen": 983413760
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003545737211634905,
      "loss": 2.8477,
      "theoretical_loss": 3.654991370628327,
      "tokens_seen": 983479296
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035456369107321966,
      "loss": 3.001,
      "theoretical_loss": 3.6549681216078502,
      "tokens_seen": 983544832
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035455366098294884,
      "loss": 2.8405,
      "theoretical_loss": 3.654944874570187,
      "tokens_seen": 983610368
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.000354543630892678,
      "loss": 2.7946,
      "theoretical_loss": 3.6549216295150364,
      "tokens_seen": 983675904
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003545336008024072,
      "loss": 2.7789,
      "theoretical_loss": 3.6548983864420967,
      "tokens_seen": 983741440
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035452357071213644,
      "loss": 2.9531,
      "theoretical_loss": 3.6548751453510677,
      "tokens_seen": 983806976
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035451354062186556,
      "loss": 2.8242,
      "theoretical_loss": 3.654851906241647,
      "tokens_seen": 983872512
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003545035105315948,
      "loss": 2.8175,
      "theoretical_loss": 3.6548286691135354,
      "tokens_seen": 983938048
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.000354493480441324,
      "loss": 2.9619,
      "theoretical_loss": 3.6548054339664304,
      "tokens_seen": 984003584
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035448345035105316,
      "loss": 2.8656,
      "theoretical_loss": 3.6547822008000326,
      "tokens_seen": 984069120
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035447342026078234,
      "loss": 2.7168,
      "theoretical_loss": 3.654758969614041,
      "tokens_seen": 984134656
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003544633901705115,
      "loss": 2.741,
      "theoretical_loss": 3.654735740408153,
      "tokens_seen": 984200192
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003544533600802407,
      "loss": 3.0168,
      "theoretical_loss": 3.6547125131820706,
      "tokens_seen": 984265728
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035444332998996994,
      "loss": 3.0056,
      "theoretical_loss": 3.654689287935492,
      "tokens_seen": 984331264
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035443329989969907,
      "loss": 2.8746,
      "theoretical_loss": 3.654666064668117,
      "tokens_seen": 984396800
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003544232698094283,
      "loss": 3.1804,
      "theoretical_loss": 3.654642843379645,
      "tokens_seen": 984462336
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035441323971915743,
      "loss": 3.1955,
      "theoretical_loss": 3.6546196240697757,
      "tokens_seen": 984527872
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035440320962888667,
      "loss": 2.9598,
      "theoretical_loss": 3.6545964067382086,
      "tokens_seen": 984593408
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035439317953861585,
      "loss": 2.938,
      "theoretical_loss": 3.654573191384644,
      "tokens_seen": 984658944
    },
    {
      "epoch": 3.03,
      "objective/train/docs_used": 1134314,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.677119493484497,
      "objective/train/theoretical_loss": 3.6545673878552805,
      "objective/train/tokens_used": 1005135328,
      "theoretical_loss": 3.6545673878552805,
      "tokens_seen": 984675328
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035438314944834503,
      "loss": 3.1099,
      "theoretical_loss": 3.654549978008782,
      "tokens_seen": 984724480
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003543731193580742,
      "loss": 3.0527,
      "theoretical_loss": 3.6545267666103207,
      "tokens_seen": 984790016
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035436308926780344,
      "loss": 3.0231,
      "theoretical_loss": 3.6545035571889626,
      "tokens_seen": 984855552
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003543530591775326,
      "loss": 2.8865,
      "theoretical_loss": 3.6544803497444063,
      "tokens_seen": 984921088
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003543430290872618,
      "loss": 2.7854,
      "theoretical_loss": 3.6544571442763516,
      "tokens_seen": 984986624
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.000354332998996991,
      "loss": 2.712,
      "theoretical_loss": 3.6544339407844997,
      "tokens_seen": 985052160
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035432296890672017,
      "loss": 2.8541,
      "theoretical_loss": 3.6544107392685503,
      "tokens_seen": 985117696
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003543129388164494,
      "loss": 2.9849,
      "theoretical_loss": 3.6543875397282037,
      "tokens_seen": 985183232
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035430290872617853,
      "loss": 2.9454,
      "theoretical_loss": 3.65436434216316,
      "tokens_seen": 985248768
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035429287863590777,
      "loss": 3.1817,
      "theoretical_loss": 3.654341146573121,
      "tokens_seen": 985314304
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003542828485456369,
      "loss": 2.9778,
      "theoretical_loss": 3.6543179529577854,
      "tokens_seen": 985379840
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035427281845536613,
      "loss": 2.9435,
      "theoretical_loss": 3.654294761316855,
      "tokens_seen": 985445376
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003542627883650953,
      "loss": 2.8006,
      "theoretical_loss": 3.6542715716500296,
      "tokens_seen": 985510912
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003542527582748245,
      "loss": 2.9925,
      "theoretical_loss": 3.6542483839570106,
      "tokens_seen": 985576448
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035424272818455367,
      "loss": 3.0758,
      "theoretical_loss": 3.6542251982374983,
      "tokens_seen": 985641984
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035423269809428285,
      "loss": 2.8809,
      "theoretical_loss": 3.6542020144911938,
      "tokens_seen": 985707520
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035422266800401203,
      "loss": 2.9749,
      "theoretical_loss": 3.6541788327177978,
      "tokens_seen": 985773056
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035421263791374127,
      "loss": 3.0246,
      "theoretical_loss": 3.6541556529170114,
      "tokens_seen": 985838592
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003542026078234704,
      "loss": 3.0333,
      "theoretical_loss": 3.6541324750885362,
      "tokens_seen": 985904128
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035419257773319963,
      "loss": 2.9751,
      "theoretical_loss": 3.654109299232072,
      "tokens_seen": 985969664
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003541825476429288,
      "loss": 3.002,
      "theoretical_loss": 3.6540861253473205,
      "tokens_seen": 986035200
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.000354172517552658,
      "loss": 3.0969,
      "theoretical_loss": 3.654062953433984,
      "tokens_seen": 986100736
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003541624874623872,
      "loss": 3.0918,
      "theoretical_loss": 3.654039783491762,
      "tokens_seen": 986166272
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035415245737211636,
      "loss": 2.738,
      "theoretical_loss": 3.654016615520357,
      "tokens_seen": 986231808
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035414242728184554,
      "loss": 2.915,
      "theoretical_loss": 3.65399344951947,
      "tokens_seen": 986297344
    },
    {
      "epoch": 3.03,
      "objective/train/docs_used": 1135754,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.907013416290283,
      "objective/train/theoretical_loss": 3.6539876583271065,
      "objective/train/tokens_used": 1006773728,
      "theoretical_loss": 3.6539876583271065,
      "tokens_seen": 986313728
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035413239719157477,
      "loss": 2.919,
      "theoretical_loss": 3.6539702854888025,
      "tokens_seen": 986362880
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003541223671013039,
      "loss": 2.9234,
      "theoretical_loss": 3.6539471234280567,
      "tokens_seen": 986428416
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035411233701103314,
      "loss": 2.8228,
      "theoretical_loss": 3.6539239633369327,
      "tokens_seen": 986493952
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035410230692076226,
      "loss": 2.9489,
      "theoretical_loss": 3.653900805215134,
      "tokens_seen": 986559488
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003540922768304915,
      "loss": 3.0417,
      "theoretical_loss": 3.6538776490623612,
      "tokens_seen": 986625024
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003540822467402207,
      "loss": 2.7938,
      "theoretical_loss": 3.6538544948783165,
      "tokens_seen": 986690560
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035407221664994986,
      "loss": 2.9534,
      "theoretical_loss": 3.6538313426627016,
      "tokens_seen": 986756096
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035406218655967904,
      "loss": 2.8762,
      "theoretical_loss": 3.6538081924152186,
      "tokens_seen": 986821632
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003540521564694082,
      "loss": 2.8072,
      "theoretical_loss": 3.653785044135569,
      "tokens_seen": 986887168
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003540421263791374,
      "loss": 3.145,
      "theoretical_loss": 3.653761897823456,
      "tokens_seen": 986952704
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035403209628886664,
      "loss": 3.0583,
      "theoretical_loss": 3.6537387534785806,
      "tokens_seen": 987018240
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035402206619859577,
      "loss": 3.0636,
      "theoretical_loss": 3.6537156111006457,
      "tokens_seen": 987083776
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.000354012036108325,
      "loss": 2.9563,
      "theoretical_loss": 3.6536924706893528,
      "tokens_seen": 987149312
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003540020060180542,
      "loss": 3.0312,
      "theoretical_loss": 3.6536693322444047,
      "tokens_seen": 987214848
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035399197592778336,
      "loss": 3.0896,
      "theoretical_loss": 3.6536461957655044,
      "tokens_seen": 987280384
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035398194583751254,
      "loss": 2.9546,
      "theoretical_loss": 3.653623061252353,
      "tokens_seen": 987345920
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003539719157472417,
      "loss": 2.9942,
      "theoretical_loss": 3.6535999287046543,
      "tokens_seen": 987411456
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003539618856569709,
      "loss": 3.0153,
      "theoretical_loss": 3.65357679812211,
      "tokens_seen": 987476992
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035395185556670014,
      "loss": 2.7674,
      "theoretical_loss": 3.6535536695044235,
      "tokens_seen": 987542528
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035394182547642927,
      "loss": 3.0445,
      "theoretical_loss": 3.653530542851297,
      "tokens_seen": 987608064
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003539317953861585,
      "loss": 3.0266,
      "theoretical_loss": 3.6535074181624334,
      "tokens_seen": 987673600
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035392176529588763,
      "loss": 2.6688,
      "theoretical_loss": 3.6534842954375355,
      "tokens_seen": 987739136
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035391173520561687,
      "loss": 2.8888,
      "theoretical_loss": 3.6534611746763064,
      "tokens_seen": 987804672
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035390170511534605,
      "loss": 3.0519,
      "theoretical_loss": 3.653438055878449,
      "tokens_seen": 987870208
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035389167502507523,
      "loss": 2.8346,
      "theoretical_loss": 3.6534149390436657,
      "tokens_seen": 987935744
    },
    {
      "epoch": 3.03,
      "objective/train/docs_used": 1136387,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.134091854095459,
      "objective/train/theoretical_loss": 3.6534091601416656,
      "objective/train/tokens_used": 1008412128,
      "theoretical_loss": 3.6534091601416656,
      "tokens_seen": 987952128
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003538816449348044,
      "loss": 2.9636,
      "theoretical_loss": 3.6533918241716608,
      "tokens_seen": 988001280
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035387161484453364,
      "loss": 2.8623,
      "theoretical_loss": 3.6533687112621367,
      "tokens_seen": 988066816
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035386158475426277,
      "loss": 3.0957,
      "theoretical_loss": 3.6533456003147964,
      "tokens_seen": 988132352
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.000353851554663992,
      "loss": 2.8578,
      "theoretical_loss": 3.653322491329344,
      "tokens_seen": 988197888
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035384152457372113,
      "loss": 2.8403,
      "theoretical_loss": 3.653299384305482,
      "tokens_seen": 988263424
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035383149448345037,
      "loss": 2.8318,
      "theoretical_loss": 3.653276279242915,
      "tokens_seen": 988328960
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035382146439317955,
      "loss": 2.8225,
      "theoretical_loss": 3.6532531761413454,
      "tokens_seen": 988394496
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035381143430290873,
      "loss": 2.7515,
      "theoretical_loss": 3.6532300750004767,
      "tokens_seen": 988460032
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003538014042126379,
      "loss": 2.8674,
      "theoretical_loss": 3.6532069758200136,
      "tokens_seen": 988525568
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003537913741223671,
      "loss": 2.8581,
      "theoretical_loss": 3.653183878599659,
      "tokens_seen": 988591104
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003537813440320963,
      "loss": 2.8162,
      "theoretical_loss": 3.6531607833391164,
      "tokens_seen": 988656640
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003537713139418255,
      "loss": 2.9214,
      "theoretical_loss": 3.6531376900380903,
      "tokens_seen": 988722176
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035376128385155464,
      "loss": 2.9007,
      "theoretical_loss": 3.653114598696284,
      "tokens_seen": 988787712
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035375125376128387,
      "loss": 3.061,
      "theoretical_loss": 3.653091509313402,
      "tokens_seen": 988853248
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.000353741223671013,
      "loss": 3.2347,
      "theoretical_loss": 3.653068421889148,
      "tokens_seen": 988918784
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035373119358074223,
      "loss": 2.7098,
      "theoretical_loss": 3.653045336423226,
      "tokens_seen": 988984320
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003537211634904714,
      "loss": 2.6685,
      "theoretical_loss": 3.65302225291534,
      "tokens_seen": 989049856
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003537111334002006,
      "loss": 3.1113,
      "theoretical_loss": 3.6529991713651944,
      "tokens_seen": 989115392
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003537011033099298,
      "loss": 2.8261,
      "theoretical_loss": 3.652976091772494,
      "tokens_seen": 989180928
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.000353691073219659,
      "loss": 2.706,
      "theoretical_loss": 3.6529530141369424,
      "tokens_seen": 989246464
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035368104312938814,
      "loss": 2.8985,
      "theoretical_loss": 3.652929938458244,
      "tokens_seen": 989312000
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003536710130391174,
      "loss": 2.8457,
      "theoretical_loss": 3.652906864736103,
      "tokens_seen": 989377536
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003536609829488465,
      "loss": 2.9615,
      "theoretical_loss": 3.6528837929702256,
      "tokens_seen": 989443072
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035365095285857574,
      "loss": 2.7433,
      "theoretical_loss": 3.6528607231603143,
      "tokens_seen": 989508608
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003536409227683049,
      "loss": 2.8964,
      "theoretical_loss": 3.6528376553060746,
      "tokens_seen": 989574144
    },
    {
      "epoch": 3.03,
      "objective/train/docs_used": 1137615,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.197035074234009,
      "objective/train/theoretical_loss": 3.6528318886480537,
      "objective/train/tokens_used": 1010050528,
      "theoretical_loss": 3.6528318886480537,
      "tokens_seen": 989590528
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003536308926780341,
      "loss": 3.1049,
      "theoretical_loss": 3.6528145894072113,
      "tokens_seen": 989639680
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003536208625877633,
      "loss": 2.9796,
      "theoretical_loss": 3.6527915254634293,
      "tokens_seen": 989705216
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035361083249749246,
      "loss": 2.7986,
      "theoretical_loss": 3.652768463474433,
      "tokens_seen": 989770752
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003536008024072217,
      "loss": 2.8468,
      "theoretical_loss": 3.652745403439928,
      "tokens_seen": 989836288
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003535907723169509,
      "loss": 2.8229,
      "theoretical_loss": 3.652722345359618,
      "tokens_seen": 989901824
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035358074222668006,
      "loss": 2.7055,
      "theoretical_loss": 3.65269928923321,
      "tokens_seen": 989967360
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035357071213640924,
      "loss": 2.593,
      "theoretical_loss": 3.652676235060407,
      "tokens_seen": 990032896
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003535606820461384,
      "loss": 2.8729,
      "theoretical_loss": 3.6526531828409157,
      "tokens_seen": 990098432
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003535506519558676,
      "loss": 2.9613,
      "theoretical_loss": 3.6526301325744406,
      "tokens_seen": 990163968
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035354062186559684,
      "loss": 3.0164,
      "theoretical_loss": 3.652607084260687,
      "tokens_seen": 990229504
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035353059177532597,
      "loss": 2.9109,
      "theoretical_loss": 3.65258403789936,
      "tokens_seen": 990295040
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003535205616850552,
      "loss": 2.8497,
      "theoretical_loss": 3.652560993490166,
      "tokens_seen": 990360576
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003535105315947844,
      "loss": 2.9282,
      "theoretical_loss": 3.6525379510328095,
      "tokens_seen": 990426112
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035350050150451356,
      "loss": 2.9462,
      "theoretical_loss": 3.652514910526997,
      "tokens_seen": 990491648
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035349047141424274,
      "loss": 2.7424,
      "theoretical_loss": 3.652491871972433,
      "tokens_seen": 990557184
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003534804413239719,
      "loss": 2.738,
      "theoretical_loss": 3.6524688353688237,
      "tokens_seen": 990622720
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003534704112337011,
      "loss": 3.0098,
      "theoretical_loss": 3.652445800715875,
      "tokens_seen": 990688256
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035346038114343034,
      "loss": 2.6713,
      "theoretical_loss": 3.6524227680132926,
      "tokens_seen": 990753792
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035345035105315947,
      "loss": 3.0378,
      "theoretical_loss": 3.652399737260782,
      "tokens_seen": 990819328
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003534403209628887,
      "loss": 3.0039,
      "theoretical_loss": 3.65237670845805,
      "tokens_seen": 990884864
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035343029087261783,
      "loss": 2.8203,
      "theoretical_loss": 3.6523536816048017,
      "tokens_seen": 990950400
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035342026078234707,
      "loss": 2.9151,
      "theoretical_loss": 3.6523306567007436,
      "tokens_seen": 991015936
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035341023069207625,
      "loss": 3.0935,
      "theoretical_loss": 3.6523076337455813,
      "tokens_seen": 991081472
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035340020060180543,
      "loss": 2.7343,
      "theoretical_loss": 3.652284612739022,
      "tokens_seen": 991147008
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003533901705115346,
      "loss": 2.9753,
      "theoretical_loss": 3.6522615936807705,
      "tokens_seen": 991212544
    },
    {
      "epoch": 3.03,
      "objective/train/docs_used": 1138406,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.1054980754852295,
      "objective/train/theoretical_loss": 3.6522558392205973,
      "objective/train/tokens_used": 1011688928,
      "theoretical_loss": 3.6522558392205973,
      "tokens_seen": 991228928
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035338014042126384,
      "loss": 2.9782,
      "theoretical_loss": 3.652238576570535,
      "tokens_seen": 991278080
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035337011033099297,
      "loss": 2.844,
      "theoretical_loss": 3.6522155614080205,
      "tokens_seen": 991343616
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003533600802407222,
      "loss": 2.9592,
      "theoretical_loss": 3.652192548192933,
      "tokens_seen": 991409152
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035335005015045133,
      "loss": 2.976,
      "theoretical_loss": 3.6521695369249807,
      "tokens_seen": 991474688
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035334002006018057,
      "loss": 2.9461,
      "theoretical_loss": 3.652146527603869,
      "tokens_seen": 991540224
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035332998996990975,
      "loss": 2.8723,
      "theoretical_loss": 3.6521235202293045,
      "tokens_seen": 991605760
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035331995987963893,
      "loss": 2.7463,
      "theoretical_loss": 3.6521005148009946,
      "tokens_seen": 991671296
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003533099297893681,
      "loss": 2.8138,
      "theoretical_loss": 3.6520775113186454,
      "tokens_seen": 991736832
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003532998996990973,
      "loss": 2.8062,
      "theoretical_loss": 3.652054509781964,
      "tokens_seen": 991802368
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003532898696088265,
      "loss": 2.9703,
      "theoretical_loss": 3.652031510190657,
      "tokens_seen": 991867904
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003532798395185557,
      "loss": 2.7611,
      "theoretical_loss": 3.6520085125444313,
      "tokens_seen": 991933440
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035326980942828484,
      "loss": 2.7765,
      "theoretical_loss": 3.651985516842995,
      "tokens_seen": 991998976
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035325977933801407,
      "loss": 2.7648,
      "theoretical_loss": 3.6519625230860537,
      "tokens_seen": 992064512
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003532497492477432,
      "loss": 3.091,
      "theoretical_loss": 3.6519395312733156,
      "tokens_seen": 992130048
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035323971915747243,
      "loss": 2.7488,
      "theoretical_loss": 3.651916541404487,
      "tokens_seen": 992195584
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003532296890672016,
      "loss": 2.7365,
      "theoretical_loss": 3.651893553479276,
      "tokens_seen": 992261120
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003532196589769308,
      "loss": 2.9087,
      "theoretical_loss": 3.6518705674973893,
      "tokens_seen": 992326656
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035320962888666,
      "loss": 2.8225,
      "theoretical_loss": 3.651847583458535,
      "tokens_seen": 992392192
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003531995987963892,
      "loss": 2.5887,
      "theoretical_loss": 3.65182460136242,
      "tokens_seen": 992457728
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035318956870611834,
      "loss": 2.8481,
      "theoretical_loss": 3.651801621208752,
      "tokens_seen": 992523264
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003531795386158476,
      "loss": 2.7536,
      "theoretical_loss": 3.651778642997238,
      "tokens_seen": 992588800
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003531695085255767,
      "loss": 2.8793,
      "theoretical_loss": 3.6517556667275866,
      "tokens_seen": 992654336
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035315947843530594,
      "loss": 2.8578,
      "theoretical_loss": 3.651732692399505,
      "tokens_seen": 992719872
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003531494483450351,
      "loss": 3.0135,
      "theoretical_loss": 3.6517097200127013,
      "tokens_seen": 992785408
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003531394182547643,
      "loss": 2.7965,
      "theoretical_loss": 3.651686749566883,
      "tokens_seen": 992850944
    },
    {
      "epoch": 3.03,
      "objective/train/docs_used": 1143666,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9665863513946533,
      "objective/train/theoretical_loss": 3.6516810072586727,
      "objective/train/tokens_used": 1013327328,
      "theoretical_loss": 3.6516810072586727,
      "tokens_seen": 992867328
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003531293881644935,
      "loss": 2.9326,
      "theoretical_loss": 3.6516637810617576,
      "tokens_seen": 992916480
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035311935807422266,
      "loss": 2.9258,
      "theoretical_loss": 3.651640814497034,
      "tokens_seen": 992982016
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035310932798395184,
      "loss": 3.1624,
      "theoretical_loss": 3.6516178498724194,
      "tokens_seen": 993047552
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003530992978936811,
      "loss": 2.8843,
      "theoretical_loss": 3.6515948871876227,
      "tokens_seen": 993113088
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003530892678034102,
      "loss": 2.7025,
      "theoretical_loss": 3.6515719264423514,
      "tokens_seen": 993178624
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035307923771313944,
      "loss": 2.7606,
      "theoretical_loss": 3.651548967636314,
      "tokens_seen": 993244160
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035306920762286857,
      "loss": 2.9293,
      "theoretical_loss": 3.651526010769219,
      "tokens_seen": 993309696
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003530591775325978,
      "loss": 2.6736,
      "theoretical_loss": 3.6515030558407737,
      "tokens_seen": 993375232
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.000353049147442327,
      "loss": 2.9679,
      "theoretical_loss": 3.651480102850688,
      "tokens_seen": 993440768
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035303911735205617,
      "loss": 2.6353,
      "theoretical_loss": 3.651457151798669,
      "tokens_seen": 993506304
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035302908726178535,
      "loss": 2.9455,
      "theoretical_loss": 3.6514342026844258,
      "tokens_seen": 993571840
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003530190571715146,
      "loss": 2.9826,
      "theoretical_loss": 3.6514112555076674,
      "tokens_seen": 993637376
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003530090270812437,
      "loss": 2.8492,
      "theoretical_loss": 3.651388310268102,
      "tokens_seen": 993702912
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035299899699097294,
      "loss": 2.8351,
      "theoretical_loss": 3.6513653669654387,
      "tokens_seen": 993768448
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035298896690070207,
      "loss": 2.9885,
      "theoretical_loss": 3.6513424255993856,
      "tokens_seen": 993833984
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003529789368104313,
      "loss": 2.7295,
      "theoretical_loss": 3.651319486169652,
      "tokens_seen": 993899520
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003529689067201605,
      "loss": 2.9383,
      "theoretical_loss": 3.6512965486759468,
      "tokens_seen": 993965056
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035295887662988967,
      "loss": 2.6859,
      "theoretical_loss": 3.651273613117979,
      "tokens_seen": 994030592
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035294884653961885,
      "loss": 2.8415,
      "theoretical_loss": 3.6512506794954573,
      "tokens_seen": 994096128
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035293881644934803,
      "loss": 2.697,
      "theoretical_loss": 3.651227747808091,
      "tokens_seen": 994161664
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003529287863590772,
      "loss": 2.8838,
      "theoretical_loss": 3.6512048180555894,
      "tokens_seen": 994227200
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035291875626880645,
      "loss": 2.9552,
      "theoretical_loss": 3.6511818902376616,
      "tokens_seen": 994292736
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003529087261785356,
      "loss": 3.0095,
      "theoretical_loss": 3.6511589643540168,
      "tokens_seen": 994358272
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003528986960882648,
      "loss": 2.6717,
      "theoretical_loss": 3.6511360404043645,
      "tokens_seen": 994423808
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035288866599799394,
      "loss": 3.1065,
      "theoretical_loss": 3.6511131183884142,
      "tokens_seen": 994489344
    },
    {
      "epoch": 3.03,
      "objective/train/docs_used": 1148851,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7076807022094727,
      "objective/train/theoretical_loss": 3.6511073881865337,
      "objective/train/tokens_used": 1014965728,
      "theoretical_loss": 3.6511073881865337,
      "tokens_seen": 994505728
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035287863590772317,
      "loss": 2.8518,
      "theoretical_loss": 3.6510901983058757,
      "tokens_seen": 994554880
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035286860581745235,
      "loss": 3.0575,
      "theoretical_loss": 3.6510672801564574,
      "tokens_seen": 994620416
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035285857572718153,
      "loss": 2.7546,
      "theoretical_loss": 3.65104436393987,
      "tokens_seen": 994685952
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035284854563691077,
      "loss": 2.8476,
      "theoretical_loss": 3.651021449655822,
      "tokens_seen": 994751488
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035283851554663995,
      "loss": 2.7623,
      "theoretical_loss": 3.650998537304025,
      "tokens_seen": 994817024
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035282848545636913,
      "loss": 2.9836,
      "theoretical_loss": 3.6509756268841866,
      "tokens_seen": 994882560
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003528184553660983,
      "loss": 2.7934,
      "theoretical_loss": 3.6509527183960184,
      "tokens_seen": 994948096
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003528084252758275,
      "loss": 2.8664,
      "theoretical_loss": 3.6509298118392293,
      "tokens_seen": 995013632
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003527983951855567,
      "loss": 2.8624,
      "theoretical_loss": 3.650906907213529,
      "tokens_seen": 995079168
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003527883650952859,
      "loss": 2.958,
      "theoretical_loss": 3.650884004518629,
      "tokens_seen": 995144704
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035277833500501504,
      "loss": 2.9275,
      "theoretical_loss": 3.650861103754238,
      "tokens_seen": 995210240
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035276830491474427,
      "loss": 3.1767,
      "theoretical_loss": 3.650838204920067,
      "tokens_seen": 995275776
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003527582748244734,
      "loss": 2.8752,
      "theoretical_loss": 3.650815308015826,
      "tokens_seen": 995341312
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035274824473420263,
      "loss": 2.9867,
      "theoretical_loss": 3.650792413041225,
      "tokens_seen": 995406848
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003527382146439318,
      "loss": 2.9126,
      "theoretical_loss": 3.6507695199959747,
      "tokens_seen": 995472384
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.000352728184553661,
      "loss": 2.957,
      "theoretical_loss": 3.650746628879785,
      "tokens_seen": 995537920
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003527181544633902,
      "loss": 2.9834,
      "theoretical_loss": 3.650723739692367,
      "tokens_seen": 995603456
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003527081243731194,
      "loss": 2.6613,
      "theoretical_loss": 3.650700852433431,
      "tokens_seen": 995668992
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035269809428284854,
      "loss": 2.4629,
      "theoretical_loss": 3.6506779671026877,
      "tokens_seen": 995734528
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003526880641925778,
      "loss": 2.7763,
      "theoretical_loss": 3.6506550836998475,
      "tokens_seen": 995800064
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003526780341023069,
      "loss": 2.7792,
      "theoretical_loss": 3.650632202224621,
      "tokens_seen": 995865600
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035266800401203614,
      "loss": 3.1098,
      "theoretical_loss": 3.6506093226767193,
      "tokens_seen": 995931136
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003526579739217653,
      "loss": 2.7007,
      "theoretical_loss": 3.650586445055853,
      "tokens_seen": 995996672
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003526479438314945,
      "loss": 2.9314,
      "theoretical_loss": 3.650563569361734,
      "tokens_seen": 996062208
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003526379137412237,
      "loss": 3.065,
      "theoretical_loss": 3.6505406955940725,
      "tokens_seen": 996127744
    },
    {
      "epoch": 3.03,
      "objective/train/docs_used": 1153874,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.706052541732788,
      "objective/train/theoretical_loss": 3.650534977453132,
      "objective/train/tokens_used": 1016604128,
      "theoretical_loss": 3.650534977453132,
      "tokens_seen": 996144128
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035262788365095286,
      "loss": 2.8504,
      "theoretical_loss": 3.650517823752579,
      "tokens_seen": 996193280
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035261785356068204,
      "loss": 2.9067,
      "theoretical_loss": 3.650494953836965,
      "tokens_seen": 996258816
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003526078234704113,
      "loss": 2.7841,
      "theoretical_loss": 3.650472085846942,
      "tokens_seen": 996324352
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003525977933801404,
      "loss": 2.7586,
      "theoretical_loss": 3.6504492197822214,
      "tokens_seen": 996389888
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035258776328986964,
      "loss": 2.6799,
      "theoretical_loss": 3.650426355642514,
      "tokens_seen": 996455424
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035257773319959877,
      "loss": 2.8452,
      "theoretical_loss": 3.6504034934275307,
      "tokens_seen": 996520960
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.000352567703109328,
      "loss": 2.7273,
      "theoretical_loss": 3.650380633136984,
      "tokens_seen": 996586496
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003525576730190572,
      "loss": 2.9257,
      "theoretical_loss": 3.6503577747705847,
      "tokens_seen": 996652032
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035254764292878637,
      "loss": 2.8902,
      "theoretical_loss": 3.6503349183280447,
      "tokens_seen": 996717568
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035253761283851555,
      "loss": 2.8466,
      "theoretical_loss": 3.650312063809075,
      "tokens_seen": 996783104
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003525275827482448,
      "loss": 2.9663,
      "theoretical_loss": 3.6502892112133876,
      "tokens_seen": 996848640
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003525175526579739,
      "loss": 2.9607,
      "theoretical_loss": 3.6502663605406944,
      "tokens_seen": 996914176
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035250752256770314,
      "loss": 2.795,
      "theoretical_loss": 3.6502435117907073,
      "tokens_seen": 996979712
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035249749247743227,
      "loss": 2.8096,
      "theoretical_loss": 3.6502206649631375,
      "tokens_seen": 997045248
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003524874623871615,
      "loss": 2.9291,
      "theoretical_loss": 3.6501978200576977,
      "tokens_seen": 997110784
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003524774322968907,
      "loss": 2.7279,
      "theoretical_loss": 3.6501749770740988,
      "tokens_seen": 997176320
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035246740220661987,
      "loss": 2.9556,
      "theoretical_loss": 3.6501521360120543,
      "tokens_seen": 997241856
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035245737211634905,
      "loss": 2.9142,
      "theoretical_loss": 3.650129296871275,
      "tokens_seen": 997307392
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035244734202607823,
      "loss": 2.9961,
      "theoretical_loss": 3.6501064596514734,
      "tokens_seen": 997372928
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003524373119358074,
      "loss": 2.7048,
      "theoretical_loss": 3.6500836243523627,
      "tokens_seen": 997438464
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035242728184553665,
      "loss": 2.8819,
      "theoretical_loss": 3.6500607909736535,
      "tokens_seen": 997504000
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003524172517552658,
      "loss": 2.9255,
      "theoretical_loss": 3.6500379595150596,
      "tokens_seen": 997569536
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.000352407221664995,
      "loss": 2.8446,
      "theoretical_loss": 3.6500151299762926,
      "tokens_seen": 997635072
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035239719157472414,
      "loss": 2.6841,
      "theoretical_loss": 3.649992302357065,
      "tokens_seen": 997700608
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035238716148445337,
      "loss": 2.8274,
      "theoretical_loss": 3.64996947665709,
      "tokens_seen": 997766144
    },
    {
      "epoch": 3.03,
      "objective/train/docs_used": 1156592,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5675315856933594,
      "objective/train/theoretical_loss": 3.6499637705319454,
      "objective/train/tokens_used": 1018242528,
      "theoretical_loss": 3.6499637705319454,
      "tokens_seen": 997782528
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035237713139418255,
      "loss": 2.9074,
      "theoretical_loss": 3.6499466528760793,
      "tokens_seen": 997831680
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035236710130391173,
      "loss": 2.8998,
      "theoretical_loss": 3.649923831013746,
      "tokens_seen": 997897216
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003523570712136409,
      "loss": 2.7522,
      "theoretical_loss": 3.6499010110698027,
      "tokens_seen": 997962752
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035234704112337015,
      "loss": 2.8103,
      "theoretical_loss": 3.6498781930439623,
      "tokens_seen": 998028288
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003523370110330993,
      "loss": 2.8486,
      "theoretical_loss": 3.6498553769359376,
      "tokens_seen": 998093824
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003523269809428285,
      "loss": 2.8552,
      "theoretical_loss": 3.6498325627454413,
      "tokens_seen": 998159360
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035231695085255764,
      "loss": 2.8996,
      "theoretical_loss": 3.649809750472187,
      "tokens_seen": 998224896
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003523069207622869,
      "loss": 2.8666,
      "theoretical_loss": 3.649786940115887,
      "tokens_seen": 998290432
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035229689067201606,
      "loss": 2.8501,
      "theoretical_loss": 3.649764131676255,
      "tokens_seen": 998355968
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035228686058174524,
      "loss": 2.7246,
      "theoretical_loss": 3.649741325153003,
      "tokens_seen": 998421504
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003522768304914744,
      "loss": 2.7134,
      "theoretical_loss": 3.6497185205458456,
      "tokens_seen": 998487040
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003522668004012036,
      "loss": 2.9018,
      "theoretical_loss": 3.6496957178544953,
      "tokens_seen": 998552576
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003522567703109328,
      "loss": 2.7131,
      "theoretical_loss": 3.649672917078666,
      "tokens_seen": 998618112
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.000352246740220662,
      "loss": 2.9995,
      "theoretical_loss": 3.6496501182180703,
      "tokens_seen": 998683648
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035223671013039114,
      "loss": 2.8589,
      "theoretical_loss": 3.649627321272422,
      "tokens_seen": 998749184
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003522266800401204,
      "loss": 2.7328,
      "theoretical_loss": 3.649604526241435,
      "tokens_seen": 998814720
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035221664994984956,
      "loss": 2.9369,
      "theoretical_loss": 3.6495817331248226,
      "tokens_seen": 998880256
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035220661985957874,
      "loss": 3.0445,
      "theoretical_loss": 3.6495589419222982,
      "tokens_seen": 998945792
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003521965897693079,
      "loss": 2.8777,
      "theoretical_loss": 3.6495361526335754,
      "tokens_seen": 999011328
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003521865596790371,
      "loss": 2.8657,
      "theoretical_loss": 3.649513365258369,
      "tokens_seen": 999076864
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003521765295887663,
      "loss": 2.7128,
      "theoretical_loss": 3.6494905797963915,
      "tokens_seen": 999142400
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003521664994984955,
      "loss": 2.704,
      "theoretical_loss": 3.649467796247358,
      "tokens_seen": 999207936
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035215646940822465,
      "loss": 2.8895,
      "theoretical_loss": 3.649445014610981,
      "tokens_seen": 999273472
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003521464393179539,
      "loss": 2.8723,
      "theoretical_loss": 3.649422234886976,
      "tokens_seen": 999339008
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.000352136409227683,
      "loss": 3.0843,
      "theoretical_loss": 3.6493994570750568,
      "tokens_seen": 999404544
    },
    {
      "epoch": 3.03,
      "objective/train/docs_used": 1159661,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.2150416374206543,
      "objective/train/theoretical_loss": 3.6493937629208064,
      "objective/train/tokens_used": 1019880928,
      "theoretical_loss": 3.6493937629208064,
      "tokens_seen": 999420928
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035212637913741224,
      "loss": 2.8963,
      "theoretical_loss": 3.649376681174936,
      "tokens_seen": 999470080
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003521163490471414,
      "loss": 3.0057,
      "theoretical_loss": 3.64935390718633,
      "tokens_seen": 999535616
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003521063189568706,
      "loss": 2.9464,
      "theoretical_loss": 3.6493311351089517,
      "tokens_seen": 999601152
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035209628886659984,
      "loss": 3.0535,
      "theoretical_loss": 3.6493083649425153,
      "tokens_seen": 999666688
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035208625877632897,
      "loss": 3.0811,
      "theoretical_loss": 3.649285596686736,
      "tokens_seen": 999732224
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003520762286860582,
      "loss": 2.69,
      "theoretical_loss": 3.6492628303413275,
      "tokens_seen": 999797760
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003520661985957874,
      "loss": 3.0515,
      "theoretical_loss": 3.649240065906005,
      "tokens_seen": 999863296
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035205616850551657,
      "loss": 2.9586,
      "theoretical_loss": 3.6492173033804827,
      "tokens_seen": 999928832
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035204613841524575,
      "loss": 3.0117,
      "theoretical_loss": 3.649194542764475,
      "tokens_seen": 999994368
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.000352036108324975,
      "loss": 2.9995,
      "theoretical_loss": 3.6491717840576974,
      "tokens_seen": 1000059904
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003520260782347041,
      "loss": 3.0554,
      "theoretical_loss": 3.6491490272598637,
      "tokens_seen": 1000125440
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035201604814443334,
      "loss": 3.0433,
      "theoretical_loss": 3.649126272370689,
      "tokens_seen": 1000190976
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035200601805416247,
      "loss": 2.8908,
      "theoretical_loss": 3.6491035193898886,
      "tokens_seen": 1000256512
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003519959879638917,
      "loss": 2.9182,
      "theoretical_loss": 3.6490807683171766,
      "tokens_seen": 1000322048
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003519859578736209,
      "loss": 2.7976,
      "theoretical_loss": 3.649058019152269,
      "tokens_seen": 1000387584
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035197592778335007,
      "loss": 2.671,
      "theoretical_loss": 3.6490352718948795,
      "tokens_seen": 1000453120
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035196589769307925,
      "loss": 2.8257,
      "theoretical_loss": 3.649012526544725,
      "tokens_seen": 1000518656
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035195586760280843,
      "loss": 2.5004,
      "theoretical_loss": 3.6489897831015194,
      "tokens_seen": 1000584192
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003519458375125376,
      "loss": 2.5683,
      "theoretical_loss": 3.648967041564978,
      "tokens_seen": 1000649728
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035193580742226685,
      "loss": 3.0786,
      "theoretical_loss": 3.648944301934817,
      "tokens_seen": 1000715264
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.000351925777331996,
      "loss": 2.8131,
      "theoretical_loss": 3.6489215642107506,
      "tokens_seen": 1000780800
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003519157472417252,
      "loss": 2.787,
      "theoretical_loss": 3.6488988283924946,
      "tokens_seen": 1000846336
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035190571715145434,
      "loss": 2.8895,
      "theoretical_loss": 3.648876094479765,
      "tokens_seen": 1000911872
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035189568706118357,
      "loss": 2.5759,
      "theoretical_loss": 3.648853362472277,
      "tokens_seen": 1000977408
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035188565697091275,
      "loss": 2.6727,
      "theoretical_loss": 3.648830632369746,
      "tokens_seen": 1001042944
    },
    {
      "epoch": 3.03,
      "objective/train/docs_used": 1160836,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.1073474884033203,
      "objective/train/theoretical_loss": 3.648824950141729,
      "objective/train/tokens_used": 1021519328,
      "theoretical_loss": 3.648824950141729,
      "tokens_seen": 1001059328
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035187562688064193,
      "loss": 3.1555,
      "theoretical_loss": 3.6488079041718877,
      "tokens_seen": 1001108480
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003518655967903711,
      "loss": 2.9301,
      "theoretical_loss": 3.648785177878418,
      "tokens_seen": 1001174016
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035185556670010035,
      "loss": 2.7881,
      "theoretical_loss": 3.648762453489053,
      "tokens_seen": 1001239552
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003518455366098295,
      "loss": 2.9136,
      "theoretical_loss": 3.6487397310035083,
      "tokens_seen": 1001305088
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003518355065195587,
      "loss": 2.9075,
      "theoretical_loss": 3.648717010421499,
      "tokens_seen": 1001370624
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035182547642928784,
      "loss": 2.561,
      "theoretical_loss": 3.6486942917427427,
      "tokens_seen": 1001436160
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003518154463390171,
      "loss": 3.1295,
      "theoretical_loss": 3.648671574966954,
      "tokens_seen": 1001501696
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035180541624874626,
      "loss": 2.7492,
      "theoretical_loss": 3.6486488600938496,
      "tokens_seen": 1001567232
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035179538615847544,
      "loss": 2.8224,
      "theoretical_loss": 3.648626147123146,
      "tokens_seen": 1001632768
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003517853560682046,
      "loss": 2.9366,
      "theoretical_loss": 3.648603436054558,
      "tokens_seen": 1001698304
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003517753259779338,
      "loss": 2.9208,
      "theoretical_loss": 3.648580726887803,
      "tokens_seen": 1001763840
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.000351765295887663,
      "loss": 2.8788,
      "theoretical_loss": 3.648558019622598,
      "tokens_seen": 1001829376
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003517552657973922,
      "loss": 2.9987,
      "theoretical_loss": 3.648535314258658,
      "tokens_seen": 1001894912
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035174523570712134,
      "loss": 3.0191,
      "theoretical_loss": 3.6485126107957004,
      "tokens_seen": 1001960448
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003517352056168506,
      "loss": 2.8817,
      "theoretical_loss": 3.648489909233441,
      "tokens_seen": 1002025984
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035172517552657976,
      "loss": 2.9599,
      "theoretical_loss": 3.648467209571597,
      "tokens_seen": 1002091520
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035171514543630894,
      "loss": 2.7182,
      "theoretical_loss": 3.6484445118098847,
      "tokens_seen": 1002157056
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003517051153460381,
      "loss": 2.9031,
      "theoretical_loss": 3.648421815948021,
      "tokens_seen": 1002222592
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003516950852557673,
      "loss": 2.6092,
      "theoretical_loss": 3.6483991219857224,
      "tokens_seen": 1002288128
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003516850551654965,
      "loss": 2.6493,
      "theoretical_loss": 3.648376429922706,
      "tokens_seen": 1002353664
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003516750250752257,
      "loss": 2.7576,
      "theoretical_loss": 3.6483537397586887,
      "tokens_seen": 1002419200
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035166499498495485,
      "loss": 2.9421,
      "theoretical_loss": 3.648331051493387,
      "tokens_seen": 1002484736
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003516549648946841,
      "loss": 2.6704,
      "theoretical_loss": 3.6483083651265185,
      "tokens_seen": 1002550272
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003516449348044132,
      "loss": 2.8942,
      "theoretical_loss": 3.6482856806578,
      "tokens_seen": 1002615808
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035163490471414244,
      "loss": 2.9315,
      "theoretical_loss": 3.648262998086948,
      "tokens_seen": 1002681344
    },
    {
      "epoch": 3.03,
      "objective/train/docs_used": 1161424,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.439594030380249,
      "objective/train/theoretical_loss": 3.648257327740744,
      "objective/train/tokens_used": 1023157728,
      "theoretical_loss": 3.648257327740744,
      "tokens_seen": 1002697728
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003516248746238716,
      "loss": 2.7253,
      "theoretical_loss": 3.6482403174136806,
      "tokens_seen": 1002746880
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003516148445336008,
      "loss": 2.9394,
      "theoretical_loss": 3.648217638637715,
      "tokens_seen": 1002812416
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035160481444333,
      "loss": 2.498,
      "theoretical_loss": 3.648194961758768,
      "tokens_seen": 1002877952
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035159478435305917,
      "loss": 2.7678,
      "theoretical_loss": 3.6481722867765574,
      "tokens_seen": 1002943488
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035158475426278835,
      "loss": 2.7227,
      "theoretical_loss": 3.6481496136908,
      "tokens_seen": 1003009024
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003515747241725176,
      "loss": 2.6544,
      "theoretical_loss": 3.6481269425012144,
      "tokens_seen": 1003074560
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003515646940822467,
      "loss": 2.7992,
      "theoretical_loss": 3.648104273207517,
      "tokens_seen": 1003140096
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035155466399197595,
      "loss": 2.8121,
      "theoretical_loss": 3.648081605809426,
      "tokens_seen": 1003205632
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035154463390170513,
      "loss": 2.8442,
      "theoretical_loss": 3.648058940306659,
      "tokens_seen": 1003271168
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003515346038114343,
      "loss": 2.91,
      "theoretical_loss": 3.648036276698934,
      "tokens_seen": 1003336704
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003515245737211635,
      "loss": 2.9568,
      "theoretical_loss": 3.6480136149859677,
      "tokens_seen": 1003402240
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035151454363089267,
      "loss": 2.7311,
      "theoretical_loss": 3.6479909551674794,
      "tokens_seen": 1003467776
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035150451354062185,
      "loss": 2.8667,
      "theoretical_loss": 3.647968297243186,
      "tokens_seen": 1003533312
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003514944834503511,
      "loss": 2.7325,
      "theoretical_loss": 3.6479456412128064,
      "tokens_seen": 1003598848
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003514844533600802,
      "loss": 2.9816,
      "theoretical_loss": 3.6479229870760577,
      "tokens_seen": 1003664384
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035147442326980945,
      "loss": 2.8746,
      "theoretical_loss": 3.6479003348326584,
      "tokens_seen": 1003729920
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003514643931795386,
      "loss": 2.9706,
      "theoretical_loss": 3.647877684482326,
      "tokens_seen": 1003795456
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003514543630892678,
      "loss": 2.769,
      "theoretical_loss": 3.64785503602478,
      "tokens_seen": 1003860992
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.000351444332998997,
      "loss": 2.815,
      "theoretical_loss": 3.647832389459738,
      "tokens_seen": 1003926528
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003514343029087262,
      "loss": 2.6748,
      "theoretical_loss": 3.647809744786918,
      "tokens_seen": 1003992064
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035142427281845536,
      "loss": 2.787,
      "theoretical_loss": 3.6477871020060393,
      "tokens_seen": 1004057600
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035141424272818454,
      "loss": 2.9168,
      "theoretical_loss": 3.647764461116819,
      "tokens_seen": 1004123136
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003514042126379137,
      "loss": 2.7364,
      "theoretical_loss": 3.6477418221189772,
      "tokens_seen": 1004188672
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035139418254764295,
      "loss": 2.7187,
      "theoretical_loss": 3.647719185012231,
      "tokens_seen": 1004254208
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003513841524573721,
      "loss": 2.7207,
      "theoretical_loss": 3.6476965497963,
      "tokens_seen": 1004319744
    },
    {
      "epoch": 3.03,
      "objective/train/docs_used": 1162777,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7652225494384766,
      "objective/train/theoretical_loss": 3.6476908912877244,
      "objective/train/tokens_used": 1024796128,
      "theoretical_loss": 3.6476908912877244,
      "tokens_seen": 1004336128
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003513741223671013,
      "loss": 2.9625,
      "theoretical_loss": 3.647673916470903,
      "tokens_seen": 1004385280
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003513640922768305,
      "loss": 2.8803,
      "theoretical_loss": 3.647651285035758,
      "tokens_seen": 1004450816
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003513540621865597,
      "loss": 2.7448,
      "theoretical_loss": 3.647628655490584,
      "tokens_seen": 1004516352
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003513440320962889,
      "loss": 3.026,
      "theoretical_loss": 3.6476060278351,
      "tokens_seen": 1004581888
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035133400200601804,
      "loss": 2.9949,
      "theoretical_loss": 3.6475834020690256,
      "tokens_seen": 1004647424
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003513239719157473,
      "loss": 3.0162,
      "theoretical_loss": 3.647560778192079,
      "tokens_seen": 1004712960
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035131394182547646,
      "loss": 2.7788,
      "theoretical_loss": 3.6475381562039795,
      "tokens_seen": 1004778496
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035130391173520564,
      "loss": 2.9032,
      "theoretical_loss": 3.647515536104446,
      "tokens_seen": 1004844032
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003512938816449348,
      "loss": 2.8642,
      "theoretical_loss": 3.6474929178931976,
      "tokens_seen": 1004909568
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.000351283851554664,
      "loss": 2.7253,
      "theoretical_loss": 3.647470301569954,
      "tokens_seen": 1004975104
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003512738214643932,
      "loss": 2.9281,
      "theoretical_loss": 3.647447687134435,
      "tokens_seen": 1005040640
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003512637913741224,
      "loss": 2.8002,
      "theoretical_loss": 3.6474250745863586,
      "tokens_seen": 1005106176
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035125376128385154,
      "loss": 2.7841,
      "theoretical_loss": 3.647402463925445,
      "tokens_seen": 1005171712
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003512437311935808,
      "loss": 2.7919,
      "theoretical_loss": 3.6473798551514136,
      "tokens_seen": 1005237248
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035123370110330996,
      "loss": 2.9419,
      "theoretical_loss": 3.647357248263984,
      "tokens_seen": 1005302784
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035122367101303914,
      "loss": 2.7465,
      "theoretical_loss": 3.6473346432628753,
      "tokens_seen": 1005368320
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003512136409227683,
      "loss": 2.811,
      "theoretical_loss": 3.647312040147808,
      "tokens_seen": 1005433856
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003512036108324975,
      "loss": 2.8367,
      "theoretical_loss": 3.6472894389185013,
      "tokens_seen": 1005499392
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003511935807422267,
      "loss": 2.87,
      "theoretical_loss": 3.647266839574675,
      "tokens_seen": 1005564928
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003511835506519559,
      "loss": 2.8118,
      "theoretical_loss": 3.6472442421160487,
      "tokens_seen": 1005630464
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.00035117352056168505,
      "loss": 2.947,
      "theoretical_loss": 3.6472216465423424,
      "tokens_seen": 1005696000
    },
    {
      "epoch": 3.03,
      "learning_rate": 0.0003511634904714143,
      "loss": 2.6745,
      "theoretical_loss": 3.647199052853277,
      "tokens_seen": 1005761536
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003511534603811434,
      "loss": 2.8839,
      "theoretical_loss": 3.6471764610485713,
      "tokens_seen": 1005827072
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00035114343029087264,
      "loss": 2.8895,
      "theoretical_loss": 3.647153871127946,
      "tokens_seen": 1005892608
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003511334002006018,
      "loss": 2.7014,
      "theoretical_loss": 3.647131283091121,
      "tokens_seen": 1005958144
    },
    {
      "epoch": 3.04,
      "objective/train/docs_used": 1163373,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.477107524871826,
      "objective/train/theoretical_loss": 3.6471256363762254,
      "objective/train/tokens_used": 1026434528,
      "theoretical_loss": 3.6471256363762254,
      "tokens_seen": 1005974528
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.000351123370110331,
      "loss": 2.8156,
      "theoretical_loss": 3.6471086969378166,
      "tokens_seen": 1006023680
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003511133400200602,
      "loss": 2.5824,
      "theoretical_loss": 3.647086112667753,
      "tokens_seen": 1006089216
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00035110330992978937,
      "loss": 2.8562,
      "theoretical_loss": 3.6470635302806507,
      "tokens_seen": 1006154752
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00035109327983951855,
      "loss": 2.9694,
      "theoretical_loss": 3.6470409497762297,
      "tokens_seen": 1006220288
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003510832497492478,
      "loss": 2.7441,
      "theoretical_loss": 3.647018371154211,
      "tokens_seen": 1006285824
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003510732196589769,
      "loss": 2.8582,
      "theoretical_loss": 3.6469957944143143,
      "tokens_seen": 1006351360
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00035106318956870615,
      "loss": 2.764,
      "theoretical_loss": 3.646973219556261,
      "tokens_seen": 1006416896
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00035105315947843533,
      "loss": 2.6505,
      "theoretical_loss": 3.6469506465797714,
      "tokens_seen": 1006482432
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003510431293881645,
      "loss": 2.6349,
      "theoretical_loss": 3.6469280754845657,
      "tokens_seen": 1006547968
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003510330992978937,
      "loss": 2.8268,
      "theoretical_loss": 3.6469055062703655,
      "tokens_seen": 1006613504
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00035102306920762287,
      "loss": 2.8678,
      "theoretical_loss": 3.6468829389368906,
      "tokens_seen": 1006679040
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00035101303911735205,
      "loss": 3.0746,
      "theoretical_loss": 3.6468603734838627,
      "tokens_seen": 1006744576
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003510030090270813,
      "loss": 2.9625,
      "theoretical_loss": 3.6468378099110024,
      "tokens_seen": 1006810112
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003509929789368104,
      "loss": 2.8395,
      "theoretical_loss": 3.6468152482180307,
      "tokens_seen": 1006875648
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00035098294884653965,
      "loss": 2.5887,
      "theoretical_loss": 3.646792688404669,
      "tokens_seen": 1006941184
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003509729187562688,
      "loss": 2.7411,
      "theoretical_loss": 3.646770130470637,
      "tokens_seen": 1007006720
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.000350962888665998,
      "loss": 3.0379,
      "theoretical_loss": 3.6467475744156577,
      "tokens_seen": 1007072256
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003509528585757272,
      "loss": 2.7484,
      "theoretical_loss": 3.646725020239451,
      "tokens_seen": 1007137792
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003509428284854564,
      "loss": 2.8597,
      "theoretical_loss": 3.646702467941739,
      "tokens_seen": 1007203328
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00035093279839518556,
      "loss": 2.8188,
      "theoretical_loss": 3.646679917522243,
      "tokens_seen": 1007268864
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00035092276830491474,
      "loss": 2.9834,
      "theoretical_loss": 3.646657368980683,
      "tokens_seen": 1007334400
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003509127382146439,
      "loss": 2.7302,
      "theoretical_loss": 3.6466348223167824,
      "tokens_seen": 1007399936
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00035090270812437315,
      "loss": 2.7753,
      "theoretical_loss": 3.646612277530261,
      "tokens_seen": 1007465472
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003508926780341023,
      "loss": 2.6536,
      "theoretical_loss": 3.646589734620842,
      "tokens_seen": 1007531008
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003508826479438315,
      "loss": 2.6738,
      "theoretical_loss": 3.646567193588246,
      "tokens_seen": 1007596544
    },
    {
      "epoch": 3.04,
      "objective/train/docs_used": 1164906,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.129481554031372,
      "objective/train/theoretical_loss": 3.6465615586233175,
      "objective/train/tokens_used": 1028072928,
      "theoretical_loss": 3.6465615586233175,
      "tokens_seen": 1007612928
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003508726178535607,
      "loss": 2.9174,
      "theoretical_loss": 3.6465446544321947,
      "tokens_seen": 1007662080
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003508625877632899,
      "loss": 2.7308,
      "theoretical_loss": 3.6465221171524105,
      "tokens_seen": 1007727616
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00035085255767301906,
      "loss": 2.8025,
      "theoretical_loss": 3.646499581748614,
      "tokens_seen": 1007793152
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00035084252758274824,
      "loss": 2.8899,
      "theoretical_loss": 3.646477048220528,
      "tokens_seen": 1007858688
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003508324974924774,
      "loss": 2.8594,
      "theoretical_loss": 3.646454516567875,
      "tokens_seen": 1007924224
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00035082246740220666,
      "loss": 2.9397,
      "theoretical_loss": 3.6464319867903754,
      "tokens_seen": 1007989760
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003508124373119358,
      "loss": 2.8024,
      "theoretical_loss": 3.646409458887752,
      "tokens_seen": 1008055296
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.000350802407221665,
      "loss": 2.8186,
      "theoretical_loss": 3.646386932859728,
      "tokens_seen": 1008120832
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00035079237713139415,
      "loss": 2.7744,
      "theoretical_loss": 3.646364408706024,
      "tokens_seen": 1008186368
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003507823470411234,
      "loss": 2.6754,
      "theoretical_loss": 3.646341886426362,
      "tokens_seen": 1008251904
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00035077231695085256,
      "loss": 2.9253,
      "theoretical_loss": 3.646319366020466,
      "tokens_seen": 1008317440
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00035076228686058174,
      "loss": 2.7929,
      "theoretical_loss": 3.646296847488057,
      "tokens_seen": 1008382976
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003507522567703109,
      "loss": 2.9397,
      "theoretical_loss": 3.6462743308288577,
      "tokens_seen": 1008448512
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00035074222668004016,
      "loss": 2.8635,
      "theoretical_loss": 3.646251816042591,
      "tokens_seen": 1008514048
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003507321965897693,
      "loss": 2.8216,
      "theoretical_loss": 3.6462293031289787,
      "tokens_seen": 1008579584
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003507221664994985,
      "loss": 2.8308,
      "theoretical_loss": 3.6462067920877437,
      "tokens_seen": 1008645120
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00035071213640922765,
      "loss": 2.9977,
      "theoretical_loss": 3.646184282918609,
      "tokens_seen": 1008710656
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003507021063189569,
      "loss": 2.8788,
      "theoretical_loss": 3.6461617756212963,
      "tokens_seen": 1008776192
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00035069207622868607,
      "loss": 2.6056,
      "theoretical_loss": 3.6461392701955297,
      "tokens_seen": 1008841728
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00035068204613841525,
      "loss": 2.8894,
      "theoretical_loss": 3.646116766641031,
      "tokens_seen": 1008907264
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00035067201604814443,
      "loss": 2.8793,
      "theoretical_loss": 3.646094264957523,
      "tokens_seen": 1008972800
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003506619859578736,
      "loss": 2.6769,
      "theoretical_loss": 3.64607176514473,
      "tokens_seen": 1009038336
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003506519558676028,
      "loss": 2.6217,
      "theoretical_loss": 3.6460492672023728,
      "tokens_seen": 1009103872
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.000350641925777332,
      "loss": 2.7837,
      "theoretical_loss": 3.6460267711301766,
      "tokens_seen": 1009169408
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00035063189568706115,
      "loss": 2.8821,
      "theoretical_loss": 3.6460042769278633,
      "tokens_seen": 1009234944
    },
    {
      "epoch": 3.04,
      "objective/train/docs_used": 1165274,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.818286657333374,
      "objective/train/theoretical_loss": 3.6459986536694218,
      "objective/train/tokens_used": 1029711328,
      "theoretical_loss": 3.6459986536694218,
      "tokens_seen": 1009251328
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003506218655967904,
      "loss": 2.8747,
      "theoretical_loss": 3.6459817845951563,
      "tokens_seen": 1009300480
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003506118355065195,
      "loss": 2.6386,
      "theoretical_loss": 3.645959294131779,
      "tokens_seen": 1009366016
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00035060180541624875,
      "loss": 2.8541,
      "theoretical_loss": 3.6459368055374544,
      "tokens_seen": 1009431552
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.000350591775325978,
      "loss": 2.925,
      "theoretical_loss": 3.6459143188119056,
      "tokens_seen": 1009497088
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003505817452357071,
      "loss": 2.7224,
      "theoretical_loss": 3.645891833954857,
      "tokens_seen": 1009562624
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00035057171514543635,
      "loss": 2.9708,
      "theoretical_loss": 3.645869350966031,
      "tokens_seen": 1009628160
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00035056168505516553,
      "loss": 2.6602,
      "theoretical_loss": 3.645846869845152,
      "tokens_seen": 1009693696
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003505516549648947,
      "loss": 2.8586,
      "theoretical_loss": 3.645824390591943,
      "tokens_seen": 1009759232
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003505416248746239,
      "loss": 2.6716,
      "theoretical_loss": 3.6458019132061272,
      "tokens_seen": 1009824768
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00035053159478435307,
      "loss": 3.0043,
      "theoretical_loss": 3.6457794376874295,
      "tokens_seen": 1009890304
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00035052156469408225,
      "loss": 2.7406,
      "theoretical_loss": 3.645756964035573,
      "tokens_seen": 1009955840
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003505115346038115,
      "loss": 2.8073,
      "theoretical_loss": 3.6457344922502815,
      "tokens_seen": 1010021376
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003505015045135406,
      "loss": 2.9574,
      "theoretical_loss": 3.6457120223312787,
      "tokens_seen": 1010086912
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00035049147442326985,
      "loss": 2.8053,
      "theoretical_loss": 3.645689554278289,
      "tokens_seen": 1010152448
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.000350481444332999,
      "loss": 2.7435,
      "theoretical_loss": 3.6456670880910362,
      "tokens_seen": 1010217984
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003504714142427282,
      "loss": 2.9574,
      "theoretical_loss": 3.645644623769244,
      "tokens_seen": 1010283520
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003504613841524574,
      "loss": 2.8229,
      "theoretical_loss": 3.645622161312637,
      "tokens_seen": 1010349056
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003504513540621866,
      "loss": 2.9436,
      "theoretical_loss": 3.645599700720939,
      "tokens_seen": 1010414592
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00035044132397191576,
      "loss": 2.8731,
      "theoretical_loss": 3.6455772419938746,
      "tokens_seen": 1010480128
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00035043129388164494,
      "loss": 2.8602,
      "theoretical_loss": 3.645554785131168,
      "tokens_seen": 1010545664
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003504212637913741,
      "loss": 2.721,
      "theoretical_loss": 3.645532330132543,
      "tokens_seen": 1010611200
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00035041123370110335,
      "loss": 2.6924,
      "theoretical_loss": 3.6455098769977248,
      "tokens_seen": 1010676736
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003504012036108325,
      "loss": 2.9447,
      "theoretical_loss": 3.645487425726437,
      "tokens_seen": 1010742272
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003503911735205617,
      "loss": 2.7244,
      "theoretical_loss": 3.6454649763184044,
      "tokens_seen": 1010807808
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003503811434302909,
      "loss": 2.8867,
      "theoretical_loss": 3.6454425287733527,
      "tokens_seen": 1010873344
    },
    {
      "epoch": 3.04,
      "objective/train/docs_used": 1166720,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7130308151245117,
      "objective/train/theoretical_loss": 3.6454369171781478,
      "objective/train/tokens_used": 1031349728,
      "theoretical_loss": 3.6454369171781478,
      "tokens_seen": 1010889728
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003503711133400201,
      "loss": 2.7468,
      "theoretical_loss": 3.645420083091005,
      "tokens_seen": 1010938880
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00035036108324974926,
      "loss": 2.7862,
      "theoretical_loss": 3.6453976392710867,
      "tokens_seen": 1011004416
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00035035105315947844,
      "loss": 2.9374,
      "theoretical_loss": 3.6453751973133226,
      "tokens_seen": 1011069952
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003503410230692076,
      "loss": 2.6861,
      "theoretical_loss": 3.645352757217437,
      "tokens_seen": 1011135488
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00035033099297893686,
      "loss": 2.7727,
      "theoretical_loss": 3.6453303189831554,
      "tokens_seen": 1011201024
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.000350320962888666,
      "loss": 2.9209,
      "theoretical_loss": 3.6453078826102026,
      "tokens_seen": 1011266560
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003503109327983952,
      "loss": 2.7929,
      "theoretical_loss": 3.6452854480983032,
      "tokens_seen": 1011332096
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00035030090270812435,
      "loss": 2.8713,
      "theoretical_loss": 3.6452630154471826,
      "tokens_seen": 1011397632
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003502908726178536,
      "loss": 2.8707,
      "theoretical_loss": 3.645240584656566,
      "tokens_seen": 1011463168
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00035028084252758276,
      "loss": 2.7903,
      "theoretical_loss": 3.645218155726178,
      "tokens_seen": 1011528704
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00035027081243731194,
      "loss": 2.8254,
      "theoretical_loss": 3.6451957286557444,
      "tokens_seen": 1011594240
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003502607823470411,
      "loss": 2.9904,
      "theoretical_loss": 3.64517330344499,
      "tokens_seen": 1011659776
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00035025075225677036,
      "loss": 2.8842,
      "theoretical_loss": 3.6451508800936407,
      "tokens_seen": 1011725312
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003502407221664995,
      "loss": 2.885,
      "theoretical_loss": 3.6451284586014214,
      "tokens_seen": 1011790848
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003502306920762287,
      "loss": 2.5298,
      "theoretical_loss": 3.6451060389680574,
      "tokens_seen": 1011856384
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00035022066198595785,
      "loss": 2.8897,
      "theoretical_loss": 3.645083621193275,
      "tokens_seen": 1011921920
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003502106318956871,
      "loss": 2.5561,
      "theoretical_loss": 3.6450612052767997,
      "tokens_seen": 1011987456
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00035020060180541627,
      "loss": 2.7428,
      "theoretical_loss": 3.645038791218356,
      "tokens_seen": 1012052992
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00035019057171514545,
      "loss": 3.0237,
      "theoretical_loss": 3.6450163790176706,
      "tokens_seen": 1012118528
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00035018054162487463,
      "loss": 2.6427,
      "theoretical_loss": 3.644993968674469,
      "tokens_seen": 1012184064
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003501705115346038,
      "loss": 3.029,
      "theoretical_loss": 3.644971560188477,
      "tokens_seen": 1012249600
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.000350160481444333,
      "loss": 2.8482,
      "theoretical_loss": 3.6449491535594207,
      "tokens_seen": 1012315136
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003501504513540622,
      "loss": 2.7981,
      "theoretical_loss": 3.644926748787025,
      "tokens_seen": 1012380672
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00035014042126379135,
      "loss": 2.9763,
      "theoretical_loss": 3.6449043458710175,
      "tokens_seen": 1012446208
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003501303911735206,
      "loss": 2.8806,
      "theoretical_loss": 3.6448819448111234,
      "tokens_seen": 1012511744
    },
    {
      "epoch": 3.04,
      "objective/train/docs_used": 1167486,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.346987724304199,
      "objective/train/theoretical_loss": 3.644876344836135,
      "objective/train/tokens_used": 1032988128,
      "theoretical_loss": 3.644876344836135,
      "tokens_seen": 1012528128
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003501203610832497,
      "loss": 2.9117,
      "theoretical_loss": 3.6448595456070683,
      "tokens_seen": 1012577280
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00035011033099297895,
      "loss": 2.748,
      "theoretical_loss": 3.644837148258579,
      "tokens_seen": 1012642816
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00035010030090270813,
      "loss": 2.9558,
      "theoretical_loss": 3.644814752765382,
      "tokens_seen": 1012708352
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003500902708124373,
      "loss": 2.8949,
      "theoretical_loss": 3.644792359127202,
      "tokens_seen": 1012773888
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003500802407221665,
      "loss": 2.8262,
      "theoretical_loss": 3.644769967343767,
      "tokens_seen": 1012839424
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00035007021063189573,
      "loss": 2.6742,
      "theoretical_loss": 3.644747577414803,
      "tokens_seen": 1012904960
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00035006018054162486,
      "loss": 2.7622,
      "theoretical_loss": 3.6447251893400363,
      "tokens_seen": 1012970496
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003500501504513541,
      "loss": 2.7589,
      "theoretical_loss": 3.644702803119193,
      "tokens_seen": 1013036032
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003500401203610832,
      "loss": 2.6765,
      "theoretical_loss": 3.6446804187520003,
      "tokens_seen": 1013101568
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00035003009027081245,
      "loss": 2.7089,
      "theoretical_loss": 3.644658036238184,
      "tokens_seen": 1013167104
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00035002006018054163,
      "loss": 2.7562,
      "theoretical_loss": 3.644635655577472,
      "tokens_seen": 1013232640
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003500100300902708,
      "loss": 2.9357,
      "theoretical_loss": 3.64461327676959,
      "tokens_seen": 1013298176
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00035,
      "loss": 2.6649,
      "theoretical_loss": 3.6445908998142653,
      "tokens_seen": 1013363712
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003499899699097292,
      "loss": 2.4005,
      "theoretical_loss": 3.6445685247112247,
      "tokens_seen": 1013429248
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034997993981945836,
      "loss": 3.0307,
      "theoretical_loss": 3.6445461514601947,
      "tokens_seen": 1013494784
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003499699097291876,
      "loss": 2.6786,
      "theoretical_loss": 3.6445237800609025,
      "tokens_seen": 1013560320
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003499598796389167,
      "loss": 2.9899,
      "theoretical_loss": 3.6445014105130755,
      "tokens_seen": 1013625856
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034994984954864596,
      "loss": 2.7386,
      "theoretical_loss": 3.64447904281644,
      "tokens_seen": 1013691392
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003499398194583751,
      "loss": 2.9648,
      "theoretical_loss": 3.6444566769707243,
      "tokens_seen": 1013756928
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003499297893681043,
      "loss": 2.7837,
      "theoretical_loss": 3.6444343129756547,
      "tokens_seen": 1013822464
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003499197592778335,
      "loss": 2.7852,
      "theoretical_loss": 3.6444119508309583,
      "tokens_seen": 1013888000
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003499097291875627,
      "loss": 2.8927,
      "theoretical_loss": 3.644389590536363,
      "tokens_seen": 1013953536
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034989969909729186,
      "loss": 2.9017,
      "theoretical_loss": 3.6443672320915956,
      "tokens_seen": 1014019072
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003498896690070211,
      "loss": 2.9671,
      "theoretical_loss": 3.644344875496384,
      "tokens_seen": 1014084608
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003498796389167502,
      "loss": 2.7448,
      "theoretical_loss": 3.644322520750456,
      "tokens_seen": 1014150144
    },
    {
      "epoch": 3.04,
      "objective/train/docs_used": 1168150,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5092766284942627,
      "objective/train/theoretical_loss": 3.6443169323528926,
      "objective/train/tokens_used": 1034626528,
      "theoretical_loss": 3.6443169323528926,
      "tokens_seen": 1014166528
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034986960882647946,
      "loss": 2.6011,
      "theoretical_loss": 3.644300167853538,
      "tokens_seen": 1014215680
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003498595787362086,
      "loss": 2.7115,
      "theoretical_loss": 3.6442778168053587,
      "tokens_seen": 1014281216
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003498495486459378,
      "loss": 2.6693,
      "theoretical_loss": 3.644255467605645,
      "tokens_seen": 1014346752
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034983951855566706,
      "loss": 2.6503,
      "theoretical_loss": 3.6442331202541256,
      "tokens_seen": 1014412288
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003498294884653962,
      "loss": 2.8226,
      "theoretical_loss": 3.644210774750527,
      "tokens_seen": 1014477824
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003498194583751254,
      "loss": 2.8758,
      "theoretical_loss": 3.6441884310945785,
      "tokens_seen": 1014543360
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034980942828485455,
      "loss": 2.9343,
      "theoretical_loss": 3.6441660892860064,
      "tokens_seen": 1014608896
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003497993981945838,
      "loss": 2.8126,
      "theoretical_loss": 3.6441437493245394,
      "tokens_seen": 1014674432
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034978936810431296,
      "loss": 2.891,
      "theoretical_loss": 3.6441214112099054,
      "tokens_seen": 1014739968
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034977933801404214,
      "loss": 2.9544,
      "theoretical_loss": 3.644099074941833,
      "tokens_seen": 1014805504
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003497693079237713,
      "loss": 2.8482,
      "theoretical_loss": 3.64407674052005,
      "tokens_seen": 1014871040
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034975927783350056,
      "loss": 2.7608,
      "theoretical_loss": 3.6440544079442843,
      "tokens_seen": 1014936576
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003497492477432297,
      "loss": 2.5706,
      "theoretical_loss": 3.644032077214264,
      "tokens_seen": 1015002112
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003497392176529589,
      "loss": 2.6036,
      "theoretical_loss": 3.644009748329718,
      "tokens_seen": 1015067648
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034972918756268805,
      "loss": 2.8312,
      "theoretical_loss": 3.6439874212903742,
      "tokens_seen": 1015133184
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003497191574724173,
      "loss": 2.8235,
      "theoretical_loss": 3.643965096095961,
      "tokens_seen": 1015198720
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034970912738214647,
      "loss": 2.8537,
      "theoretical_loss": 3.6439427727462075,
      "tokens_seen": 1015264256
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034969909729187565,
      "loss": 2.8265,
      "theoretical_loss": 3.643920451240841,
      "tokens_seen": 1015329792
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034968906720160483,
      "loss": 2.8553,
      "theoretical_loss": 3.6438981315795917,
      "tokens_seen": 1015395328
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.000349679037111334,
      "loss": 2.6262,
      "theoretical_loss": 3.643875813762187,
      "tokens_seen": 1015460864
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003496690070210632,
      "loss": 2.93,
      "theoretical_loss": 3.6438534977883554,
      "tokens_seen": 1015526400
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003496589769307924,
      "loss": 2.9229,
      "theoretical_loss": 3.643831183657827,
      "tokens_seen": 1015591936
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034964894684052155,
      "loss": 3.0048,
      "theoretical_loss": 3.6438088713703296,
      "tokens_seen": 1015657472
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003496389167502508,
      "loss": 2.8082,
      "theoretical_loss": 3.6437865609255917,
      "tokens_seen": 1015723008
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003496288866599799,
      "loss": 2.6069,
      "theoretical_loss": 3.6437642523233436,
      "tokens_seen": 1015788544
    },
    {
      "epoch": 3.04,
      "objective/train/docs_used": 1169618,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.381481885910034,
      "objective/train/theoretical_loss": 3.6437586754606386,
      "objective/train/tokens_used": 1036264928,
      "theoretical_loss": 3.6437586754606386,
      "tokens_seen": 1015804928
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034961885656970915,
      "loss": 2.815,
      "theoretical_loss": 3.6437419455633133,
      "tokens_seen": 1015854080
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034960882647943833,
      "loss": 2.7737,
      "theoretical_loss": 3.64371964064523,
      "tokens_seen": 1015919616
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003495987963891675,
      "loss": 2.606,
      "theoretical_loss": 3.6436973375688226,
      "tokens_seen": 1015985152
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003495887662988967,
      "loss": 2.8198,
      "theoretical_loss": 3.6436750363338204,
      "tokens_seen": 1016050688
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034957873620862593,
      "loss": 2.8019,
      "theoretical_loss": 3.643652736939953,
      "tokens_seen": 1016116224
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034956870611835506,
      "loss": 2.8078,
      "theoretical_loss": 3.643630439386949,
      "tokens_seen": 1016181760
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003495586760280843,
      "loss": 2.7872,
      "theoretical_loss": 3.6436081436745384,
      "tokens_seen": 1016247296
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003495486459378134,
      "loss": 2.7448,
      "theoretical_loss": 3.6435858498024505,
      "tokens_seen": 1016312832
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034953861584754265,
      "loss": 2.9181,
      "theoretical_loss": 3.643563557770414,
      "tokens_seen": 1016378368
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034952858575727183,
      "loss": 2.8828,
      "theoretical_loss": 3.643541267578159,
      "tokens_seen": 1016443904
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.000349518555667001,
      "loss": 2.9088,
      "theoretical_loss": 3.643518979225415,
      "tokens_seen": 1016509440
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003495085255767302,
      "loss": 2.7972,
      "theoretical_loss": 3.6434966927119117,
      "tokens_seen": 1016574976
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003494984954864594,
      "loss": 2.7734,
      "theoretical_loss": 3.6434744080373784,
      "tokens_seen": 1016640512
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034948846539618856,
      "loss": 2.6963,
      "theoretical_loss": 3.643452125201545,
      "tokens_seen": 1016706048
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003494784353059178,
      "loss": 2.9125,
      "theoretical_loss": 3.6434298442041415,
      "tokens_seen": 1016771584
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003494684052156469,
      "loss": 2.8687,
      "theoretical_loss": 3.643407565044898,
      "tokens_seen": 1016837120
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034945837512537616,
      "loss": 2.6719,
      "theoretical_loss": 3.6433852877235435,
      "tokens_seen": 1016902656
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003494483450351053,
      "loss": 2.8081,
      "theoretical_loss": 3.6433630122398086,
      "tokens_seen": 1016968192
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003494383149448345,
      "loss": 2.8221,
      "theoretical_loss": 3.643340738593423,
      "tokens_seen": 1017033728
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003494282848545637,
      "loss": 2.6224,
      "theoretical_loss": 3.643318466784117,
      "tokens_seen": 1017099264
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003494182547642929,
      "loss": 2.7745,
      "theoretical_loss": 3.6432961968116206,
      "tokens_seen": 1017164800
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034940822467402206,
      "loss": 2.7596,
      "theoretical_loss": 3.643273928675664,
      "tokens_seen": 1017230336
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003493981945837513,
      "loss": 2.5713,
      "theoretical_loss": 3.643251662375978,
      "tokens_seen": 1017295872
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003493881644934804,
      "loss": 2.7915,
      "theoretical_loss": 3.643229397912292,
      "tokens_seen": 1017361408
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034937813440320966,
      "loss": 2.6503,
      "theoretical_loss": 3.6432071352843365,
      "tokens_seen": 1017426944
    },
    {
      "epoch": 3.04,
      "objective/train/docs_used": 1170378,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.776099681854248,
      "objective/train/theoretical_loss": 3.643201569914149,
      "objective/train/tokens_used": 1037903328,
      "theoretical_loss": 3.643201569914149,
      "tokens_seen": 1017443328
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003493681043129388,
      "loss": 3.0182,
      "theoretical_loss": 3.6431848744918423,
      "tokens_seen": 1017492480
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.000349358074222668,
      "loss": 2.9773,
      "theoretical_loss": 3.6431626155345396,
      "tokens_seen": 1017558016
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003493480441323972,
      "loss": 2.822,
      "theoretical_loss": 3.6431403584121593,
      "tokens_seen": 1017623552
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003493380140421264,
      "loss": 2.7286,
      "theoretical_loss": 3.643118103124431,
      "tokens_seen": 1017689088
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034932798395185557,
      "loss": 2.7543,
      "theoretical_loss": 3.643095849671087,
      "tokens_seen": 1017754624
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034931795386158475,
      "loss": 2.7698,
      "theoretical_loss": 3.6430735980518563,
      "tokens_seen": 1017820160
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034930792377131393,
      "loss": 2.4205,
      "theoretical_loss": 3.6430513482664706,
      "tokens_seen": 1017885696
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034929789368104316,
      "loss": 2.6192,
      "theoretical_loss": 3.643029100314661,
      "tokens_seen": 1017951232
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003492878635907723,
      "loss": 2.8012,
      "theoretical_loss": 3.643006854196157,
      "tokens_seen": 1018016768
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003492778335005015,
      "loss": 2.6011,
      "theoretical_loss": 3.642984609910691,
      "tokens_seen": 1018082304
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034926780341023065,
      "loss": 2.6748,
      "theoretical_loss": 3.6429623674579927,
      "tokens_seen": 1018147840
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003492577733199599,
      "loss": 2.789,
      "theoretical_loss": 3.642940126837795,
      "tokens_seen": 1018213376
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034924774322968907,
      "loss": 2.691,
      "theoretical_loss": 3.6429178880498263,
      "tokens_seen": 1018278912
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034923771313941825,
      "loss": 2.813,
      "theoretical_loss": 3.64289565109382,
      "tokens_seen": 1018344448
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034922768304914743,
      "loss": 2.7174,
      "theoretical_loss": 3.6428734159695066,
      "tokens_seen": 1018409984
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034921765295887667,
      "loss": 2.7577,
      "theoretical_loss": 3.6428511826766172,
      "tokens_seen": 1018475520
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003492076228686058,
      "loss": 2.6778,
      "theoretical_loss": 3.6428289512148835,
      "tokens_seen": 1018541056
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034919759277833503,
      "loss": 2.6804,
      "theoretical_loss": 3.6428067215840363,
      "tokens_seen": 1018606592
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034918756268806416,
      "loss": 2.679,
      "theoretical_loss": 3.6427844937838074,
      "tokens_seen": 1018672128
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003491775325977934,
      "loss": 2.8358,
      "theoretical_loss": 3.6427622678139286,
      "tokens_seen": 1018737664
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034916750250752257,
      "loss": 2.7548,
      "theoretical_loss": 3.642740043674131,
      "tokens_seen": 1018803200
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034915747241725175,
      "loss": 2.7795,
      "theoretical_loss": 3.642717821364146,
      "tokens_seen": 1018868736
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034914744232698093,
      "loss": 2.7847,
      "theoretical_loss": 3.6426956008837053,
      "tokens_seen": 1018934272
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003491374122367101,
      "loss": 2.8055,
      "theoretical_loss": 3.642673382232541,
      "tokens_seen": 1018999808
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003491273821464393,
      "loss": 2.8508,
      "theoretical_loss": 3.642651165410385,
      "tokens_seen": 1019065344
    },
    {
      "epoch": 3.04,
      "objective/train/docs_used": 1171971,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9963715076446533,
      "objective/train/theoretical_loss": 3.6426456114905967,
      "objective/train/tokens_used": 1039541728,
      "theoretical_loss": 3.6426456114905967,
      "tokens_seen": 1019081728
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034911735205616853,
      "loss": 2.9141,
      "theoretical_loss": 3.642628950416969,
      "tokens_seen": 1019130880
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034910732196589766,
      "loss": 2.7184,
      "theoretical_loss": 3.6426067372520246,
      "tokens_seen": 1019196416
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003490972918756269,
      "loss": 2.6736,
      "theoretical_loss": 3.6425845259152836,
      "tokens_seen": 1019261952
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034908726178535613,
      "loss": 2.6398,
      "theoretical_loss": 3.642562316406478,
      "tokens_seen": 1019327488
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034907723169508526,
      "loss": 2.7368,
      "theoretical_loss": 3.642540108725341,
      "tokens_seen": 1019393024
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003490672016048145,
      "loss": 2.6245,
      "theoretical_loss": 3.6425179028716035,
      "tokens_seen": 1019458560
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003490571715145436,
      "loss": 2.7568,
      "theoretical_loss": 3.642495698844998,
      "tokens_seen": 1019524096
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034904714142427285,
      "loss": 2.9744,
      "theoretical_loss": 3.6424734966452568,
      "tokens_seen": 1019589632
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034903711133400203,
      "loss": 2.8797,
      "theoretical_loss": 3.642451296272112,
      "tokens_seen": 1019655168
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003490270812437312,
      "loss": 2.6659,
      "theoretical_loss": 3.642429097725296,
      "tokens_seen": 1019720704
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003490170511534604,
      "loss": 2.8605,
      "theoretical_loss": 3.642406901004542,
      "tokens_seen": 1019786240
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003490070210631896,
      "loss": 2.6881,
      "theoretical_loss": 3.6423847061095813,
      "tokens_seen": 1019851776
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034899699097291876,
      "loss": 2.6267,
      "theoretical_loss": 3.6423625130401467,
      "tokens_seen": 1019917312
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.000348986960882648,
      "loss": 2.8009,
      "theoretical_loss": 3.6423403217959707,
      "tokens_seen": 1019982848
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003489769307923771,
      "loss": 2.8462,
      "theoretical_loss": 3.6423181323767864,
      "tokens_seen": 1020048384
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034896690070210636,
      "loss": 2.9541,
      "theoretical_loss": 3.642295944782326,
      "tokens_seen": 1020113920
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003489568706118355,
      "loss": 2.6883,
      "theoretical_loss": 3.6422737590123226,
      "tokens_seen": 1020179456
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003489468405215647,
      "loss": 2.6475,
      "theoretical_loss": 3.6422515750665085,
      "tokens_seen": 1020244992
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003489368104312939,
      "loss": 2.7722,
      "theoretical_loss": 3.6422293929446172,
      "tokens_seen": 1020310528
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003489267803410231,
      "loss": 2.7078,
      "theoretical_loss": 3.6422072126463814,
      "tokens_seen": 1020376064
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034891675025075226,
      "loss": 2.7822,
      "theoretical_loss": 3.6421850341715336,
      "tokens_seen": 1020441600
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003489067201604815,
      "loss": 2.6382,
      "theoretical_loss": 3.642162857519807,
      "tokens_seen": 1020507136
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003488966900702106,
      "loss": 2.6841,
      "theoretical_loss": 3.6421406826909353,
      "tokens_seen": 1020572672
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034888665997993986,
      "loss": 2.8371,
      "theoretical_loss": 3.642118509684651,
      "tokens_seen": 1020638208
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.000348876629889669,
      "loss": 2.7937,
      "theoretical_loss": 3.642096338500687,
      "tokens_seen": 1020703744
    },
    {
      "epoch": 3.04,
      "objective/train/docs_used": 1172849,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6258201599121094,
      "objective/train/theoretical_loss": 3.6420907959894024,
      "objective/train/tokens_used": 1041180128,
      "theoretical_loss": 3.6420907959894024,
      "tokens_seen": 1020720128
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003488665997993982,
      "loss": 2.9837,
      "theoretical_loss": 3.6420741691387772,
      "tokens_seen": 1020769280
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003488565697091274,
      "loss": 2.67,
      "theoretical_loss": 3.642052001598655,
      "tokens_seen": 1020834816
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003488465396188566,
      "loss": 2.7801,
      "theoretical_loss": 3.6420298358800527,
      "tokens_seen": 1020900352
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034883650952858577,
      "loss": 2.6078,
      "theoretical_loss": 3.6420076719827046,
      "tokens_seen": 1020965888
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034882647943831495,
      "loss": 2.8249,
      "theoretical_loss": 3.6419855099063443,
      "tokens_seen": 1021031424
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034881644934804413,
      "loss": 2.9316,
      "theoretical_loss": 3.641963349650705,
      "tokens_seen": 1021096960
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034880641925777336,
      "loss": 2.6782,
      "theoretical_loss": 3.64194119121552,
      "tokens_seen": 1021162496
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003487963891675025,
      "loss": 2.8126,
      "theoretical_loss": 3.641919034600523,
      "tokens_seen": 1021228032
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003487863590772317,
      "loss": 2.7651,
      "theoretical_loss": 3.6418968798054485,
      "tokens_seen": 1021293568
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034877632898696085,
      "loss": 2.8099,
      "theoretical_loss": 3.6418747268300296,
      "tokens_seen": 1021359104
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003487662988966901,
      "loss": 2.6514,
      "theoretical_loss": 3.6418525756739992,
      "tokens_seen": 1021424640
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034875626880641927,
      "loss": 2.7745,
      "theoretical_loss": 3.6418304263370933,
      "tokens_seen": 1021490176
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034874623871614845,
      "loss": 2.7871,
      "theoretical_loss": 3.6418082788190436,
      "tokens_seen": 1021555712
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034873620862587763,
      "loss": 2.4823,
      "theoretical_loss": 3.6417861331195853,
      "tokens_seen": 1021621248
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034872617853560687,
      "loss": 2.7061,
      "theoretical_loss": 3.6417639892384526,
      "tokens_seen": 1021686784
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.000348716148445336,
      "loss": 2.7506,
      "theoretical_loss": 3.6417418471753793,
      "tokens_seen": 1021752320
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034870611835506523,
      "loss": 2.7926,
      "theoretical_loss": 3.6417197069300986,
      "tokens_seen": 1021817856
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034869608826479436,
      "loss": 2.745,
      "theoretical_loss": 3.6416975685023463,
      "tokens_seen": 1021883392
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003486860581745236,
      "loss": 2.4741,
      "theoretical_loss": 3.6416754318918554,
      "tokens_seen": 1021948928
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034867602808425277,
      "loss": 2.8679,
      "theoretical_loss": 3.641653297098361,
      "tokens_seen": 1022014464
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034866599799398195,
      "loss": 2.9177,
      "theoretical_loss": 3.6416311641215966,
      "tokens_seen": 1022080000
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034865596790371113,
      "loss": 2.7663,
      "theoretical_loss": 3.6416090329612976,
      "tokens_seen": 1022145536
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003486459378134403,
      "loss": 2.794,
      "theoretical_loss": 3.6415869036171973,
      "tokens_seen": 1022211072
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003486359077231695,
      "loss": 2.6762,
      "theoretical_loss": 3.641564776089032,
      "tokens_seen": 1022276608
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034862587763289873,
      "loss": 2.6958,
      "theoretical_loss": 3.6415426503765342,
      "tokens_seen": 1022342144
    },
    {
      "epoch": 3.04,
      "objective/train/docs_used": 1174219,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4945321083068848,
      "objective/train/theoretical_loss": 3.641537119232077,
      "objective/train/tokens_used": 1042818528,
      "theoretical_loss": 3.641537119232077,
      "tokens_seen": 1022358528
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034861584754262786,
      "loss": 2.6434,
      "theoretical_loss": 3.6415205264794404,
      "tokens_seen": 1022407680
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003486058174523571,
      "loss": 2.8606,
      "theoretical_loss": 3.641498404397484,
      "tokens_seen": 1022473216
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003485957873620862,
      "loss": 2.6391,
      "theoretical_loss": 3.6414762841304,
      "tokens_seen": 1022538752
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034858575727181546,
      "loss": 2.6538,
      "theoretical_loss": 3.6414541656779233,
      "tokens_seen": 1022604288
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034857572718154464,
      "loss": 2.664,
      "theoretical_loss": 3.6414320490397896,
      "tokens_seen": 1022669824
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003485656970912738,
      "loss": 2.7767,
      "theoretical_loss": 3.6414099342157327,
      "tokens_seen": 1022735360
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.000348555667001003,
      "loss": 2.6541,
      "theoretical_loss": 3.6413878212054875,
      "tokens_seen": 1022800896
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034854563691073223,
      "loss": 2.902,
      "theoretical_loss": 3.64136571000879,
      "tokens_seen": 1022866432
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034853560682046136,
      "loss": 2.764,
      "theoretical_loss": 3.641343600625375,
      "tokens_seen": 1022931968
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003485255767301906,
      "loss": 2.7848,
      "theoretical_loss": 3.641321493054977,
      "tokens_seen": 1022997504
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003485155466399197,
      "loss": 2.6366,
      "theoretical_loss": 3.6412993872973316,
      "tokens_seen": 1023063040
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034850551654964896,
      "loss": 2.6445,
      "theoretical_loss": 3.6412772833521743,
      "tokens_seen": 1023128576
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034849548645937814,
      "loss": 2.8639,
      "theoretical_loss": 3.6412551812192397,
      "tokens_seen": 1023194112
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003484854563691073,
      "loss": 2.8851,
      "theoretical_loss": 3.641233080898264,
      "tokens_seen": 1023259648
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003484754262788365,
      "loss": 2.7964,
      "theoretical_loss": 3.6412109823889818,
      "tokens_seen": 1023325184
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003484653961885657,
      "loss": 2.8088,
      "theoretical_loss": 3.64118888569113,
      "tokens_seen": 1023390720
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034845536609829486,
      "loss": 2.6254,
      "theoretical_loss": 3.641166790804442,
      "tokens_seen": 1023456256
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003484453360080241,
      "loss": 2.6364,
      "theoretical_loss": 3.641144697728655,
      "tokens_seen": 1023521792
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003484353059177532,
      "loss": 2.822,
      "theoretical_loss": 3.6411226064635045,
      "tokens_seen": 1023587328
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034842527582748246,
      "loss": 2.6328,
      "theoretical_loss": 3.6411005170087254,
      "tokens_seen": 1023652864
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034841524573721164,
      "loss": 2.8881,
      "theoretical_loss": 3.641078429364054,
      "tokens_seen": 1023718400
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003484052156469408,
      "loss": 2.8342,
      "theoretical_loss": 3.6410563435292262,
      "tokens_seen": 1023783936
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034839518555667,
      "loss": 2.63,
      "theoretical_loss": 3.6410342595039777,
      "tokens_seen": 1023849472
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003483851554663992,
      "loss": 2.7483,
      "theoretical_loss": 3.6410121772880446,
      "tokens_seen": 1023915008
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034837512537612837,
      "loss": 2.6859,
      "theoretical_loss": 3.640990096881162,
      "tokens_seen": 1023980544
    },
    {
      "epoch": 3.04,
      "objective/train/docs_used": 1174839,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9092800617218018,
      "objective/train/theoretical_loss": 3.640984577062075,
      "objective/train/tokens_used": 1044456928,
      "theoretical_loss": 3.640984577062075,
      "tokens_seen": 1023996928
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003483650952858576,
      "loss": 2.8045,
      "theoretical_loss": 3.6409680182830675,
      "tokens_seen": 1024046080
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034835506519558673,
      "loss": 2.8116,
      "theoretical_loss": 3.6409459414934955,
      "tokens_seen": 1024111616
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034834503510531597,
      "loss": 2.7341,
      "theoretical_loss": 3.6409238665121832,
      "tokens_seen": 1024177152
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034833500501504515,
      "loss": 2.4435,
      "theoretical_loss": 3.640901793338867,
      "tokens_seen": 1024242688
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034832497492477433,
      "loss": 2.6926,
      "theoretical_loss": 3.640879721973282,
      "tokens_seen": 1024308224
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034831494483450356,
      "loss": 2.8559,
      "theoretical_loss": 3.6408576524151655,
      "tokens_seen": 1024373760
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003483049147442327,
      "loss": 2.7453,
      "theoretical_loss": 3.6408355846642535,
      "tokens_seen": 1024439296
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003482948846539619,
      "loss": 2.7137,
      "theoretical_loss": 3.6408135187202824,
      "tokens_seen": 1024504832
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034828485456369105,
      "loss": 2.573,
      "theoretical_loss": 3.640791454582989,
      "tokens_seen": 1024570368
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003482748244734203,
      "loss": 2.6886,
      "theoretical_loss": 3.640769392252109,
      "tokens_seen": 1024635904
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034826479438314947,
      "loss": 2.9169,
      "theoretical_loss": 3.64074733172738,
      "tokens_seen": 1024701440
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034825476429287865,
      "loss": 2.6807,
      "theoretical_loss": 3.640725273008538,
      "tokens_seen": 1024766976
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034824473420260783,
      "loss": 2.8173,
      "theoretical_loss": 3.64070321609532,
      "tokens_seen": 1024832512
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034823470411233707,
      "loss": 2.8759,
      "theoretical_loss": 3.6406811609874628,
      "tokens_seen": 1024898048
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003482246740220662,
      "loss": 2.6937,
      "theoretical_loss": 3.640659107684703,
      "tokens_seen": 1024963584
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034821464393179543,
      "loss": 2.8059,
      "theoretical_loss": 3.640637056186777,
      "tokens_seen": 1025029120
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034820461384152456,
      "loss": 2.7328,
      "theoretical_loss": 3.6406150064934226,
      "tokens_seen": 1025094656
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003481945837512538,
      "loss": 2.8426,
      "theoretical_loss": 3.6405929586043766,
      "tokens_seen": 1025160192
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034818455366098297,
      "loss": 2.5166,
      "theoretical_loss": 3.640570912519375,
      "tokens_seen": 1025225728
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034817452357071215,
      "loss": 2.7644,
      "theoretical_loss": 3.640548868238157,
      "tokens_seen": 1025291264
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034816449348044133,
      "loss": 2.7304,
      "theoretical_loss": 3.640526825760457,
      "tokens_seen": 1025356800
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003481544633901705,
      "loss": 2.6769,
      "theoretical_loss": 3.6405047850860144,
      "tokens_seen": 1025422336
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003481444332998997,
      "loss": 2.8822,
      "theoretical_loss": 3.6404827462145652,
      "tokens_seen": 1025487872
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034813440320962893,
      "loss": 2.8437,
      "theoretical_loss": 3.6404607091458474,
      "tokens_seen": 1025553408
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034812437311935806,
      "loss": 2.6955,
      "theoretical_loss": 3.640438673879598,
      "tokens_seen": 1025618944
    },
    {
      "epoch": 3.04,
      "objective/train/docs_used": 1176332,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5118682384490967,
      "objective/train/theoretical_loss": 3.640433165344641,
      "objective/train/tokens_used": 1046095328,
      "theoretical_loss": 3.640433165344641,
      "tokens_seen": 1025635328
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003481143430290873,
      "loss": 2.6492,
      "theoretical_loss": 3.6404166404155545,
      "tokens_seen": 1025684480
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003481043129388164,
      "loss": 2.468,
      "theoretical_loss": 3.6403946087534544,
      "tokens_seen": 1025750016
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034809428284854566,
      "loss": 2.7733,
      "theoretical_loss": 3.6403725788930354,
      "tokens_seen": 1025815552
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034808425275827484,
      "loss": 2.8966,
      "theoretical_loss": 3.6403505508340346,
      "tokens_seen": 1025881088
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.000348074222668004,
      "loss": 2.8089,
      "theoretical_loss": 3.6403285245761903,
      "tokens_seen": 1025946624
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003480641925777332,
      "loss": 2.7961,
      "theoretical_loss": 3.640306500119239,
      "tokens_seen": 1026012160
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034805416248746243,
      "loss": 2.691,
      "theoretical_loss": 3.64028447746292,
      "tokens_seen": 1026077696
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034804413239719156,
      "loss": 2.7155,
      "theoretical_loss": 3.64026245660697,
      "tokens_seen": 1026143232
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003480341023069208,
      "loss": 2.7533,
      "theoretical_loss": 3.6402404375511273,
      "tokens_seen": 1026208768
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003480240722166499,
      "loss": 2.8656,
      "theoretical_loss": 3.6402184202951293,
      "tokens_seen": 1026274304
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034801404212637916,
      "loss": 2.9105,
      "theoretical_loss": 3.640196404838715,
      "tokens_seen": 1026339840
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034800401203610834,
      "loss": 2.545,
      "theoretical_loss": 3.6401743911816213,
      "tokens_seen": 1026405376
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003479939819458375,
      "loss": 2.7461,
      "theoretical_loss": 3.6401523793235864,
      "tokens_seen": 1026470912
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003479839518555667,
      "loss": 2.6097,
      "theoretical_loss": 3.640130369264349,
      "tokens_seen": 1026536448
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003479739217652959,
      "loss": 2.6789,
      "theoretical_loss": 3.640108361003647,
      "tokens_seen": 1026601984
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034796389167502506,
      "loss": 2.8538,
      "theoretical_loss": 3.6400863545412188,
      "tokens_seen": 1026667520
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003479538615847543,
      "loss": 2.9017,
      "theoretical_loss": 3.640064349876803,
      "tokens_seen": 1026733056
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034794383149448343,
      "loss": 2.6433,
      "theoretical_loss": 3.6400423470101364,
      "tokens_seen": 1026798592
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034793380140421266,
      "loss": 2.8106,
      "theoretical_loss": 3.6400203459409592,
      "tokens_seen": 1026864128
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034792377131394184,
      "loss": 2.5955,
      "theoretical_loss": 3.6399983466690085,
      "tokens_seen": 1026929664
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.000347913741223671,
      "loss": 2.7773,
      "theoretical_loss": 3.639976349194024,
      "tokens_seen": 1026995200
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003479037111334002,
      "loss": 2.7691,
      "theoretical_loss": 3.6399543535157433,
      "tokens_seen": 1027060736
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003478936810431294,
      "loss": 2.9683,
      "theoretical_loss": 3.639932359633906,
      "tokens_seen": 1027126272
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034788365095285857,
      "loss": 2.7609,
      "theoretical_loss": 3.63991036754825,
      "tokens_seen": 1027191808
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003478736208625878,
      "loss": 2.8229,
      "theoretical_loss": 3.639888377258514,
      "tokens_seen": 1027257344
    },
    {
      "epoch": 3.04,
      "objective/train/docs_used": 1176787,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9666268825531006,
      "objective/train/theoretical_loss": 3.639882879966662,
      "objective/train/tokens_used": 1047733728,
      "theoretical_loss": 3.639882879966662,
      "tokens_seen": 1027273728
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034786359077231693,
      "loss": 2.7606,
      "theoretical_loss": 3.639866388764437,
      "tokens_seen": 1027322880
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034785356068204617,
      "loss": 2.8858,
      "theoretical_loss": 3.6398444020657577,
      "tokens_seen": 1027388416
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003478435305917753,
      "loss": 2.8103,
      "theoretical_loss": 3.6398224171622156,
      "tokens_seen": 1027453952
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034783350050150453,
      "loss": 2.8517,
      "theoretical_loss": 3.639800434053549,
      "tokens_seen": 1027519488
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003478234704112337,
      "loss": 2.7297,
      "theoretical_loss": 3.6397784527394967,
      "tokens_seen": 1027585024
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003478134403209629,
      "loss": 3.0084,
      "theoretical_loss": 3.6397564732197987,
      "tokens_seen": 1027650560
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034780341023069207,
      "loss": 2.8202,
      "theoretical_loss": 3.639734495494193,
      "tokens_seen": 1027716096
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034779338014042125,
      "loss": 2.7186,
      "theoretical_loss": 3.6397125195624196,
      "tokens_seen": 1027781632
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034778335005015043,
      "loss": 2.8238,
      "theoretical_loss": 3.6396905454242177,
      "tokens_seen": 1027847168
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034777331995987967,
      "loss": 2.8337,
      "theoretical_loss": 3.6396685730793257,
      "tokens_seen": 1027912704
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003477632898696088,
      "loss": 2.7286,
      "theoretical_loss": 3.6396466025274834,
      "tokens_seen": 1027978240
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034775325977933803,
      "loss": 2.846,
      "theoretical_loss": 3.6396246337684306,
      "tokens_seen": 1028043776
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003477432296890672,
      "loss": 2.7812,
      "theoretical_loss": 3.639602666801906,
      "tokens_seen": 1028109312
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003477331995987964,
      "loss": 2.6417,
      "theoretical_loss": 3.6395807016276502,
      "tokens_seen": 1028174848
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003477231695085256,
      "loss": 2.6741,
      "theoretical_loss": 3.6395587382454018,
      "tokens_seen": 1028240384
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034771313941825476,
      "loss": 2.9862,
      "theoretical_loss": 3.639536776654901,
      "tokens_seen": 1028305920
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034770310932798394,
      "loss": 2.9435,
      "theoretical_loss": 3.6395148168558866,
      "tokens_seen": 1028371456
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034769307923771317,
      "loss": 2.5491,
      "theoretical_loss": 3.639492858848099,
      "tokens_seen": 1028436992
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003476830491474423,
      "loss": 2.8525,
      "theoretical_loss": 3.6394709026312775,
      "tokens_seen": 1028502528
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034767301905717153,
      "loss": 2.7869,
      "theoretical_loss": 3.6394489482051626,
      "tokens_seen": 1028568064
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034766298896690066,
      "loss": 2.9154,
      "theoretical_loss": 3.639426995569494,
      "tokens_seen": 1028633600
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003476529588766299,
      "loss": 2.6809,
      "theoretical_loss": 3.6394050447240107,
      "tokens_seen": 1028699136
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003476429287863591,
      "loss": 2.9911,
      "theoretical_loss": 3.6393830956684536,
      "tokens_seen": 1028764672
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034763289869608826,
      "loss": 2.6227,
      "theoretical_loss": 3.6393611484025628,
      "tokens_seen": 1028830208
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034762286860581744,
      "loss": 2.775,
      "theoretical_loss": 3.639339202926078,
      "tokens_seen": 1028895744
    },
    {
      "epoch": 3.04,
      "objective/train/docs_used": 1178361,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4620256423950195,
      "objective/train/theoretical_loss": 3.6393337168365205,
      "objective/train/tokens_used": 1049372128,
      "theoretical_loss": 3.6393337168365205,
      "tokens_seen": 1028912128
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003476128385155466,
      "loss": 2.7881,
      "theoretical_loss": 3.639317259238739,
      "tokens_seen": 1028961280
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003476028084252758,
      "loss": 2.8292,
      "theoretical_loss": 3.639295317340287,
      "tokens_seen": 1029026816
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034759277833500504,
      "loss": 2.6422,
      "theoretical_loss": 3.6392733772304613,
      "tokens_seen": 1029092352
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003475827482447342,
      "loss": 2.971,
      "theoretical_loss": 3.639251438909003,
      "tokens_seen": 1029157888
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003475727181544634,
      "loss": 2.6485,
      "theoretical_loss": 3.6392295023756516,
      "tokens_seen": 1029223424
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034756268806419263,
      "loss": 2.9472,
      "theoretical_loss": 3.6392075676301485,
      "tokens_seen": 1029288960
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034755265797392176,
      "loss": 2.6777,
      "theoretical_loss": 3.6391856346722333,
      "tokens_seen": 1029354496
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.000347542627883651,
      "loss": 2.6996,
      "theoretical_loss": 3.639163703501647,
      "tokens_seen": 1029420032
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003475325977933801,
      "loss": 2.8,
      "theoretical_loss": 3.63914177411813,
      "tokens_seen": 1029485568
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034752256770310936,
      "loss": 2.7562,
      "theoretical_loss": 3.6391198465214227,
      "tokens_seen": 1029551104
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034751253761283854,
      "loss": 2.6978,
      "theoretical_loss": 3.6390979207112664,
      "tokens_seen": 1029616640
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003475025075225677,
      "loss": 2.6952,
      "theoretical_loss": 3.639075996687401,
      "tokens_seen": 1029682176
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003474924774322969,
      "loss": 2.7032,
      "theoretical_loss": 3.6390540744495685,
      "tokens_seen": 1029747712
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003474824473420261,
      "loss": 2.6098,
      "theoretical_loss": 3.639032153997509,
      "tokens_seen": 1029813248
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034747241725175527,
      "loss": 2.9094,
      "theoretical_loss": 3.6390102353309626,
      "tokens_seen": 1029878784
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003474623871614845,
      "loss": 2.6849,
      "theoretical_loss": 3.6389883184496723,
      "tokens_seen": 1029944320
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034745235707121363,
      "loss": 2.8267,
      "theoretical_loss": 3.638966403353377,
      "tokens_seen": 1030009856
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034744232698094286,
      "loss": 2.8079,
      "theoretical_loss": 3.6389444900418186,
      "tokens_seen": 1030075392
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034743229689067204,
      "loss": 2.6925,
      "theoretical_loss": 3.6389225785147388,
      "tokens_seen": 1030140928
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003474222668004012,
      "loss": 2.8071,
      "theoretical_loss": 3.638900668771878,
      "tokens_seen": 1030206464
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003474122367101304,
      "loss": 2.8804,
      "theoretical_loss": 3.638878760812978,
      "tokens_seen": 1030272000
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003474022066198596,
      "loss": 2.6228,
      "theoretical_loss": 3.6388568546377793,
      "tokens_seen": 1030337536
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034739217652958877,
      "loss": 2.6782,
      "theoretical_loss": 3.6388349502460238,
      "tokens_seen": 1030403072
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.000347382146439318,
      "loss": 2.8143,
      "theoretical_loss": 3.638813047637453,
      "tokens_seen": 1030468608
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034737211634904713,
      "loss": 2.5234,
      "theoretical_loss": 3.6387911468118075,
      "tokens_seen": 1030534144
    },
    {
      "epoch": 3.04,
      "objective/train/docs_used": 1179088,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.1466615200042725,
      "objective/train/theoretical_loss": 3.6387856718839484,
      "objective/train/tokens_used": 1051010528,
      "theoretical_loss": 3.6387856718839484,
      "tokens_seen": 1030550528
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034736208625877637,
      "loss": 2.859,
      "theoretical_loss": 3.6387692477688303,
      "tokens_seen": 1030599680
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003473520561685055,
      "loss": 2.7063,
      "theoretical_loss": 3.6387473505082615,
      "tokens_seen": 1030665216
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034734202607823473,
      "loss": 2.4725,
      "theoretical_loss": 3.6387254550298436,
      "tokens_seen": 1030730752
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003473319959879639,
      "loss": 2.8637,
      "theoretical_loss": 3.6387035613333176,
      "tokens_seen": 1030796288
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003473219658976931,
      "loss": 2.9199,
      "theoretical_loss": 3.6386816694184256,
      "tokens_seen": 1030861824
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034731193580742227,
      "loss": 2.8224,
      "theoretical_loss": 3.6386597792849096,
      "tokens_seen": 1030927360
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034730190571715145,
      "loss": 2.6093,
      "theoretical_loss": 3.638637890932511,
      "tokens_seen": 1030992896
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034729187562688063,
      "loss": 2.783,
      "theoretical_loss": 3.638616004360972,
      "tokens_seen": 1031058432
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034728184553660987,
      "loss": 2.7837,
      "theoretical_loss": 3.6385941195700338,
      "tokens_seen": 1031123968
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.000347271815446339,
      "loss": 2.644,
      "theoretical_loss": 3.638572236559439,
      "tokens_seen": 1031189504
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034726178535606823,
      "loss": 2.708,
      "theoretical_loss": 3.63855035532893,
      "tokens_seen": 1031255040
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003472517552657974,
      "loss": 2.8875,
      "theoretical_loss": 3.638528475878248,
      "tokens_seen": 1031320576
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003472417251755266,
      "loss": 2.712,
      "theoretical_loss": 3.6385065982071354,
      "tokens_seen": 1031386112
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003472316950852558,
      "loss": 2.7669,
      "theoretical_loss": 3.6384847223153347,
      "tokens_seen": 1031451648
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034722166499498496,
      "loss": 2.8086,
      "theoretical_loss": 3.6384628482025887,
      "tokens_seen": 1031517184
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034721163490471414,
      "loss": 2.6404,
      "theoretical_loss": 3.6384409758686385,
      "tokens_seen": 1031582720
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034720160481444337,
      "loss": 2.6375,
      "theoretical_loss": 3.6384191053132264,
      "tokens_seen": 1031648256
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003471915747241725,
      "loss": 2.8026,
      "theoretical_loss": 3.638397236536096,
      "tokens_seen": 1031713792
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034718154463390173,
      "loss": 2.5592,
      "theoretical_loss": 3.6383753695369894,
      "tokens_seen": 1031779328
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034717151454363086,
      "loss": 2.6478,
      "theoretical_loss": 3.6383535043156483,
      "tokens_seen": 1031844864
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003471614844533601,
      "loss": 2.977,
      "theoretical_loss": 3.638331640871816,
      "tokens_seen": 1031910400
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003471514543630893,
      "loss": 2.4206,
      "theoretical_loss": 3.638309779205235,
      "tokens_seen": 1031975936
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034714142427281846,
      "loss": 2.4946,
      "theoretical_loss": 3.6382879193156485,
      "tokens_seen": 1032041472
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034713139418254764,
      "loss": 2.6483,
      "theoretical_loss": 3.6382660612027977,
      "tokens_seen": 1032107008
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003471213640922768,
      "loss": 2.6798,
      "theoretical_loss": 3.638244204866427,
      "tokens_seen": 1032172544
    },
    {
      "epoch": 3.04,
      "objective/train/docs_used": 1180390,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7738821506500244,
      "objective/train/theoretical_loss": 3.6382387410598787,
      "objective/train/tokens_used": 1052648928,
      "theoretical_loss": 3.6382387410598787,
      "tokens_seen": 1032188928
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.000347111334002006,
      "loss": 2.8191,
      "theoretical_loss": 3.638222350306278,
      "tokens_seen": 1032238080
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034710130391173524,
      "loss": 2.733,
      "theoretical_loss": 3.638200497522095,
      "tokens_seen": 1032303616
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034709127382146436,
      "loss": 2.7866,
      "theoretical_loss": 3.6381786465136194,
      "tokens_seen": 1032369152
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003470812437311936,
      "loss": 2.7378,
      "theoretical_loss": 3.6381567972805957,
      "tokens_seen": 1032434688
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003470712136409228,
      "loss": 2.5182,
      "theoretical_loss": 3.6381349498227653,
      "tokens_seen": 1032500224
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034706118355065196,
      "loss": 2.5097,
      "theoretical_loss": 3.6381131041398724,
      "tokens_seen": 1032565760
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034705115346038114,
      "loss": 2.8444,
      "theoretical_loss": 3.6380912602316604,
      "tokens_seen": 1032631296
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003470411233701103,
      "loss": 2.9808,
      "theoretical_loss": 3.6380694180978717,
      "tokens_seen": 1032696832
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003470310932798395,
      "loss": 2.7574,
      "theoretical_loss": 3.6380475777382504,
      "tokens_seen": 1032762368
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034702106318956874,
      "loss": 2.7757,
      "theoretical_loss": 3.638025739152539,
      "tokens_seen": 1032827904
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034701103309929787,
      "loss": 2.676,
      "theoretical_loss": 3.638003902340481,
      "tokens_seen": 1032893440
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003470010030090271,
      "loss": 2.6281,
      "theoretical_loss": 3.63798206730182,
      "tokens_seen": 1032958976
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034699097291875623,
      "loss": 2.5487,
      "theoretical_loss": 3.6379602340363,
      "tokens_seen": 1033024512
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034698094282848547,
      "loss": 2.8011,
      "theoretical_loss": 3.637938402543664,
      "tokens_seen": 1033090048
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034697091273821465,
      "loss": 2.71,
      "theoretical_loss": 3.6379165728236553,
      "tokens_seen": 1033155584
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034696088264794383,
      "loss": 2.7223,
      "theoretical_loss": 3.637894744876018,
      "tokens_seen": 1033221120
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.000346950852557673,
      "loss": 2.4191,
      "theoretical_loss": 3.637872918700496,
      "tokens_seen": 1033286656
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034694082246740224,
      "loss": 2.9634,
      "theoretical_loss": 3.6378510942968325,
      "tokens_seen": 1033352192
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034693079237713137,
      "loss": 2.8381,
      "theoretical_loss": 3.6378292716647715,
      "tokens_seen": 1033417728
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003469207622868606,
      "loss": 2.7367,
      "theoretical_loss": 3.637807450804057,
      "tokens_seen": 1033483264
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034691073219658973,
      "loss": 2.5099,
      "theoretical_loss": 3.637785631714433,
      "tokens_seen": 1033548800
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034690070210631897,
      "loss": 2.8073,
      "theoretical_loss": 3.637763814395643,
      "tokens_seen": 1033614336
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034689067201604815,
      "loss": 2.9439,
      "theoretical_loss": 3.6377419988474315,
      "tokens_seen": 1033679872
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034688064192577733,
      "loss": 2.8287,
      "theoretical_loss": 3.6377201850695418,
      "tokens_seen": 1033745408
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003468706118355065,
      "loss": 2.6707,
      "theoretical_loss": 3.6376983730617187,
      "tokens_seen": 1033810944
    },
    {
      "epoch": 3.04,
      "objective/train/docs_used": 1181093,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.383265256881714,
      "objective/train/theoretical_loss": 3.637692920336306,
      "objective/train/tokens_used": 1054287328,
      "theoretical_loss": 3.637692920336306,
      "tokens_seen": 1033827328
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003468605817452357,
      "loss": 2.6716,
      "theoretical_loss": 3.6376765628237067,
      "tokens_seen": 1033876480
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003468505516549649,
      "loss": 2.4495,
      "theoretical_loss": 3.6376547543552493,
      "tokens_seen": 1033942016
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003468405215646941,
      "loss": 2.8331,
      "theoretical_loss": 3.6376329476560914,
      "tokens_seen": 1034007552
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003468304914744233,
      "loss": 2.8199,
      "theoretical_loss": 3.6376111427259765,
      "tokens_seen": 1034073088
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034682046138415247,
      "loss": 2.6642,
      "theoretical_loss": 3.6375893395646495,
      "tokens_seen": 1034138624
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034681043129388165,
      "loss": 2.6186,
      "theoretical_loss": 3.637567538171855,
      "tokens_seen": 1034204160
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034680040120361083,
      "loss": 2.8684,
      "theoretical_loss": 3.637545738547337,
      "tokens_seen": 1034269696
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034679037111334007,
      "loss": 2.6562,
      "theoretical_loss": 3.6375239406908406,
      "tokens_seen": 1034335232
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003467803410230692,
      "loss": 2.7554,
      "theoretical_loss": 3.6375021446021103,
      "tokens_seen": 1034400768
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034677031093279843,
      "loss": 3.0065,
      "theoretical_loss": 3.63748035028089,
      "tokens_seen": 1034466304
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003467602808425276,
      "loss": 2.8307,
      "theoretical_loss": 3.6374585577269256,
      "tokens_seen": 1034531840
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003467502507522568,
      "loss": 2.647,
      "theoretical_loss": 3.6374367669399605,
      "tokens_seen": 1034597376
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.000346740220661986,
      "loss": 2.8376,
      "theoretical_loss": 3.637414977919741,
      "tokens_seen": 1034662912
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034673019057171516,
      "loss": 2.7634,
      "theoretical_loss": 3.6373931906660113,
      "tokens_seen": 1034728448
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034672016048144434,
      "loss": 2.6724,
      "theoretical_loss": 3.6373714051785155,
      "tokens_seen": 1034793984
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034671013039117357,
      "loss": 2.6667,
      "theoretical_loss": 3.637349621457,
      "tokens_seen": 1034859520
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003467001003009027,
      "loss": 3.0254,
      "theoretical_loss": 3.6373278395012085,
      "tokens_seen": 1034925056
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034669007021063193,
      "loss": 2.7656,
      "theoretical_loss": 3.6373060593108875,
      "tokens_seen": 1034990592
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034668004012036106,
      "loss": 2.6329,
      "theoretical_loss": 3.6372842808857806,
      "tokens_seen": 1035056128
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003466700100300903,
      "loss": 2.7772,
      "theoretical_loss": 3.6372625042256344,
      "tokens_seen": 1035121664
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003466599799398195,
      "loss": 2.7761,
      "theoretical_loss": 3.637240729330193,
      "tokens_seen": 1035187200
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034664994984954866,
      "loss": 2.7121,
      "theoretical_loss": 3.637218956199202,
      "tokens_seen": 1035252736
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034663991975927784,
      "loss": 2.8773,
      "theoretical_loss": 3.637197184832407,
      "tokens_seen": 1035318272
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.000346629889669007,
      "loss": 2.7028,
      "theoretical_loss": 3.6371754152295535,
      "tokens_seen": 1035383808
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003466198595787362,
      "loss": 2.7789,
      "theoretical_loss": 3.6371536473903863,
      "tokens_seen": 1035449344
    },
    {
      "epoch": 3.04,
      "objective/train/docs_used": 1182527,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.771360158920288,
      "objective/train/theoretical_loss": 3.637148205706141,
      "objective/train/tokens_used": 1055925728,
      "theoretical_loss": 3.637148205706141,
      "tokens_seen": 1035465728
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034660982948846544,
      "loss": 2.901,
      "theoretical_loss": 3.6371318813146516,
      "tokens_seen": 1035514880
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034659979939819456,
      "loss": 2.9041,
      "theoretical_loss": 3.637110117002095,
      "tokens_seen": 1035580416
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003465897693079238,
      "loss": 2.9562,
      "theoretical_loss": 3.6370883544524615,
      "tokens_seen": 1035645952
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.000346579739217653,
      "loss": 2.843,
      "theoretical_loss": 3.637066593665497,
      "tokens_seen": 1035711488
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034656970912738216,
      "loss": 2.7811,
      "theoretical_loss": 3.637044834640947,
      "tokens_seen": 1035777024
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034655967903711134,
      "loss": 2.5779,
      "theoretical_loss": 3.637023077378558,
      "tokens_seen": 1035842560
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003465496489468405,
      "loss": 2.7218,
      "theoretical_loss": 3.637001321878075,
      "tokens_seen": 1035908096
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003465396188565697,
      "loss": 2.7265,
      "theoretical_loss": 3.6369795681392443,
      "tokens_seen": 1035973632
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034652958876629894,
      "loss": 2.8509,
      "theoretical_loss": 3.6369578161618117,
      "tokens_seen": 1036039168
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034651955867602807,
      "loss": 2.6018,
      "theoretical_loss": 3.636936065945523,
      "tokens_seen": 1036104704
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003465095285857573,
      "loss": 2.5405,
      "theoretical_loss": 3.6369143174901244,
      "tokens_seen": 1036170240
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034649949849548643,
      "loss": 2.644,
      "theoretical_loss": 3.6368925707953625,
      "tokens_seen": 1036235776
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034648946840521567,
      "loss": 2.7706,
      "theoretical_loss": 3.6368708258609823,
      "tokens_seen": 1036301312
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034647943831494485,
      "loss": 2.7791,
      "theoretical_loss": 3.636849082686731,
      "tokens_seen": 1036366848
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034646940822467403,
      "loss": 2.528,
      "theoretical_loss": 3.6368273412723546,
      "tokens_seen": 1036432384
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003464593781344032,
      "loss": 2.7708,
      "theoretical_loss": 3.6368056016175987,
      "tokens_seen": 1036497920
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034644934804413244,
      "loss": 2.8057,
      "theoretical_loss": 3.6367838637222105,
      "tokens_seen": 1036563456
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034643931795386157,
      "loss": 2.8408,
      "theoretical_loss": 3.6367621275859356,
      "tokens_seen": 1036628992
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003464292878635908,
      "loss": 2.7432,
      "theoretical_loss": 3.6367403932085214,
      "tokens_seen": 1036694528
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034641925777331993,
      "loss": 2.7451,
      "theoretical_loss": 3.636718660589714,
      "tokens_seen": 1036760064
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034640922768304917,
      "loss": 2.7478,
      "theoretical_loss": 3.6366969297292595,
      "tokens_seen": 1036825600
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034639919759277835,
      "loss": 2.71,
      "theoretical_loss": 3.6366752006269056,
      "tokens_seen": 1036891136
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034638916750250753,
      "loss": 2.8828,
      "theoretical_loss": 3.636653473282397,
      "tokens_seen": 1036956672
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003463791374122367,
      "loss": 2.5511,
      "theoretical_loss": 3.636631747695483,
      "tokens_seen": 1037022208
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003463691073219659,
      "loss": 2.2055,
      "theoretical_loss": 3.6366100238659085,
      "tokens_seen": 1037087744
    },
    {
      "epoch": 3.04,
      "objective/train/docs_used": 1183367,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4552834033966064,
      "objective/train/theoretical_loss": 3.636604593183069,
      "objective/train/tokens_used": 1057564128,
      "theoretical_loss": 3.636604593183069,
      "tokens_seen": 1037104128
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003463590772316951,
      "loss": 2.5444,
      "theoretical_loss": 3.63658830179342,
      "tokens_seen": 1037153280
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003463490471414243,
      "loss": 2.6151,
      "theoretical_loss": 3.6365665814777666,
      "tokens_seen": 1037218816
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034633901705115344,
      "loss": 2.7626,
      "theoretical_loss": 3.6365448629186927,
      "tokens_seen": 1037284352
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034632898696088267,
      "loss": 2.8501,
      "theoretical_loss": 3.6365231461159473,
      "tokens_seen": 1037349888
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003463189568706118,
      "loss": 2.8395,
      "theoretical_loss": 3.6365014310692754,
      "tokens_seen": 1037415424
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034630892678034103,
      "loss": 2.8243,
      "theoretical_loss": 3.636479717778426,
      "tokens_seen": 1037480960
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003462988966900702,
      "loss": 2.8931,
      "theoretical_loss": 3.6364580062431453,
      "tokens_seen": 1037546496
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003462888665997994,
      "loss": 2.4737,
      "theoretical_loss": 3.6364362964631805,
      "tokens_seen": 1037612032
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003462788365095286,
      "loss": 2.6304,
      "theoretical_loss": 3.636414588438279,
      "tokens_seen": 1037677568
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003462688064192578,
      "loss": 2.7085,
      "theoretical_loss": 3.636392882168188,
      "tokens_seen": 1037743104
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034625877632898694,
      "loss": 2.7709,
      "theoretical_loss": 3.636371177652655,
      "tokens_seen": 1037808640
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003462487462387162,
      "loss": 2.7462,
      "theoretical_loss": 3.6363494748914267,
      "tokens_seen": 1037874176
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003462387161484453,
      "loss": 2.7757,
      "theoretical_loss": 3.6363277738842514,
      "tokens_seen": 1037939712
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034622868605817454,
      "loss": 2.7106,
      "theoretical_loss": 3.636306074630876,
      "tokens_seen": 1038005248
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003462186559679037,
      "loss": 2.5829,
      "theoretical_loss": 3.636284377131049,
      "tokens_seen": 1038070784
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003462086258776329,
      "loss": 2.6035,
      "theoretical_loss": 3.636262681384517,
      "tokens_seen": 1038136320
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003461985957873621,
      "loss": 2.6727,
      "theoretical_loss": 3.636240987391028,
      "tokens_seen": 1038201856
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034618856569709126,
      "loss": 2.8162,
      "theoretical_loss": 3.6362192951503296,
      "tokens_seen": 1038267392
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034617853560682044,
      "loss": 2.7681,
      "theoretical_loss": 3.6361976046621702,
      "tokens_seen": 1038332928
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003461685055165497,
      "loss": 2.7701,
      "theoretical_loss": 3.6361759159262963,
      "tokens_seen": 1038398464
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003461584754262788,
      "loss": 2.7035,
      "theoretical_loss": 3.636154228942457,
      "tokens_seen": 1038464000
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034614844533600804,
      "loss": 2.986,
      "theoretical_loss": 3.6361325437103993,
      "tokens_seen": 1038529536
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.00034613841524573717,
      "loss": 2.9617,
      "theoretical_loss": 3.636110860229872,
      "tokens_seen": 1038595072
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003461283851554664,
      "loss": 2.6108,
      "theoretical_loss": 3.636089178500623,
      "tokens_seen": 1038660608
    },
    {
      "epoch": 3.04,
      "learning_rate": 0.0003461183550651956,
      "loss": 2.639,
      "theoretical_loss": 3.6360674985223995,
      "tokens_seen": 1038726144
    },
    {
      "epoch": 3.04,
      "objective/train/docs_used": 1184791,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.915973663330078,
      "objective/train/theoretical_loss": 3.636062078801412,
      "objective/train/tokens_used": 1059202528,
      "theoretical_loss": 3.636062078801412,
      "tokens_seen": 1038742528
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034610832497492476,
      "loss": 2.8912,
      "theoretical_loss": 3.6360458202949504,
      "tokens_seen": 1038791680
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034609829488465395,
      "loss": 2.5813,
      "theoretical_loss": 3.636024143818024,
      "tokens_seen": 1038857216
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003460882647943832,
      "loss": 2.761,
      "theoretical_loss": 3.636002469091368,
      "tokens_seen": 1038922752
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034607823470411236,
      "loss": 2.6937,
      "theoretical_loss": 3.6359807961147306,
      "tokens_seen": 1038988288
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034606820461384154,
      "loss": 2.5175,
      "theoretical_loss": 3.6359591248878607,
      "tokens_seen": 1039053824
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003460581745235707,
      "loss": 2.9226,
      "theoretical_loss": 3.635937455410507,
      "tokens_seen": 1039119360
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003460481444332999,
      "loss": 2.6023,
      "theoretical_loss": 3.6359157876824164,
      "tokens_seen": 1039184896
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034603811434302914,
      "loss": 2.7549,
      "theoretical_loss": 3.6358941217033385,
      "tokens_seen": 1039250432
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034602808425275827,
      "loss": 2.6038,
      "theoretical_loss": 3.635872457473022,
      "tokens_seen": 1039315968
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003460180541624875,
      "loss": 2.8857,
      "theoretical_loss": 3.635850794991215,
      "tokens_seen": 1039381504
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034600802407221663,
      "loss": 2.6358,
      "theoretical_loss": 3.6358291342576665,
      "tokens_seen": 1039447040
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034599799398194587,
      "loss": 2.8824,
      "theoretical_loss": 3.635807475272125,
      "tokens_seen": 1039512576
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034598796389167505,
      "loss": 2.8954,
      "theoretical_loss": 3.635785818034339,
      "tokens_seen": 1039578112
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034597793380140423,
      "loss": 2.682,
      "theoretical_loss": 3.6357641625440578,
      "tokens_seen": 1039643648
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003459679037111334,
      "loss": 2.6801,
      "theoretical_loss": 3.6357425088010302,
      "tokens_seen": 1039709184
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034595787362086264,
      "loss": 2.5574,
      "theoretical_loss": 3.635720856805005,
      "tokens_seen": 1039774720
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034594784353059177,
      "loss": 2.7004,
      "theoretical_loss": 3.63569920655573,
      "tokens_seen": 1039840256
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.000345937813440321,
      "loss": 2.4317,
      "theoretical_loss": 3.635677558052956,
      "tokens_seen": 1039905792
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034592778335005013,
      "loss": 2.69,
      "theoretical_loss": 3.635655911296431,
      "tokens_seen": 1039971328
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034591775325977937,
      "loss": 2.6949,
      "theoretical_loss": 3.635634266285905,
      "tokens_seen": 1040036864
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034590772316950855,
      "loss": 2.9375,
      "theoretical_loss": 3.6356126230211263,
      "tokens_seen": 1040102400
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034589769307923773,
      "loss": 2.5751,
      "theoretical_loss": 3.635590981501845,
      "tokens_seen": 1040167936
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003458876629889669,
      "loss": 2.4966,
      "theoretical_loss": 3.635569341727809,
      "tokens_seen": 1040233472
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003458776328986961,
      "loss": 2.6717,
      "theoretical_loss": 3.6355477036987685,
      "tokens_seen": 1040299008
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003458676028084253,
      "loss": 2.7597,
      "theoretical_loss": 3.6355260674144727,
      "tokens_seen": 1040364544
    },
    {
      "epoch": 3.05,
      "objective/train/docs_used": 1185522,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9252572059631348,
      "objective/train/theoretical_loss": 3.6355206586159863,
      "objective/train/tokens_used": 1060840928,
      "theoretical_loss": 3.6355206586159863,
      "tokens_seen": 1040380928
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003458575727181545,
      "loss": 2.8402,
      "theoretical_loss": 3.6355044328746713,
      "tokens_seen": 1040430080
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034584754262788364,
      "loss": 2.8633,
      "theoretical_loss": 3.6354828000791137,
      "tokens_seen": 1040495616
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034583751253761287,
      "loss": 2.6076,
      "theoretical_loss": 3.6354611690275496,
      "tokens_seen": 1040561152
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.000345827482447342,
      "loss": 2.67,
      "theoretical_loss": 3.6354395397197274,
      "tokens_seen": 1040626688
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034581745235707123,
      "loss": 2.7122,
      "theoretical_loss": 3.6354179121553987,
      "tokens_seen": 1040692224
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003458074222668004,
      "loss": 2.7741,
      "theoretical_loss": 3.635396286334311,
      "tokens_seen": 1040757760
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003457973921765296,
      "loss": 2.7348,
      "theoretical_loss": 3.6353746622562157,
      "tokens_seen": 1040823296
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003457873620862588,
      "loss": 2.8063,
      "theoretical_loss": 3.6353530399208616,
      "tokens_seen": 1040888832
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.000345777331995988,
      "loss": 2.4041,
      "theoretical_loss": 3.6353314193279993,
      "tokens_seen": 1040954368
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034576730190571714,
      "loss": 2.9207,
      "theoretical_loss": 3.635309800477378,
      "tokens_seen": 1041019904
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003457572718154464,
      "loss": 2.7454,
      "theoretical_loss": 3.6352881833687483,
      "tokens_seen": 1041085440
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003457472417251755,
      "loss": 2.7788,
      "theoretical_loss": 3.6352665680018594,
      "tokens_seen": 1041150976
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034573721163490474,
      "loss": 2.909,
      "theoretical_loss": 3.6352449543764624,
      "tokens_seen": 1041216512
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003457271815446339,
      "loss": 2.8034,
      "theoretical_loss": 3.6352233424923064,
      "tokens_seen": 1041282048
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003457171514543631,
      "loss": 2.8243,
      "theoretical_loss": 3.6352017323491417,
      "tokens_seen": 1041347584
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003457071213640923,
      "loss": 2.6581,
      "theoretical_loss": 3.635180123946719,
      "tokens_seen": 1041413120
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034569709127382146,
      "loss": 2.7839,
      "theoretical_loss": 3.6351585172847884,
      "tokens_seen": 1041478656
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034568706118355064,
      "loss": 2.6316,
      "theoretical_loss": 3.6351369123631,
      "tokens_seen": 1041544192
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003456770310932799,
      "loss": 2.5668,
      "theoretical_loss": 3.635115309181404,
      "tokens_seen": 1041609728
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.000345667001003009,
      "loss": 2.7055,
      "theoretical_loss": 3.6350937077394514,
      "tokens_seen": 1041675264
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034565697091273824,
      "loss": 2.7145,
      "theoretical_loss": 3.635072108036992,
      "tokens_seen": 1041740800
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034564694082246737,
      "loss": 2.8442,
      "theoretical_loss": 3.6350505100737767,
      "tokens_seen": 1041806336
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003456369107321966,
      "loss": 2.6584,
      "theoretical_loss": 3.6350289138495557,
      "tokens_seen": 1041871872
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003456268806419258,
      "loss": 2.7529,
      "theoretical_loss": 3.63500731936408,
      "tokens_seen": 1041937408
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034561685055165496,
      "loss": 2.7175,
      "theoretical_loss": 3.6349857266171,
      "tokens_seen": 1042002944
    },
    {
      "epoch": 3.05,
      "objective/train/docs_used": 1186074,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.702031135559082,
      "objective/train/theoretical_loss": 3.6349803287019657,
      "objective/train/tokens_used": 1062479328,
      "theoretical_loss": 3.6349803287019657,
      "tokens_seen": 1042019328
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034560682046138415,
      "loss": 2.7838,
      "theoretical_loss": 3.634964135608367,
      "tokens_seen": 1042068480
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003455967903711134,
      "loss": 2.7338,
      "theoretical_loss": 3.634942546337631,
      "tokens_seen": 1042134016
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003455867602808425,
      "loss": 2.6965,
      "theoretical_loss": 3.634920958804643,
      "tokens_seen": 1042199552
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034557673019057174,
      "loss": 2.7916,
      "theoretical_loss": 3.634899373009154,
      "tokens_seen": 1042265088
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034556670010030087,
      "loss": 2.4773,
      "theoretical_loss": 3.6348777889509147,
      "tokens_seen": 1042330624
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003455566700100301,
      "loss": 2.3448,
      "theoretical_loss": 3.634856206629677,
      "tokens_seen": 1042396160
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003455466399197593,
      "loss": 2.6472,
      "theoretical_loss": 3.6348346260451905,
      "tokens_seen": 1042461696
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034553660982948847,
      "loss": 2.8073,
      "theoretical_loss": 3.634813047197207,
      "tokens_seen": 1042527232
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034552657973921765,
      "loss": 2.7546,
      "theoretical_loss": 3.6347914700854775,
      "tokens_seen": 1042592768
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034551654964894683,
      "loss": 2.8483,
      "theoretical_loss": 3.6347698947097538,
      "tokens_seen": 1042658304
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.000345506519558676,
      "loss": 2.6587,
      "theoretical_loss": 3.634748321069786,
      "tokens_seen": 1042723840
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034549648946840525,
      "loss": 2.675,
      "theoretical_loss": 3.634726749165326,
      "tokens_seen": 1042789376
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003454864593781344,
      "loss": 2.642,
      "theoretical_loss": 3.634705178996125,
      "tokens_seen": 1042854912
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003454764292878636,
      "loss": 2.4778,
      "theoretical_loss": 3.6346836105619347,
      "tokens_seen": 1042920448
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034546639919759274,
      "loss": 2.4542,
      "theoretical_loss": 3.634662043862506,
      "tokens_seen": 1042985984
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034545636910732197,
      "loss": 2.4854,
      "theoretical_loss": 3.634640478897591,
      "tokens_seen": 1043051520
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034544633901705115,
      "loss": 2.834,
      "theoretical_loss": 3.6346189156669406,
      "tokens_seen": 1043117056
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034543630892678033,
      "loss": 3.0203,
      "theoretical_loss": 3.6345973541703067,
      "tokens_seen": 1043182592
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003454262788365095,
      "loss": 2.9097,
      "theoretical_loss": 3.634575794407441,
      "tokens_seen": 1043248128
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034541624874623875,
      "loss": 2.8772,
      "theoretical_loss": 3.6345542363780954,
      "tokens_seen": 1043313664
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003454062186559679,
      "loss": 2.6594,
      "theoretical_loss": 3.634532680082021,
      "tokens_seen": 1043379200
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003453961885656971,
      "loss": 2.8194,
      "theoretical_loss": 3.63451112551897,
      "tokens_seen": 1043444736
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034538615847542624,
      "loss": 2.8311,
      "theoretical_loss": 3.634489572688694,
      "tokens_seen": 1043510272
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003453761283851555,
      "loss": 2.6407,
      "theoretical_loss": 3.634468021590945,
      "tokens_seen": 1043575808
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034536609829488466,
      "loss": 2.7576,
      "theoretical_loss": 3.6344464722254757,
      "tokens_seen": 1043641344
    },
    {
      "epoch": 3.05,
      "objective/train/docs_used": 1186076,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4055583477020264,
      "objective/train/theoretical_loss": 3.634441085154748,
      "objective/train/tokens_used": 1064117728,
      "theoretical_loss": 3.634441085154748,
      "tokens_seen": 1043657728
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034535606820461384,
      "loss": 2.7181,
      "theoretical_loss": 3.634424924592037,
      "tokens_seen": 1043706880
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.000345346038114343,
      "loss": 2.7111,
      "theoretical_loss": 3.634403378690381,
      "tokens_seen": 1043772416
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003453360080240722,
      "loss": 2.8271,
      "theoretical_loss": 3.6343818345202603,
      "tokens_seen": 1043837952
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034532597793380143,
      "loss": 2.7099,
      "theoretical_loss": 3.634360292081427,
      "tokens_seen": 1043903488
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003453159478435306,
      "loss": 2.8388,
      "theoretical_loss": 3.634338751373633,
      "tokens_seen": 1043969024
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003453059177532598,
      "loss": 2.695,
      "theoretical_loss": 3.634317212396631,
      "tokens_seen": 1044034560
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.000345295887662989,
      "loss": 2.5767,
      "theoretical_loss": 3.634295675150173,
      "tokens_seen": 1044100096
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003452858575727182,
      "loss": 2.8101,
      "theoretical_loss": 3.634274139634011,
      "tokens_seen": 1044165632
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034527582748244734,
      "loss": 3.153,
      "theoretical_loss": 3.634252605847898,
      "tokens_seen": 1044231168
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003452657973921766,
      "loss": 2.9676,
      "theoretical_loss": 3.634231073791586,
      "tokens_seen": 1044296704
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003452557673019057,
      "loss": 2.6099,
      "theoretical_loss": 3.634209543464828,
      "tokens_seen": 1044362240
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034524573721163494,
      "loss": 2.8402,
      "theoretical_loss": 3.6341880148673766,
      "tokens_seen": 1044427776
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003452357071213641,
      "loss": 2.9848,
      "theoretical_loss": 3.6341664879989835,
      "tokens_seen": 1044493312
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003452256770310933,
      "loss": 2.7283,
      "theoretical_loss": 3.634144962859402,
      "tokens_seen": 1044558848
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003452156469408225,
      "loss": 2.6106,
      "theoretical_loss": 3.634123439448385,
      "tokens_seen": 1044624384
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034520561685055166,
      "loss": 2.696,
      "theoretical_loss": 3.634101917765685,
      "tokens_seen": 1044689920
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034519558676028084,
      "loss": 2.7361,
      "theoretical_loss": 3.6340803978110543,
      "tokens_seen": 1044755456
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003451855566700101,
      "loss": 2.813,
      "theoretical_loss": 3.634058879584247,
      "tokens_seen": 1044820992
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003451755265797392,
      "loss": 2.8119,
      "theoretical_loss": 3.6340373630850147,
      "tokens_seen": 1044886528
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034516549648946844,
      "loss": 2.7531,
      "theoretical_loss": 3.6340158483131115,
      "tokens_seen": 1044952064
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034515546639919757,
      "loss": 2.8369,
      "theoretical_loss": 3.6339943352682895,
      "tokens_seen": 1045017600
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003451454363089268,
      "loss": 2.7817,
      "theoretical_loss": 3.633972823950302,
      "tokens_seen": 1045083136
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.000345135406218656,
      "loss": 2.572,
      "theoretical_loss": 3.633951314358903,
      "tokens_seen": 1045148672
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034512537612838517,
      "loss": 3.0036,
      "theoretical_loss": 3.6339298064938443,
      "tokens_seen": 1045214208
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034511534603811435,
      "loss": 2.9289,
      "theoretical_loss": 3.6339083003548796,
      "tokens_seen": 1045279744
    },
    {
      "epoch": 3.05,
      "objective/train/docs_used": 1186840,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6896257400512695,
      "objective/train/theoretical_loss": 3.6339029240898117,
      "objective/train/tokens_used": 1065756128,
      "theoretical_loss": 3.6339029240898117,
      "tokens_seen": 1045296128
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003451053159478436,
      "loss": 2.7279,
      "theoretical_loss": 3.6338867959417622,
      "tokens_seen": 1045345280
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003450952858575727,
      "loss": 2.8756,
      "theoretical_loss": 3.633865293254246,
      "tokens_seen": 1045410816
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034508525576730194,
      "loss": 2.6824,
      "theoretical_loss": 3.633843792292083,
      "tokens_seen": 1045476352
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034507522567703107,
      "loss": 2.7099,
      "theoretical_loss": 3.6338222930550286,
      "tokens_seen": 1045541888
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003450651955867603,
      "loss": 2.6207,
      "theoretical_loss": 3.6338007955428346,
      "tokens_seen": 1045607424
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003450551654964895,
      "loss": 2.7052,
      "theoretical_loss": 3.633779299755255,
      "tokens_seen": 1045672960
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034504513540621867,
      "loss": 2.7676,
      "theoretical_loss": 3.633757805692044,
      "tokens_seen": 1045738496
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034503510531594785,
      "loss": 2.7697,
      "theoretical_loss": 3.633736313352954,
      "tokens_seen": 1045804032
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034502507522567703,
      "loss": 2.6363,
      "theoretical_loss": 3.6337148227377396,
      "tokens_seen": 1045869568
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003450150451354062,
      "loss": 2.8698,
      "theoretical_loss": 3.633693333846154,
      "tokens_seen": 1045935104
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034500501504513545,
      "loss": 2.7141,
      "theoretical_loss": 3.6336718466779514,
      "tokens_seen": 1046000640
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003449949849548646,
      "loss": 2.8788,
      "theoretical_loss": 3.6336503612328857,
      "tokens_seen": 1046066176
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003449849548645938,
      "loss": 2.9623,
      "theoretical_loss": 3.63362887751071,
      "tokens_seen": 1046131712
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034497492477432294,
      "loss": 2.7589,
      "theoretical_loss": 3.633607395511179,
      "tokens_seen": 1046197248
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034496489468405217,
      "loss": 2.716,
      "theoretical_loss": 3.6335859152340464,
      "tokens_seen": 1046262784
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034495486459378135,
      "loss": 2.9333,
      "theoretical_loss": 3.633564436679066,
      "tokens_seen": 1046328320
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034494483450351053,
      "loss": 2.7096,
      "theoretical_loss": 3.6335429598459927,
      "tokens_seen": 1046393856
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003449348044132397,
      "loss": 2.5753,
      "theoretical_loss": 3.6335214847345796,
      "tokens_seen": 1046459392
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034492477432296895,
      "loss": 2.7974,
      "theoretical_loss": 3.633500011344582,
      "tokens_seen": 1046524928
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003449147442326981,
      "loss": 2.733,
      "theoretical_loss": 3.6334785396757527,
      "tokens_seen": 1046590464
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003449047141424273,
      "loss": 2.6909,
      "theoretical_loss": 3.633457069727847,
      "tokens_seen": 1046656000
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034489468405215644,
      "loss": 2.966,
      "theoretical_loss": 3.6334356015006186,
      "tokens_seen": 1046721536
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003448846539618857,
      "loss": 2.6867,
      "theoretical_loss": 3.633414134993823,
      "tokens_seen": 1046787072
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034487462387161486,
      "loss": 2.9336,
      "theoretical_loss": 3.6333926702072135,
      "tokens_seen": 1046852608
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034486459378134404,
      "loss": 2.6975,
      "theoretical_loss": 3.6333712071405446,
      "tokens_seen": 1046918144
    },
    {
      "epoch": 3.05,
      "objective/train/docs_used": 1188135,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7929697036743164,
      "objective/train/theoretical_loss": 3.63336584164259,
      "objective/train/tokens_used": 1067394528,
      "theoretical_loss": 3.63336584164259,
      "tokens_seen": 1046934528
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003448545636910732,
      "loss": 2.5809,
      "theoretical_loss": 3.6333497457935717,
      "tokens_seen": 1046983680
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003448445336008024,
      "loss": 2.8726,
      "theoretical_loss": 3.6333282861660487,
      "tokens_seen": 1047049216
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003448345035105316,
      "loss": 2.6964,
      "theoretical_loss": 3.63330682825773,
      "tokens_seen": 1047114752
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003448244734202608,
      "loss": 2.8283,
      "theoretical_loss": 3.6332853720683715,
      "tokens_seen": 1047180288
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034481444332998994,
      "loss": 3.0839,
      "theoretical_loss": 3.633263917597726,
      "tokens_seen": 1047245824
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003448044132397192,
      "loss": 2.8138,
      "theoretical_loss": 3.6332424648455506,
      "tokens_seen": 1047311360
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034479438314944836,
      "loss": 2.6576,
      "theoretical_loss": 3.633221013811598,
      "tokens_seen": 1047376896
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034478435305917754,
      "loss": 2.7789,
      "theoretical_loss": 3.6331995644956243,
      "tokens_seen": 1047442432
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003447743229689067,
      "loss": 2.8961,
      "theoretical_loss": 3.6331781168973842,
      "tokens_seen": 1047507968
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003447642928786359,
      "loss": 2.6799,
      "theoretical_loss": 3.6331566710166325,
      "tokens_seen": 1047573504
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003447542627883651,
      "loss": 2.9108,
      "theoretical_loss": 3.6331352268531245,
      "tokens_seen": 1047639040
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003447442326980943,
      "loss": 2.7632,
      "theoretical_loss": 3.633113784406615,
      "tokens_seen": 1047704576
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034473420260782345,
      "loss": 2.8166,
      "theoretical_loss": 3.6330923436768594,
      "tokens_seen": 1047770112
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003447241725175527,
      "loss": 2.8229,
      "theoretical_loss": 3.6330709046636125,
      "tokens_seen": 1047835648
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003447141424272818,
      "loss": 3.0599,
      "theoretical_loss": 3.63304946736663,
      "tokens_seen": 1047901184
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034470411233701104,
      "loss": 2.8003,
      "theoretical_loss": 3.633028031785667,
      "tokens_seen": 1047966720
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003446940822467402,
      "loss": 2.8111,
      "theoretical_loss": 3.633006597920479,
      "tokens_seen": 1048032256
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003446840521564694,
      "loss": 2.7593,
      "theoretical_loss": 3.6329851657708208,
      "tokens_seen": 1048097792
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003446740220661986,
      "loss": 2.6875,
      "theoretical_loss": 3.632963735336448,
      "tokens_seen": 1048163328
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034466399197592777,
      "loss": 2.8356,
      "theoretical_loss": 3.632942306617117,
      "tokens_seen": 1048228864
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034465396188565695,
      "loss": 2.591,
      "theoretical_loss": 3.6329208796125823,
      "tokens_seen": 1048294400
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003446439317953862,
      "loss": 2.6914,
      "theoretical_loss": 3.6328994543226,
      "tokens_seen": 1048359936
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003446339017051153,
      "loss": 2.757,
      "theoretical_loss": 3.632878030746925,
      "tokens_seen": 1048425472
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034462387161484455,
      "loss": 2.9459,
      "theoretical_loss": 3.632856608885314,
      "tokens_seen": 1048491008
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034461384152457373,
      "loss": 2.9613,
      "theoretical_loss": 3.6328351887375225,
      "tokens_seen": 1048556544
    },
    {
      "epoch": 3.05,
      "objective/train/docs_used": 1188820,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7827885150909424,
      "objective/train/theoretical_loss": 3.63282983396833,
      "objective/train/tokens_used": 1069032928,
      "theoretical_loss": 3.63282983396833,
      "tokens_seen": 1048572928
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003446038114343029,
      "loss": 2.8882,
      "theoretical_loss": 3.6328137703033057,
      "tokens_seen": 1048622080
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034459378134403214,
      "loss": 2.6796,
      "theoretical_loss": 3.63279235358242,
      "tokens_seen": 1048687616
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034458375125376127,
      "loss": 2.7243,
      "theoretical_loss": 3.632770938574621,
      "tokens_seen": 1048753152
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003445737211634905,
      "loss": 2.835,
      "theoretical_loss": 3.6327495252796647,
      "tokens_seen": 1048818688
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003445636910732197,
      "loss": 2.756,
      "theoretical_loss": 3.632728113697307,
      "tokens_seen": 1048884224
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034455366098294887,
      "loss": 2.7928,
      "theoretical_loss": 3.6327067038273047,
      "tokens_seen": 1048949760
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034454363089267805,
      "loss": 2.5859,
      "theoretical_loss": 3.632685295669413,
      "tokens_seen": 1049015296
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034453360080240723,
      "loss": 2.7753,
      "theoretical_loss": 3.6326638892233873,
      "tokens_seen": 1049080832
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003445235707121364,
      "loss": 2.7935,
      "theoretical_loss": 3.6326424844889855,
      "tokens_seen": 1049146368
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034451354062186565,
      "loss": 2.6532,
      "theoretical_loss": 3.6326210814659636,
      "tokens_seen": 1049211904
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003445035105315948,
      "loss": 2.7596,
      "theoretical_loss": 3.632599680154077,
      "tokens_seen": 1049277440
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.000344493480441324,
      "loss": 2.7922,
      "theoretical_loss": 3.632578280553082,
      "tokens_seen": 1049342976
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034448345035105314,
      "loss": 2.6624,
      "theoretical_loss": 3.632556882662736,
      "tokens_seen": 1049408512
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034447342026078237,
      "loss": 2.6397,
      "theoretical_loss": 3.6325354864827943,
      "tokens_seen": 1049474048
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034446339017051155,
      "loss": 2.7459,
      "theoretical_loss": 3.6325140920130146,
      "tokens_seen": 1049539584
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034445336008024073,
      "loss": 2.6949,
      "theoretical_loss": 3.6324926992531523,
      "tokens_seen": 1049605120
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003444433299899699,
      "loss": 2.4783,
      "theoretical_loss": 3.6324713082029643,
      "tokens_seen": 1049670656
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034443329989969915,
      "loss": 2.788,
      "theoretical_loss": 3.632449918862208,
      "tokens_seen": 1049736192
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003444232698094283,
      "loss": 2.8112,
      "theoretical_loss": 3.632428531230639,
      "tokens_seen": 1049801728
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003444132397191575,
      "loss": 2.7413,
      "theoretical_loss": 3.6324071453080147,
      "tokens_seen": 1049867264
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034440320962888664,
      "loss": 2.6558,
      "theoretical_loss": 3.6323857610940915,
      "tokens_seen": 1049932800
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003443931795386159,
      "loss": 2.7811,
      "theoretical_loss": 3.6323643785886266,
      "tokens_seen": 1049998336
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034438314944834506,
      "loss": 2.9847,
      "theoretical_loss": 3.6323429977913766,
      "tokens_seen": 1050063872
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034437311935807424,
      "loss": 2.6683,
      "theoretical_loss": 3.6323216187020986,
      "tokens_seen": 1050129408
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003443630892678034,
      "loss": 2.7443,
      "theoretical_loss": 3.632300241320549,
      "tokens_seen": 1050194944
    },
    {
      "epoch": 3.05,
      "objective/train/docs_used": 1190162,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4736685752868652,
      "objective/train/theoretical_loss": 3.6322948972419664,
      "objective/train/tokens_used": 1070671328,
      "theoretical_loss": 3.6322948972419664,
      "tokens_seen": 1050211328
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003443530591775326,
      "loss": 2.7721,
      "theoretical_loss": 3.6322788656464864,
      "tokens_seen": 1050260480
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003443430290872618,
      "loss": 2.5936,
      "theoretical_loss": 3.632257491679666,
      "tokens_seen": 1050326016
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.000344332998996991,
      "loss": 2.5859,
      "theoretical_loss": 3.6322361194198463,
      "tokens_seen": 1050391552
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034432296890672014,
      "loss": 2.5561,
      "theoretical_loss": 3.632214748866784,
      "tokens_seen": 1050457088
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003443129388164494,
      "loss": 2.7315,
      "theoretical_loss": 3.6321933800202357,
      "tokens_seen": 1050522624
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034430290872617856,
      "loss": 2.6789,
      "theoretical_loss": 3.63217201287996,
      "tokens_seen": 1050588160
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034429287863590774,
      "loss": 2.7474,
      "theoretical_loss": 3.632150647445713,
      "tokens_seen": 1050653696
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003442828485456369,
      "loss": 2.8795,
      "theoretical_loss": 3.6321292837172523,
      "tokens_seen": 1050719232
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003442728184553661,
      "loss": 2.7036,
      "theoretical_loss": 3.6321079216943364,
      "tokens_seen": 1050784768
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003442627883650953,
      "loss": 2.6738,
      "theoretical_loss": 3.6320865613767213,
      "tokens_seen": 1050850304
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003442527582748245,
      "loss": 2.6909,
      "theoretical_loss": 3.6320652027641653,
      "tokens_seen": 1050915840
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034424272818455365,
      "loss": 2.6339,
      "theoretical_loss": 3.6320438458564266,
      "tokens_seen": 1050981376
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003442326980942829,
      "loss": 2.8449,
      "theoretical_loss": 3.632022490653261,
      "tokens_seen": 1051046912
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.000344222668004012,
      "loss": 2.759,
      "theoretical_loss": 3.6320011371544285,
      "tokens_seen": 1051112448
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034421263791374124,
      "loss": 2.7718,
      "theoretical_loss": 3.631979785359685,
      "tokens_seen": 1051177984
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003442026078234704,
      "loss": 2.5683,
      "theoretical_loss": 3.6319584352687895,
      "tokens_seen": 1051243520
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003441925777331996,
      "loss": 2.87,
      "theoretical_loss": 3.6319370868814986,
      "tokens_seen": 1051309056
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003441825476429288,
      "loss": 2.9015,
      "theoretical_loss": 3.631915740197571,
      "tokens_seen": 1051374592
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034417251755265797,
      "loss": 2.717,
      "theoretical_loss": 3.6318943952167646,
      "tokens_seen": 1051440128
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034416248746238715,
      "loss": 2.6497,
      "theoretical_loss": 3.631873051938837,
      "tokens_seen": 1051505664
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003441524573721164,
      "loss": 2.7191,
      "theoretical_loss": 3.631851710363547,
      "tokens_seen": 1051571200
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003441424272818455,
      "loss": 2.7562,
      "theoretical_loss": 3.6318303704906514,
      "tokens_seen": 1051636736
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034413239719157475,
      "loss": 2.6796,
      "theoretical_loss": 3.6318090323199095,
      "tokens_seen": 1051702272
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034412236710130393,
      "loss": 2.7178,
      "theoretical_loss": 3.631787695851079,
      "tokens_seen": 1051767808
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003441123370110331,
      "loss": 2.9961,
      "theoretical_loss": 3.6317663610839177,
      "tokens_seen": 1051833344
    },
    {
      "epoch": 3.05,
      "objective/train/docs_used": 1190623,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.515566825866699,
      "objective/train/theoretical_loss": 3.6317610276579853,
      "objective/train/tokens_used": 1072309728,
      "theoretical_loss": 3.6317610276579853,
      "tokens_seen": 1051849728
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003441023069207623,
      "loss": 2.5649,
      "theoretical_loss": 3.6317450280181847,
      "tokens_seen": 1051898880
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034409227683049147,
      "loss": 2.9343,
      "theoretical_loss": 3.631723696653638,
      "tokens_seen": 1051964416
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034408224674022065,
      "loss": 2.8018,
      "theoretical_loss": 3.631702366990036,
      "tokens_seen": 1052029952
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003440722166499499,
      "loss": 2.5757,
      "theoretical_loss": 3.6316810390271366,
      "tokens_seen": 1052095488
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.000344062186559679,
      "loss": 2.7342,
      "theoretical_loss": 3.631659712764699,
      "tokens_seen": 1052161024
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034405215646940825,
      "loss": 2.7447,
      "theoretical_loss": 3.6316383882024814,
      "tokens_seen": 1052226560
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003440421263791374,
      "loss": 2.5221,
      "theoretical_loss": 3.631617065340243,
      "tokens_seen": 1052292096
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003440320962888666,
      "loss": 2.8617,
      "theoretical_loss": 3.6315957441777407,
      "tokens_seen": 1052357632
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003440220661985958,
      "loss": 2.8462,
      "theoretical_loss": 3.631574424714735,
      "tokens_seen": 1052423168
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.000344012036108325,
      "loss": 2.7823,
      "theoretical_loss": 3.6315531069509834,
      "tokens_seen": 1052488704
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034400200601805416,
      "loss": 2.7497,
      "theoretical_loss": 3.6315317908862457,
      "tokens_seen": 1052554240
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034399197592778334,
      "loss": 2.7485,
      "theoretical_loss": 3.6315104765202797,
      "tokens_seen": 1052619776
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003439819458375125,
      "loss": 2.6464,
      "theoretical_loss": 3.6314891638528453,
      "tokens_seen": 1052685312
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034397191574724175,
      "loss": 2.9019,
      "theoretical_loss": 3.6314678528837003,
      "tokens_seen": 1052750848
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003439618856569709,
      "loss": 2.8454,
      "theoretical_loss": 3.6314465436126047,
      "tokens_seen": 1052816384
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003439518555667001,
      "loss": 2.9119,
      "theoretical_loss": 3.6314252360393167,
      "tokens_seen": 1052881920
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003439418254764293,
      "loss": 2.8862,
      "theoretical_loss": 3.6314039301635956,
      "tokens_seen": 1052947456
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003439317953861585,
      "loss": 2.6123,
      "theoretical_loss": 3.631382625985201,
      "tokens_seen": 1053012992
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034392176529588766,
      "loss": 2.746,
      "theoretical_loss": 3.6313613235038913,
      "tokens_seen": 1053078528
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034391173520561684,
      "loss": 2.8266,
      "theoretical_loss": 3.6313400227194266,
      "tokens_seen": 1053144064
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.000343901705115346,
      "loss": 2.7095,
      "theoretical_loss": 3.631318723631565,
      "tokens_seen": 1053209600
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034389167502507526,
      "loss": 2.7095,
      "theoretical_loss": 3.631297426240067,
      "tokens_seen": 1053275136
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003438816449348044,
      "loss": 2.7669,
      "theoretical_loss": 3.6312761305446912,
      "tokens_seen": 1053340672
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003438716148445336,
      "loss": 2.7469,
      "theoretical_loss": 3.631254836545197,
      "tokens_seen": 1053406208
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034386158475426275,
      "loss": 2.6697,
      "theoretical_loss": 3.6312335442413444,
      "tokens_seen": 1053471744
    },
    {
      "epoch": 3.05,
      "objective/train/docs_used": 1192122,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.786816120147705,
      "objective/train/theoretical_loss": 3.6312282214302964,
      "objective/train/tokens_used": 1073948128,
      "theoretical_loss": 3.6312282214302964,
      "tokens_seen": 1053488128
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.000343851554663992,
      "loss": 2.7388,
      "theoretical_loss": 3.631212253632892,
      "tokens_seen": 1053537280
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003438415245737212,
      "loss": 2.8104,
      "theoretical_loss": 3.6311909647196003,
      "tokens_seen": 1053602816
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034383149448345034,
      "loss": 2.8182,
      "theoretical_loss": 3.6311696775012283,
      "tokens_seen": 1053668352
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003438214643931796,
      "loss": 2.8246,
      "theoretical_loss": 3.6311483919775362,
      "tokens_seen": 1053733888
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034381143430290876,
      "loss": 2.7451,
      "theoretical_loss": 3.6311271081482834,
      "tokens_seen": 1053799424
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034380140421263794,
      "loss": 2.7069,
      "theoretical_loss": 3.631105826013229,
      "tokens_seen": 1053864960
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003437913741223671,
      "loss": 2.5447,
      "theoretical_loss": 3.631084545572134,
      "tokens_seen": 1053930496
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003437813440320963,
      "loss": 2.8172,
      "theoretical_loss": 3.6310632668247576,
      "tokens_seen": 1053996032
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003437713139418255,
      "loss": 2.8105,
      "theoretical_loss": 3.6310419897708597,
      "tokens_seen": 1054061568
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003437612838515547,
      "loss": 2.7633,
      "theoretical_loss": 3.6310207144102,
      "tokens_seen": 1054127104
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034375125376128385,
      "loss": 2.7244,
      "theoretical_loss": 3.630999440742539,
      "tokens_seen": 1054192640
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003437412236710131,
      "loss": 2.6294,
      "theoretical_loss": 3.6309781687676366,
      "tokens_seen": 1054258176
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003437311935807422,
      "loss": 2.688,
      "theoretical_loss": 3.6309568984852527,
      "tokens_seen": 1054323712
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034372116349047144,
      "loss": 2.7278,
      "theoretical_loss": 3.6309356298951476,
      "tokens_seen": 1054389248
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003437111334002006,
      "loss": 2.8325,
      "theoretical_loss": 3.630914362997082,
      "tokens_seen": 1054454784
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003437011033099298,
      "loss": 2.5317,
      "theoretical_loss": 3.630893097790815,
      "tokens_seen": 1054520320
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.000343691073219659,
      "loss": 2.6591,
      "theoretical_loss": 3.6308718342761073,
      "tokens_seen": 1054585856
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034368104312938817,
      "loss": 2.8177,
      "theoretical_loss": 3.6308505724527196,
      "tokens_seen": 1054651392
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034367101303911735,
      "loss": 2.5762,
      "theoretical_loss": 3.6308293123204125,
      "tokens_seen": 1054716928
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003436609829488466,
      "loss": 2.7574,
      "theoretical_loss": 3.630808053878946,
      "tokens_seen": 1054782464
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003436509528585757,
      "loss": 2.8421,
      "theoretical_loss": 3.6307867971280796,
      "tokens_seen": 1054848000
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034364092276830495,
      "loss": 2.7248,
      "theoretical_loss": 3.630765542067576,
      "tokens_seen": 1054913536
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034363089267803413,
      "loss": 2.6222,
      "theoretical_loss": 3.630744288697194,
      "tokens_seen": 1054979072
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003436208625877633,
      "loss": 2.7392,
      "theoretical_loss": 3.630723037016695,
      "tokens_seen": 1055044608
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003436108324974925,
      "loss": 2.685,
      "theoretical_loss": 3.63070178702584,
      "tokens_seen": 1055110144
    },
    {
      "epoch": 3.05,
      "objective/train/docs_used": 1192817,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.987161874771118,
      "objective/train/theoretical_loss": 3.6306964747921047,
      "objective/train/tokens_used": 1075586528,
      "theoretical_loss": 3.6306964747921047,
      "tokens_seen": 1055126528
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034360080240722167,
      "loss": 2.8193,
      "theoretical_loss": 3.6306805387243886,
      "tokens_seen": 1055175680
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034359077231695085,
      "loss": 2.6656,
      "theoretical_loss": 3.6306592921121026,
      "tokens_seen": 1055241216
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003435807422266801,
      "loss": 2.8565,
      "theoretical_loss": 3.630638047188742,
      "tokens_seen": 1055306752
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003435707121364092,
      "loss": 2.8089,
      "theoretical_loss": 3.630616803954069,
      "tokens_seen": 1055372288
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034356068204613845,
      "loss": 2.7897,
      "theoretical_loss": 3.6305955624078434,
      "tokens_seen": 1055437824
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003435506519558676,
      "loss": 2.8819,
      "theoretical_loss": 3.630574322549826,
      "tokens_seen": 1055503360
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003435406218655968,
      "loss": 2.6583,
      "theoretical_loss": 3.6305530843797786,
      "tokens_seen": 1055568896
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.000343530591775326,
      "loss": 2.8231,
      "theoretical_loss": 3.630531847897462,
      "tokens_seen": 1055634432
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003435205616850552,
      "loss": 2.5255,
      "theoretical_loss": 3.6305106131026363,
      "tokens_seen": 1055699968
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034351053159478436,
      "loss": 2.8292,
      "theoretical_loss": 3.6304893799950646,
      "tokens_seen": 1055765504
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034350050150451354,
      "loss": 2.8406,
      "theoretical_loss": 3.630468148574507,
      "tokens_seen": 1055831040
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003434904714142427,
      "loss": 2.6635,
      "theoretical_loss": 3.6304469188407245,
      "tokens_seen": 1055896576
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034348044132397195,
      "loss": 2.5332,
      "theoretical_loss": 3.630425690793479,
      "tokens_seen": 1055962112
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003434704112337011,
      "loss": 2.7283,
      "theoretical_loss": 3.6304044644325315,
      "tokens_seen": 1056027648
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003434603811434303,
      "loss": 2.8458,
      "theoretical_loss": 3.6303832397576437,
      "tokens_seen": 1056093184
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003434503510531595,
      "loss": 2.8332,
      "theoretical_loss": 3.6303620167685766,
      "tokens_seen": 1056158720
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003434403209628887,
      "loss": 2.8633,
      "theoretical_loss": 3.6303407954650924,
      "tokens_seen": 1056224256
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034343029087261786,
      "loss": 2.8437,
      "theoretical_loss": 3.630319575846952,
      "tokens_seen": 1056289792
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034342026078234704,
      "loss": 2.837,
      "theoretical_loss": 3.6302983579139174,
      "tokens_seen": 1056355328
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003434102306920762,
      "loss": 2.7527,
      "theoretical_loss": 3.6302771416657498,
      "tokens_seen": 1056420864
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034340020060180546,
      "loss": 2.8107,
      "theoretical_loss": 3.6302559271022115,
      "tokens_seen": 1056486400
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003433901705115346,
      "loss": 2.751,
      "theoretical_loss": 3.6302347142230644,
      "tokens_seen": 1056551936
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003433801404212638,
      "loss": 2.834,
      "theoretical_loss": 3.6302135030280693,
      "tokens_seen": 1056617472
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034337011033099295,
      "loss": 2.6986,
      "theoretical_loss": 3.6301922935169886,
      "tokens_seen": 1056683008
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003433600802407222,
      "loss": 2.7787,
      "theoretical_loss": 3.630171085689584,
      "tokens_seen": 1056748544
    },
    {
      "epoch": 3.05,
      "objective/train/docs_used": 1193467,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6657423973083496,
      "objective/train/theoretical_loss": 3.6301657839957793,
      "objective/train/tokens_used": 1077224928,
      "theoretical_loss": 3.6301657839957793,
      "tokens_seen": 1056764928
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034335005015045136,
      "loss": 2.6549,
      "theoretical_loss": 3.630149879545618,
      "tokens_seen": 1056814080
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034334002006018054,
      "loss": 2.6641,
      "theoretical_loss": 3.630128675084852,
      "tokens_seen": 1056879616
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003433299899699097,
      "loss": 2.6756,
      "theoretical_loss": 3.6301074723070483,
      "tokens_seen": 1056945152
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003433199598796389,
      "loss": 2.8726,
      "theoretical_loss": 3.6300862712119693,
      "tokens_seen": 1057010688
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003433099297893681,
      "loss": 2.8562,
      "theoretical_loss": 3.6300650717993763,
      "tokens_seen": 1057076224
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003432998996990973,
      "loss": 2.4702,
      "theoretical_loss": 3.630043874069032,
      "tokens_seen": 1057141760
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034328986960882645,
      "loss": 2.7899,
      "theoretical_loss": 3.630022678020699,
      "tokens_seen": 1057207296
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003432798395185557,
      "loss": 2.6209,
      "theoretical_loss": 3.630001483654139,
      "tokens_seen": 1057272832
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034326980942828487,
      "loss": 2.8341,
      "theoretical_loss": 3.6299802909691143,
      "tokens_seen": 1057338368
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034325977933801405,
      "loss": 2.6814,
      "theoretical_loss": 3.6299590999653875,
      "tokens_seen": 1057403904
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034324974924774323,
      "loss": 2.6038,
      "theoretical_loss": 3.629937910642721,
      "tokens_seen": 1057469440
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003432397191574724,
      "loss": 2.6944,
      "theoretical_loss": 3.6299167230008775,
      "tokens_seen": 1057534976
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003432296890672016,
      "loss": 2.8614,
      "theoretical_loss": 3.6298955370396193,
      "tokens_seen": 1057600512
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003432196589769308,
      "loss": 2.753,
      "theoretical_loss": 3.6298743527587085,
      "tokens_seen": 1057666048
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034320962888665995,
      "loss": 2.697,
      "theoretical_loss": 3.629853170157909,
      "tokens_seen": 1057731584
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003431995987963892,
      "loss": 2.6929,
      "theoretical_loss": 3.6298319892369815,
      "tokens_seen": 1057797120
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003431895687061183,
      "loss": 2.7009,
      "theoretical_loss": 3.629810809995691,
      "tokens_seen": 1057862656
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034317953861584755,
      "loss": 2.9626,
      "theoretical_loss": 3.6297896324337984,
      "tokens_seen": 1057928192
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034316950852557673,
      "loss": 2.6848,
      "theoretical_loss": 3.6297684565510675,
      "tokens_seen": 1057993728
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003431594784353059,
      "loss": 2.7164,
      "theoretical_loss": 3.629747282347261,
      "tokens_seen": 1058059264
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003431494483450351,
      "loss": 2.4662,
      "theoretical_loss": 3.6297261098221414,
      "tokens_seen": 1058124800
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034313941825476433,
      "loss": 2.8341,
      "theoretical_loss": 3.629704938975472,
      "tokens_seen": 1058190336
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034312938816449345,
      "loss": 2.8161,
      "theoretical_loss": 3.6296837698070163,
      "tokens_seen": 1058255872
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003431193580742227,
      "loss": 3.0423,
      "theoretical_loss": 3.6296626023165364,
      "tokens_seen": 1058321408
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003431093279839518,
      "loss": 2.9207,
      "theoretical_loss": 3.6296414365037957,
      "tokens_seen": 1058386944
    },
    {
      "epoch": 3.05,
      "objective/train/docs_used": 1194725,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.449052095413208,
      "objective/train/theoretical_loss": 3.6296361453127295,
      "objective/train/tokens_used": 1078863328,
      "theoretical_loss": 3.6296361453127295,
      "tokens_seen": 1058403328
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034309929789368105,
      "loss": 2.7992,
      "theoretical_loss": 3.6296202723685576,
      "tokens_seen": 1058452480
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003430892678034103,
      "loss": 2.9941,
      "theoretical_loss": 3.6295991099105853,
      "tokens_seen": 1058518016
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003430792377131394,
      "loss": 2.6418,
      "theoretical_loss": 3.629577949129642,
      "tokens_seen": 1058583552
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034306920762286865,
      "loss": 2.7596,
      "theoretical_loss": 3.6295567900254904,
      "tokens_seen": 1058649088
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003430591775325978,
      "loss": 2.7151,
      "theoretical_loss": 3.6295356325978947,
      "tokens_seen": 1058714624
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.000343049147442327,
      "loss": 2.839,
      "theoretical_loss": 3.6295144768466177,
      "tokens_seen": 1058780160
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003430391173520562,
      "loss": 2.7535,
      "theoretical_loss": 3.629493322771423,
      "tokens_seen": 1058845696
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003430290872617854,
      "loss": 2.7381,
      "theoretical_loss": 3.6294721703720745,
      "tokens_seen": 1058911232
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034301905717151456,
      "loss": 2.8283,
      "theoretical_loss": 3.629451019648335,
      "tokens_seen": 1058976768
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034300902708124374,
      "loss": 2.5179,
      "theoretical_loss": 3.6294298705999686,
      "tokens_seen": 1059042304
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003429989969909729,
      "loss": 2.844,
      "theoretical_loss": 3.6294087232267387,
      "tokens_seen": 1059107840
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034298896690070215,
      "loss": 2.883,
      "theoretical_loss": 3.629387577528409,
      "tokens_seen": 1059173376
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003429789368104313,
      "loss": 2.6924,
      "theoretical_loss": 3.6293664335047433,
      "tokens_seen": 1059238912
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003429689067201605,
      "loss": 2.8918,
      "theoretical_loss": 3.6293452911555057,
      "tokens_seen": 1059304448
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003429588766298897,
      "loss": 2.8112,
      "theoretical_loss": 3.6293241504804596,
      "tokens_seen": 1059369984
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003429488465396189,
      "loss": 2.9203,
      "theoretical_loss": 3.6293030114793687,
      "tokens_seen": 1059435520
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034293881644934806,
      "loss": 2.7102,
      "theoretical_loss": 3.629281874151997,
      "tokens_seen": 1059501056
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034292878635907724,
      "loss": 2.6518,
      "theoretical_loss": 3.6292607384981093,
      "tokens_seen": 1059566592
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003429187562688064,
      "loss": 2.6567,
      "theoretical_loss": 3.629239604517468,
      "tokens_seen": 1059632128
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034290872617853566,
      "loss": 2.9782,
      "theoretical_loss": 3.629218472209839,
      "tokens_seen": 1059697664
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003428986960882648,
      "loss": 2.6975,
      "theoretical_loss": 3.6291973415749847,
      "tokens_seen": 1059763200
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.000342888665997994,
      "loss": 2.7407,
      "theoretical_loss": 3.6291762126126708,
      "tokens_seen": 1059828736
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034287863590772315,
      "loss": 2.9337,
      "theoretical_loss": 3.6291550853226604,
      "tokens_seen": 1059894272
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003428686058174524,
      "loss": 2.6709,
      "theoretical_loss": 3.629133959704718,
      "tokens_seen": 1059959808
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034285857572718156,
      "loss": 2.6925,
      "theoretical_loss": 3.629112835758608,
      "tokens_seen": 1060025344
    },
    {
      "epoch": 3.05,
      "objective/train/docs_used": 1195453,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8471500873565674,
      "objective/train/theoretical_loss": 3.629107555033277,
      "objective/train/tokens_used": 1080501728,
      "theoretical_loss": 3.629107555033277,
      "tokens_seen": 1060041728
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034284854563691074,
      "loss": 2.9138,
      "theoretical_loss": 3.629091713484095,
      "tokens_seen": 1060090880
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003428385155466399,
      "loss": 2.6919,
      "theoretical_loss": 3.629070592880943,
      "tokens_seen": 1060156416
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003428284854563691,
      "loss": 2.7959,
      "theoretical_loss": 3.6290494739489167,
      "tokens_seen": 1060221952
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003428184553660983,
      "loss": 2.7872,
      "theoretical_loss": 3.62902835668778,
      "tokens_seen": 1060287488
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003428084252758275,
      "loss": 2.6846,
      "theoretical_loss": 3.6290072410972987,
      "tokens_seen": 1060353024
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034279839518555665,
      "loss": 2.8737,
      "theoretical_loss": 3.628986127177236,
      "tokens_seen": 1060418560
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003427883650952859,
      "loss": 2.9303,
      "theoretical_loss": 3.6289650149273576,
      "tokens_seen": 1060484096
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034277833500501507,
      "loss": 2.7662,
      "theoretical_loss": 3.628943904347427,
      "tokens_seen": 1060549632
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034276830491474425,
      "loss": 2.8635,
      "theoretical_loss": 3.6289227954372105,
      "tokens_seen": 1060615168
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034275827482447343,
      "loss": 2.7954,
      "theoretical_loss": 3.6289016881964713,
      "tokens_seen": 1060680704
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003427482447342026,
      "loss": 2.7364,
      "theoretical_loss": 3.6288805826249755,
      "tokens_seen": 1060746240
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003427382146439318,
      "loss": 2.9048,
      "theoretical_loss": 3.628859478722487,
      "tokens_seen": 1060811776
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.000342728184553661,
      "loss": 2.9484,
      "theoretical_loss": 3.6288383764887713,
      "tokens_seen": 1060877312
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034271815446339015,
      "loss": 2.8514,
      "theoretical_loss": 3.628817275923593,
      "tokens_seen": 1060942848
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003427081243731194,
      "loss": 2.9913,
      "theoretical_loss": 3.628796177026718,
      "tokens_seen": 1061008384
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003426980942828485,
      "loss": 2.7835,
      "theoretical_loss": 3.6287750797979097,
      "tokens_seen": 1061073920
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034268806419257775,
      "loss": 2.787,
      "theoretical_loss": 3.628753984236935,
      "tokens_seen": 1061139456
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034267803410230693,
      "loss": 2.6012,
      "theoretical_loss": 3.628732890343558,
      "tokens_seen": 1061204992
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003426680040120361,
      "loss": 2.8233,
      "theoretical_loss": 3.6287117981175436,
      "tokens_seen": 1061270528
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003426579739217653,
      "loss": 2.9159,
      "theoretical_loss": 3.6286907075586585,
      "tokens_seen": 1061336064
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034264794383149453,
      "loss": 2.754,
      "theoretical_loss": 3.6286696186666667,
      "tokens_seen": 1061401600
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034263791374122366,
      "loss": 2.5168,
      "theoretical_loss": 3.6286485314413337,
      "tokens_seen": 1061467136
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003426278836509529,
      "loss": 2.8263,
      "theoretical_loss": 3.6286274458824255,
      "tokens_seen": 1061532672
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.000342617853560682,
      "loss": 2.6702,
      "theoretical_loss": 3.6286063619897067,
      "tokens_seen": 1061598208
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034260782347041125,
      "loss": 3.0237,
      "theoretical_loss": 3.6285852797629436,
      "tokens_seen": 1061663744
    },
    {
      "epoch": 3.05,
      "objective/train/docs_used": 1196907,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8292043209075928,
      "objective/train/theoretical_loss": 3.628580009466531,
      "objective/train/tokens_used": 1082140128,
      "theoretical_loss": 3.628580009466531,
      "tokens_seen": 1061680128
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034259779338014043,
      "loss": 2.798,
      "theoretical_loss": 3.628564199201901,
      "tokens_seen": 1061729280
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003425877632898696,
      "loss": 2.5776,
      "theoretical_loss": 3.628543120306346,
      "tokens_seen": 1061794816
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003425777331995988,
      "loss": 2.8096,
      "theoretical_loss": 3.628522043076042,
      "tokens_seen": 1061860352
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.000342567703109328,
      "loss": 2.8131,
      "theoretical_loss": 3.6285009675107567,
      "tokens_seen": 1061925888
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034255767301905716,
      "loss": 2.8516,
      "theoretical_loss": 3.628479893610254,
      "tokens_seen": 1061991424
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003425476429287864,
      "loss": 2.6119,
      "theoretical_loss": 3.628458821374301,
      "tokens_seen": 1062056960
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003425376128385155,
      "loss": 2.7599,
      "theoretical_loss": 3.6284377508026635,
      "tokens_seen": 1062122496
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034252758274824476,
      "loss": 2.6915,
      "theoretical_loss": 3.6284166818951067,
      "tokens_seen": 1062188032
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003425175526579739,
      "loss": 2.7241,
      "theoretical_loss": 3.6283956146513967,
      "tokens_seen": 1062253568
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003425075225677031,
      "loss": 2.6911,
      "theoretical_loss": 3.6283745490712995,
      "tokens_seen": 1062319104
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003424974924774323,
      "loss": 2.8022,
      "theoretical_loss": 3.628353485154581,
      "tokens_seen": 1062384640
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003424874623871615,
      "loss": 2.5889,
      "theoretical_loss": 3.6283324229010083,
      "tokens_seen": 1062450176
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034247743229689066,
      "loss": 2.8766,
      "theoretical_loss": 3.6283113623103462,
      "tokens_seen": 1062515712
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003424674022066199,
      "loss": 2.7854,
      "theoretical_loss": 3.6282903033823612,
      "tokens_seen": 1062581248
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.000342457372116349,
      "loss": 2.838,
      "theoretical_loss": 3.62826924611682,
      "tokens_seen": 1062646784
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034244734202607826,
      "loss": 2.7359,
      "theoretical_loss": 3.628248190513488,
      "tokens_seen": 1062712320
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003424373119358074,
      "loss": 2.9226,
      "theoretical_loss": 3.628227136572132,
      "tokens_seen": 1062777856
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003424272818455366,
      "loss": 2.942,
      "theoretical_loss": 3.6282060842925183,
      "tokens_seen": 1062843392
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003424172517552658,
      "loss": 2.8092,
      "theoretical_loss": 3.6281850336744137,
      "tokens_seen": 1062908928
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.000342407221664995,
      "loss": 2.8666,
      "theoretical_loss": 3.6281639847175837,
      "tokens_seen": 1062974464
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034239719157472416,
      "loss": 2.6878,
      "theoretical_loss": 3.6281429374217953,
      "tokens_seen": 1063040000
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034238716148445335,
      "loss": 2.7101,
      "theoretical_loss": 3.628121891786815,
      "tokens_seen": 1063105536
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003423771313941825,
      "loss": 2.9153,
      "theoretical_loss": 3.62810084781241,
      "tokens_seen": 1063171072
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034236710130391176,
      "loss": 2.8423,
      "theoretical_loss": 3.628079805498346,
      "tokens_seen": 1063236608
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003423570712136409,
      "loss": 2.5257,
      "theoretical_loss": 3.62805876484439,
      "tokens_seen": 1063302144
    },
    {
      "epoch": 3.05,
      "objective/train/docs_used": 1197465,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.75803279876709,
      "objective/train/theoretical_loss": 3.628053504940265,
      "objective/train/tokens_used": 1083778528,
      "theoretical_loss": 3.628053504940265,
      "tokens_seen": 1063318528
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003423470411233701,
      "loss": 2.5156,
      "theoretical_loss": 3.6280377258503087,
      "tokens_seen": 1063367680
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003423370110330993,
      "loss": 2.8494,
      "theoretical_loss": 3.628016688515869,
      "tokens_seen": 1063433216
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003423269809428285,
      "loss": 2.8854,
      "theoretical_loss": 3.6279956528408377,
      "tokens_seen": 1063498752
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003423169508525577,
      "loss": 2.6486,
      "theoretical_loss": 3.627974618824981,
      "tokens_seen": 1063564288
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034230692076228685,
      "loss": 2.765,
      "theoretical_loss": 3.6279535864680668,
      "tokens_seen": 1063629824
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003422968906720161,
      "loss": 2.6579,
      "theoretical_loss": 3.6279325557698616,
      "tokens_seen": 1063695360
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034228686058174527,
      "loss": 2.7489,
      "theoretical_loss": 3.6279115267301325,
      "tokens_seen": 1063760896
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034227683049147445,
      "loss": 2.7437,
      "theoretical_loss": 3.6278904993486467,
      "tokens_seen": 1063826432
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034226680040120363,
      "loss": 2.9753,
      "theoretical_loss": 3.6278694736251706,
      "tokens_seen": 1063891968
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003422567703109328,
      "loss": 2.8276,
      "theoretical_loss": 3.627848449559472,
      "tokens_seen": 1063957504
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.000342246740220662,
      "loss": 2.6558,
      "theoretical_loss": 3.6278274271513182,
      "tokens_seen": 1064023040
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003422367101303912,
      "loss": 2.8123,
      "theoretical_loss": 3.627806406400476,
      "tokens_seen": 1064088576
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034222668004012035,
      "loss": 2.515,
      "theoretical_loss": 3.6277853873067127,
      "tokens_seen": 1064154112
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003422166499498496,
      "loss": 2.9208,
      "theoretical_loss": 3.627764369869796,
      "tokens_seen": 1064219648
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003422066198595787,
      "loss": 2.8514,
      "theoretical_loss": 3.6277433540894926,
      "tokens_seen": 1064285184
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034219658976930795,
      "loss": 2.6061,
      "theoretical_loss": 3.6277223399655707,
      "tokens_seen": 1064350720
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034218655967903713,
      "loss": 2.8817,
      "theoretical_loss": 3.627701327497798,
      "tokens_seen": 1064416256
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003421765295887663,
      "loss": 2.7702,
      "theoretical_loss": 3.627680316685941,
      "tokens_seen": 1064481792
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003421664994984955,
      "loss": 2.9302,
      "theoretical_loss": 3.6276593075297674,
      "tokens_seen": 1064547328
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034215646940822473,
      "loss": 2.8216,
      "theoretical_loss": 3.6276383000290457,
      "tokens_seen": 1064612864
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034214643931795386,
      "loss": 2.6398,
      "theoretical_loss": 3.627617294183543,
      "tokens_seen": 1064678400
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003421364092276831,
      "loss": 2.873,
      "theoretical_loss": 3.6275962899930265,
      "tokens_seen": 1064743936
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003421263791374122,
      "loss": 2.887,
      "theoretical_loss": 3.627575287457265,
      "tokens_seen": 1064809472
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034211634904714145,
      "loss": 2.7337,
      "theoretical_loss": 3.6275542865760255,
      "tokens_seen": 1064875008
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034210631895687063,
      "loss": 3.067,
      "theoretical_loss": 3.6275332873490767,
      "tokens_seen": 1064940544
    },
    {
      "epoch": 3.05,
      "objective/train/docs_used": 1198703,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5579044818878174,
      "objective/train/theoretical_loss": 3.627528037800795,
      "objective/train/tokens_used": 1085416928,
      "theoretical_loss": 3.627528037800795,
      "tokens_seen": 1064956928
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003420962888665998,
      "loss": 2.7133,
      "theoretical_loss": 3.627512289776185,
      "tokens_seen": 1065006080
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.000342086258776329,
      "loss": 2.7935,
      "theoretical_loss": 3.62749129385712,
      "tokens_seen": 1065071616
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003420762286860582,
      "loss": 2.6849,
      "theoretical_loss": 3.6274702995916486,
      "tokens_seen": 1065137152
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034206619859578736,
      "loss": 2.8924,
      "theoretical_loss": 3.627449306979539,
      "tokens_seen": 1065202688
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003420561685055166,
      "loss": 2.7858,
      "theoretical_loss": 3.62742831602056,
      "tokens_seen": 1065268224
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003420461384152457,
      "loss": 2.8484,
      "theoretical_loss": 3.6274073267144793,
      "tokens_seen": 1065333760
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034203610832497496,
      "loss": 2.7231,
      "theoretical_loss": 3.627386339061064,
      "tokens_seen": 1065399296
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003420260782347041,
      "loss": 2.8332,
      "theoretical_loss": 3.6273653530600845,
      "tokens_seen": 1065464832
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003420160481444333,
      "loss": 2.7879,
      "theoretical_loss": 3.627344368711307,
      "tokens_seen": 1065530368
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003420060180541625,
      "loss": 2.4803,
      "theoretical_loss": 3.6273233860145013,
      "tokens_seen": 1065595904
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003419959879638917,
      "loss": 2.8658,
      "theoretical_loss": 3.627302404969435,
      "tokens_seen": 1065661440
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034198595787362086,
      "loss": 2.7391,
      "theoretical_loss": 3.6272814255758763,
      "tokens_seen": 1065726976
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003419759277833501,
      "loss": 2.9654,
      "theoretical_loss": 3.6272604478335944,
      "tokens_seen": 1065792512
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003419658976930792,
      "loss": 2.5866,
      "theoretical_loss": 3.6272394717423575,
      "tokens_seen": 1065858048
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034195586760280846,
      "loss": 2.7491,
      "theoretical_loss": 3.627218497301934,
      "tokens_seen": 1065923584
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003419458375125376,
      "loss": 2.94,
      "theoretical_loss": 3.627197524512092,
      "tokens_seen": 1065989120
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003419358074222668,
      "loss": 2.7148,
      "theoretical_loss": 3.6271765533726015,
      "tokens_seen": 1066054656
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.000341925777331996,
      "loss": 3.0207,
      "theoretical_loss": 3.62715558388323,
      "tokens_seen": 1066120192
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003419157472417252,
      "loss": 2.8715,
      "theoretical_loss": 3.6271346160437465,
      "tokens_seen": 1066185728
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034190571715145436,
      "loss": 2.8406,
      "theoretical_loss": 3.6271136498539205,
      "tokens_seen": 1066251264
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034189568706118355,
      "loss": 2.6654,
      "theoretical_loss": 3.6270926853135195,
      "tokens_seen": 1066316800
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003418856569709127,
      "loss": 2.7278,
      "theoretical_loss": 3.627071722422314,
      "tokens_seen": 1066382336
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034187562688064196,
      "loss": 2.7196,
      "theoretical_loss": 3.6270507611800715,
      "tokens_seen": 1066447872
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003418655967903711,
      "loss": 2.8842,
      "theoretical_loss": 3.6270298015865614,
      "tokens_seen": 1066513408
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003418555667001003,
      "loss": 2.5783,
      "theoretical_loss": 3.627008843641553,
      "tokens_seen": 1066578944
    },
    {
      "epoch": 3.05,
      "objective/train/docs_used": 1199964,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9109983444213867,
      "objective/train/theoretical_loss": 3.6270036044128524,
      "objective/train/tokens_used": 1087055328,
      "theoretical_loss": 3.6270036044128524,
      "tokens_seen": 1066595328
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034184553660982945,
      "loss": 3.0025,
      "theoretical_loss": 3.626987887344815,
      "tokens_seen": 1066644480
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003418355065195587,
      "loss": 2.7713,
      "theoretical_loss": 3.626966932696117,
      "tokens_seen": 1066710016
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034182547642928787,
      "loss": 2.7343,
      "theoretical_loss": 3.626945979695228,
      "tokens_seen": 1066775552
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034181544633901705,
      "loss": 2.942,
      "theoretical_loss": 3.626925028341917,
      "tokens_seen": 1066841088
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034180541624874623,
      "loss": 3.002,
      "theoretical_loss": 3.6269040786359534,
      "tokens_seen": 1066906624
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034179538615847547,
      "loss": 2.9148,
      "theoretical_loss": 3.626883130577106,
      "tokens_seen": 1066972160
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003417853560682046,
      "loss": 2.864,
      "theoretical_loss": 3.6268621841651445,
      "tokens_seen": 1067037696
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034177532597793383,
      "loss": 2.8888,
      "theoretical_loss": 3.626841239399839,
      "tokens_seen": 1067103232
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034176529588766295,
      "loss": 2.9927,
      "theoretical_loss": 3.626820296280958,
      "tokens_seen": 1067168768
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003417552657973922,
      "loss": 2.6687,
      "theoretical_loss": 3.626799354808271,
      "tokens_seen": 1067234304
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034174523570712137,
      "loss": 2.7331,
      "theoretical_loss": 3.626778414981548,
      "tokens_seen": 1067299840
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034173520561685055,
      "loss": 2.6517,
      "theoretical_loss": 3.626757476800558,
      "tokens_seen": 1067365376
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034172517552657973,
      "loss": 3.0107,
      "theoretical_loss": 3.6267365402650715,
      "tokens_seen": 1067430912
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003417151454363089,
      "loss": 3.1109,
      "theoretical_loss": 3.6267156053748577,
      "tokens_seen": 1067496448
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003417051153460381,
      "loss": 2.6219,
      "theoretical_loss": 3.6266946721296858,
      "tokens_seen": 1067561984
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034169508525576733,
      "loss": 2.7849,
      "theoretical_loss": 3.626673740529326,
      "tokens_seen": 1067627520
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034168505516549646,
      "loss": 2.7899,
      "theoretical_loss": 3.6266528105735487,
      "tokens_seen": 1067693056
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003416750250752257,
      "loss": 2.6937,
      "theoretical_loss": 3.6266318822621226,
      "tokens_seen": 1067758592
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003416649949849548,
      "loss": 2.7576,
      "theoretical_loss": 3.6266109555948187,
      "tokens_seen": 1067824128
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034165496489468406,
      "loss": 2.7506,
      "theoretical_loss": 3.626590030571406,
      "tokens_seen": 1067889664
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034164493480441324,
      "loss": 2.7177,
      "theoretical_loss": 3.626569107191655,
      "tokens_seen": 1067955200
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003416349047141424,
      "loss": 2.6327,
      "theoretical_loss": 3.6265481854553356,
      "tokens_seen": 1068020736
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003416248746238716,
      "loss": 2.8429,
      "theoretical_loss": 3.6265272653622183,
      "tokens_seen": 1068086272
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034161484453360083,
      "loss": 2.7674,
      "theoretical_loss": 3.6265063469120724,
      "tokens_seen": 1068151808
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034160481444332996,
      "loss": 2.796,
      "theoretical_loss": 3.626485430104669,
      "tokens_seen": 1068217344
    },
    {
      "epoch": 3.05,
      "objective/train/docs_used": 1200699,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.70048451423645,
      "objective/train/theoretical_loss": 3.6264802011594695,
      "objective/train/tokens_used": 1088693728,
      "theoretical_loss": 3.6264802011594695,
      "tokens_seen": 1068233728
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003415947843530592,
      "loss": 2.6839,
      "theoretical_loss": 3.6264645149397774,
      "tokens_seen": 1068282880
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003415847542627884,
      "loss": 2.7033,
      "theoretical_loss": 3.6264436014171686,
      "tokens_seen": 1068348416
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034157472417251756,
      "loss": 2.6647,
      "theoretical_loss": 3.6264226895366125,
      "tokens_seen": 1068413952
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003415646940822468,
      "loss": 2.8639,
      "theoretical_loss": 3.62640177929788,
      "tokens_seen": 1068479488
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003415546639919759,
      "loss": 2.8761,
      "theoretical_loss": 3.626380870700741,
      "tokens_seen": 1068545024
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034154463390170516,
      "loss": 2.8552,
      "theoretical_loss": 3.6263599637449655,
      "tokens_seen": 1068610560
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003415346038114343,
      "loss": 2.8734,
      "theoretical_loss": 3.6263390584303252,
      "tokens_seen": 1068676096
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003415245737211635,
      "loss": 2.7517,
      "theoretical_loss": 3.6263181547565893,
      "tokens_seen": 1068741632
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003415145436308927,
      "loss": 2.7747,
      "theoretical_loss": 3.62629725272353,
      "tokens_seen": 1068807168
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003415045135406219,
      "loss": 2.8721,
      "theoretical_loss": 3.6262763523309163,
      "tokens_seen": 1068872704
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034149448345035106,
      "loss": 2.7056,
      "theoretical_loss": 3.62625545357852,
      "tokens_seen": 1068938240
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003414844533600803,
      "loss": 2.6699,
      "theoretical_loss": 3.6262345564661116,
      "tokens_seen": 1069003776
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003414744232698094,
      "loss": 2.8384,
      "theoretical_loss": 3.6262136609934617,
      "tokens_seen": 1069069312
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034146439317953866,
      "loss": 2.6113,
      "theoretical_loss": 3.626192767160341,
      "tokens_seen": 1069134848
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003414543630892678,
      "loss": 2.8847,
      "theoretical_loss": 3.6261718749665204,
      "tokens_seen": 1069200384
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.000341444332998997,
      "loss": 2.9789,
      "theoretical_loss": 3.6261509844117708,
      "tokens_seen": 1069265920
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003414343029087262,
      "loss": 2.69,
      "theoretical_loss": 3.626130095495864,
      "tokens_seen": 1069331456
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003414242728184554,
      "loss": 2.8193,
      "theoretical_loss": 3.62610920821857,
      "tokens_seen": 1069396992
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034141424272818456,
      "loss": 2.7256,
      "theoretical_loss": 3.62608832257966,
      "tokens_seen": 1069462528
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034140421263791375,
      "loss": 2.8196,
      "theoretical_loss": 3.626067438578905,
      "tokens_seen": 1069528064
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034139418254764293,
      "loss": 2.8286,
      "theoretical_loss": 3.626046556216077,
      "tokens_seen": 1069593600
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034138415245737216,
      "loss": 2.8624,
      "theoretical_loss": 3.626025675490946,
      "tokens_seen": 1069659136
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003413741223671013,
      "loss": 2.8349,
      "theoretical_loss": 3.6260047964032847,
      "tokens_seen": 1069724672
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003413640922768305,
      "loss": 2.8984,
      "theoretical_loss": 3.625983918952863,
      "tokens_seen": 1069790208
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034135406218655965,
      "loss": 2.7387,
      "theoretical_loss": 3.6259630431394525,
      "tokens_seen": 1069855744
    },
    {
      "epoch": 3.05,
      "objective/train/docs_used": 1201919,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 1.7883902788162231,
      "objective/train/theoretical_loss": 3.625957824441856,
      "objective/train/tokens_used": 1090332128,
      "theoretical_loss": 3.625957824441856,
      "tokens_seen": 1069872128
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003413440320962889,
      "loss": 2.4208,
      "theoretical_loss": 3.6259421689628253,
      "tokens_seen": 1069921280
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034133400200601807,
      "loss": 2.8423,
      "theoretical_loss": 3.625921296422752,
      "tokens_seen": 1069986816
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034132397191574725,
      "loss": 2.7369,
      "theoretical_loss": 3.625900425519004,
      "tokens_seen": 1070052352
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034131394182547643,
      "loss": 2.9399,
      "theoretical_loss": 3.6258795562513537,
      "tokens_seen": 1070117888
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034130391173520567,
      "loss": 2.7031,
      "theoretical_loss": 3.6258586886195725,
      "tokens_seen": 1070183424
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003412938816449348,
      "loss": 2.8105,
      "theoretical_loss": 3.6258378226234314,
      "tokens_seen": 1070248960
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034128385155466403,
      "loss": 2.6984,
      "theoretical_loss": 3.6258169582627024,
      "tokens_seen": 1070314496
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034127382146439315,
      "loss": 2.7699,
      "theoretical_loss": 3.625796095537157,
      "tokens_seen": 1070380032
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003412637913741224,
      "loss": 2.79,
      "theoretical_loss": 3.625775234446567,
      "tokens_seen": 1070445568
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034125376128385157,
      "loss": 2.5291,
      "theoretical_loss": 3.6257543749907044,
      "tokens_seen": 1070511104
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034124373119358075,
      "loss": 2.7412,
      "theoretical_loss": 3.625733517169341,
      "tokens_seen": 1070576640
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034123370110330993,
      "loss": 2.7322,
      "theoretical_loss": 3.6257126609822485,
      "tokens_seen": 1070642176
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003412236710130391,
      "loss": 2.7944,
      "theoretical_loss": 3.6256918064291987,
      "tokens_seen": 1070707712
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003412136409227683,
      "loss": 2.6609,
      "theoretical_loss": 3.625670953509964,
      "tokens_seen": 1070773248
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034120361083249753,
      "loss": 2.6465,
      "theoretical_loss": 3.625650102224316,
      "tokens_seen": 1070838784
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034119358074222666,
      "loss": 2.5707,
      "theoretical_loss": 3.6256292525720273,
      "tokens_seen": 1070904320
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003411835506519559,
      "loss": 2.4912,
      "theoretical_loss": 3.625608404552869,
      "tokens_seen": 1070969856
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.000341173520561685,
      "loss": 2.5813,
      "theoretical_loss": 3.6255875581666146,
      "tokens_seen": 1071035392
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034116349047141426,
      "loss": 2.8009,
      "theoretical_loss": 3.6255667134130354,
      "tokens_seen": 1071100928
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034115346038114344,
      "loss": 2.7488,
      "theoretical_loss": 3.6255458702919037,
      "tokens_seen": 1071166464
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003411434302908726,
      "loss": 2.8851,
      "theoretical_loss": 3.6255250288029917,
      "tokens_seen": 1071232000
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003411334002006018,
      "loss": 2.879,
      "theoretical_loss": 3.6255041889460724,
      "tokens_seen": 1071297536
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034112337011033103,
      "loss": 2.8509,
      "theoretical_loss": 3.625483350720918,
      "tokens_seen": 1071363072
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034111334002006016,
      "loss": 2.7807,
      "theoretical_loss": 3.6254625141273,
      "tokens_seen": 1071428608
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003411033099297894,
      "loss": 2.8458,
      "theoretical_loss": 3.625441679164992,
      "tokens_seen": 1071494144
    },
    {
      "epoch": 3.05,
      "objective/train/docs_used": 1202634,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.0612995624542236,
      "objective/train/theoretical_loss": 3.6254364706792805,
      "objective/train/tokens_used": 1091970528,
      "theoretical_loss": 3.6254364706792805,
      "tokens_seen": 1071510528
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003410932798395185,
      "loss": 2.8313,
      "theoretical_loss": 3.6254208458337662,
      "tokens_seen": 1071559680
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034108324974924776,
      "loss": 2.7383,
      "theoretical_loss": 3.625400014133395,
      "tokens_seen": 1071625216
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.00034107321965897694,
      "loss": 2.9735,
      "theoretical_loss": 3.6253791840636507,
      "tokens_seen": 1071690752
    },
    {
      "epoch": 3.05,
      "learning_rate": 0.0003410631895687061,
      "loss": 2.9058,
      "theoretical_loss": 3.6253583556243063,
      "tokens_seen": 1071756288
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003410531594784353,
      "loss": 2.8198,
      "theoretical_loss": 3.625337528815135,
      "tokens_seen": 1071821824
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003410431293881645,
      "loss": 2.8492,
      "theoretical_loss": 3.6253167036359084,
      "tokens_seen": 1071887360
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00034103309929789366,
      "loss": 2.7865,
      "theoretical_loss": 3.6252958800864006,
      "tokens_seen": 1071952896
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003410230692076229,
      "loss": 2.9489,
      "theoretical_loss": 3.625275058166384,
      "tokens_seen": 1072018432
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.000341013039117352,
      "loss": 2.99,
      "theoretical_loss": 3.6252542378756303,
      "tokens_seen": 1072083968
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00034100300902708126,
      "loss": 2.6486,
      "theoretical_loss": 3.6252334192139144,
      "tokens_seen": 1072149504
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00034099297893681044,
      "loss": 2.6888,
      "theoretical_loss": 3.6252126021810076,
      "tokens_seen": 1072215040
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003409829488465396,
      "loss": 2.738,
      "theoretical_loss": 3.625191786776684,
      "tokens_seen": 1072280576
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003409729187562688,
      "loss": 2.8253,
      "theoretical_loss": 3.6251709730007162,
      "tokens_seen": 1072346112
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.000340962888665998,
      "loss": 2.5057,
      "theoretical_loss": 3.6251501608528773,
      "tokens_seen": 1072411648
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00034095285857572717,
      "loss": 2.8812,
      "theoretical_loss": 3.6251293503329407,
      "tokens_seen": 1072477184
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003409428284854564,
      "loss": 2.9102,
      "theoretical_loss": 3.625108541440679,
      "tokens_seen": 1072542720
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00034093279839518553,
      "loss": 2.6922,
      "theoretical_loss": 3.6250877341758665,
      "tokens_seen": 1072608256
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00034092276830491477,
      "loss": 2.8168,
      "theoretical_loss": 3.6250669285382755,
      "tokens_seen": 1072673792
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003409127382146439,
      "loss": 2.7439,
      "theoretical_loss": 3.62504612452768,
      "tokens_seen": 1072739328
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00034090270812437313,
      "loss": 2.9342,
      "theoretical_loss": 3.625025322143853,
      "tokens_seen": 1072804864
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003408926780341023,
      "loss": 2.8336,
      "theoretical_loss": 3.625004521386568,
      "tokens_seen": 1072870400
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003408826479438315,
      "loss": 2.8846,
      "theoretical_loss": 3.6249837222555987,
      "tokens_seen": 1072935936
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00034087261785356067,
      "loss": 2.703,
      "theoretical_loss": 3.624962924750718,
      "tokens_seen": 1073001472
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00034086258776328985,
      "loss": 2.6356,
      "theoretical_loss": 3.6249421288717,
      "tokens_seen": 1073067008
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00034085255767301903,
      "loss": 2.8443,
      "theoretical_loss": 3.6249213346183184,
      "tokens_seen": 1073132544
    },
    {
      "epoch": 3.06,
      "objective/train/docs_used": 1203719,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.52689266204834,
      "objective/train/theoretical_loss": 3.6249161363089524,
      "objective/train/tokens_used": 1093608928,
      "theoretical_loss": 3.6249161363089524,
      "tokens_seen": 1073148928
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00034084252758274827,
      "loss": 2.6909,
      "theoretical_loss": 3.624900541990346,
      "tokens_seen": 1073198080
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00034083249749247745,
      "loss": 2.9078,
      "theoretical_loss": 3.6248797509875583,
      "tokens_seen": 1073263616
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00034082246740220663,
      "loss": 2.8621,
      "theoretical_loss": 3.6248589616097275,
      "tokens_seen": 1073329152
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00034081243731193587,
      "loss": 2.7082,
      "theoretical_loss": 3.6248381738566278,
      "tokens_seen": 1073394688
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.000340802407221665,
      "loss": 2.8163,
      "theoretical_loss": 3.624817387728033,
      "tokens_seen": 1073460224
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00034079237713139423,
      "loss": 2.8633,
      "theoretical_loss": 3.624796603223717,
      "tokens_seen": 1073525760
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00034078234704112335,
      "loss": 2.6537,
      "theoretical_loss": 3.6247758203434537,
      "tokens_seen": 1073591296
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003407723169508526,
      "loss": 2.843,
      "theoretical_loss": 3.624755039087017,
      "tokens_seen": 1073656832
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00034076228686058177,
      "loss": 2.9657,
      "theoretical_loss": 3.624734259454182,
      "tokens_seen": 1073722368
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00034075225677031095,
      "loss": 2.7164,
      "theoretical_loss": 3.6247134814447204,
      "tokens_seen": 1073787904
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00034074222668004013,
      "loss": 2.6927,
      "theoretical_loss": 3.6246927050584086,
      "tokens_seen": 1073853440
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003407321965897693,
      "loss": 2.7393,
      "theoretical_loss": 3.62467193029502,
      "tokens_seen": 1073918976
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003407221664994985,
      "loss": 2.8091,
      "theoretical_loss": 3.624651157154328,
      "tokens_seen": 1073984512
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00034071213640922773,
      "loss": 2.6501,
      "theoretical_loss": 3.6246303856361077,
      "tokens_seen": 1074050048
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00034070210631895686,
      "loss": 2.7927,
      "theoretical_loss": 3.6246096157401335,
      "tokens_seen": 1074115584
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003406920762286861,
      "loss": 2.8565,
      "theoretical_loss": 3.6245888474661796,
      "tokens_seen": 1074181120
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003406820461384152,
      "loss": 2.7706,
      "theoretical_loss": 3.62456808081402,
      "tokens_seen": 1074246656
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00034067201604814446,
      "loss": 2.7304,
      "theoretical_loss": 3.6245473157834294,
      "tokens_seen": 1074312192
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00034066198595787364,
      "loss": 2.8795,
      "theoretical_loss": 3.6245265523741823,
      "tokens_seen": 1074377728
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003406519558676028,
      "loss": 2.8952,
      "theoretical_loss": 3.6245057905860527,
      "tokens_seen": 1074443264
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.000340641925777332,
      "loss": 2.854,
      "theoretical_loss": 3.624485030418816,
      "tokens_seen": 1074508800
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00034063189568706123,
      "loss": 2.7502,
      "theoretical_loss": 3.624464271872246,
      "tokens_seen": 1074574336
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00034062186559679036,
      "loss": 2.5958,
      "theoretical_loss": 3.6244435149461185,
      "tokens_seen": 1074639872
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003406118355065196,
      "loss": 2.8345,
      "theoretical_loss": 3.624422759640207,
      "tokens_seen": 1074705408
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003406018054162487,
      "loss": 2.8674,
      "theoretical_loss": 3.6244020059542867,
      "tokens_seen": 1074770944
    },
    {
      "epoch": 3.06,
      "objective/train/docs_used": 1204483,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6257479190826416,
      "objective/train/theoretical_loss": 3.624396817785904,
      "objective/train/tokens_used": 1095247328,
      "theoretical_loss": 3.624396817785904,
      "tokens_seen": 1074787328
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00034059177532597796,
      "loss": 2.763,
      "theoretical_loss": 3.6243812538881324,
      "tokens_seen": 1074836480
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00034058174523570714,
      "loss": 2.8225,
      "theoretical_loss": 3.624360503441519,
      "tokens_seen": 1074902016
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003405717151454363,
      "loss": 2.7167,
      "theoretical_loss": 3.6243397546142218,
      "tokens_seen": 1074967552
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003405616850551655,
      "loss": 2.7695,
      "theoretical_loss": 3.6243190074060143,
      "tokens_seen": 1075033088
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003405516549648947,
      "loss": 2.664,
      "theoretical_loss": 3.624298261816673,
      "tokens_seen": 1075098624
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00034054162487462386,
      "loss": 2.8657,
      "theoretical_loss": 3.6242775178459716,
      "tokens_seen": 1075164160
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003405315947843531,
      "loss": 2.707,
      "theoretical_loss": 3.624256775493687,
      "tokens_seen": 1075229696
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003405215646940822,
      "loss": 2.8292,
      "theoretical_loss": 3.6242360347595923,
      "tokens_seen": 1075295232
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00034051153460381146,
      "loss": 2.7094,
      "theoretical_loss": 3.6242152956434635,
      "tokens_seen": 1075360768
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00034050150451354064,
      "loss": 2.9164,
      "theoretical_loss": 3.624194558145076,
      "tokens_seen": 1075426304
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003404914744232698,
      "loss": 2.6311,
      "theoretical_loss": 3.624173822264205,
      "tokens_seen": 1075491840
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.000340481444332999,
      "loss": 2.5965,
      "theoretical_loss": 3.624153088000625,
      "tokens_seen": 1075557376
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003404714142427282,
      "loss": 2.8016,
      "theoretical_loss": 3.6241323553541127,
      "tokens_seen": 1075622912
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00034046138415245737,
      "loss": 2.7171,
      "theoretical_loss": 3.624111624324442,
      "tokens_seen": 1075688448
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003404513540621866,
      "loss": 2.8304,
      "theoretical_loss": 3.6240908949113897,
      "tokens_seen": 1075753984
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00034044132397191573,
      "loss": 2.9595,
      "theoretical_loss": 3.62407016711473,
      "tokens_seen": 1075819520
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00034043129388164497,
      "loss": 2.6081,
      "theoretical_loss": 3.6240494409342396,
      "tokens_seen": 1075885056
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003404212637913741,
      "loss": 3.0219,
      "theoretical_loss": 3.6240287163696934,
      "tokens_seen": 1075950592
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00034041123370110333,
      "loss": 2.8945,
      "theoretical_loss": 3.624007993420866,
      "tokens_seen": 1076016128
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003404012036108325,
      "loss": 2.9308,
      "theoretical_loss": 3.6239872720875352,
      "tokens_seen": 1076081664
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003403911735205617,
      "loss": 2.6346,
      "theoretical_loss": 3.623966552369475,
      "tokens_seen": 1076147200
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00034038114343029087,
      "loss": 2.754,
      "theoretical_loss": 3.623945834266462,
      "tokens_seen": 1076212736
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00034037111334002005,
      "loss": 2.5849,
      "theoretical_loss": 3.6239251177782714,
      "tokens_seen": 1076278272
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00034036108324974923,
      "loss": 2.8053,
      "theoretical_loss": 3.6239044029046794,
      "tokens_seen": 1076343808
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00034035105315947847,
      "loss": 2.7989,
      "theoretical_loss": 3.623883689645462,
      "tokens_seen": 1076409344
    },
    {
      "epoch": 3.06,
      "objective/train/docs_used": 1205647,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.873117446899414,
      "objective/train/theoretical_loss": 3.623878511582877,
      "objective/train/tokens_used": 1096885728,
      "theoretical_loss": 3.623878511582877,
      "tokens_seen": 1076425728
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003403410230692076,
      "loss": 2.8016,
      "theoretical_loss": 3.6238629780003944,
      "tokens_seen": 1076474880
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00034033099297893683,
      "loss": 2.8755,
      "theoretical_loss": 3.6238422679692532,
      "tokens_seen": 1076540416
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.000340320962888666,
      "loss": 2.6786,
      "theoretical_loss": 3.623821559551814,
      "tokens_seen": 1076605952
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003403109327983952,
      "loss": 2.5573,
      "theoretical_loss": 3.623800852747854,
      "tokens_seen": 1076671488
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003403009027081244,
      "loss": 2.8492,
      "theoretical_loss": 3.623780147557147,
      "tokens_seen": 1076737024
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00034029087261785356,
      "loss": 2.7657,
      "theoretical_loss": 3.623759443979471,
      "tokens_seen": 1076802560
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00034028084252758274,
      "loss": 2.8066,
      "theoretical_loss": 3.623738742014602,
      "tokens_seen": 1076868096
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00034027081243731197,
      "loss": 2.7101,
      "theoretical_loss": 3.6237180416623156,
      "tokens_seen": 1076933632
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003402607823470411,
      "loss": 2.8152,
      "theoretical_loss": 3.623697342922388,
      "tokens_seen": 1076999168
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00034025075225677033,
      "loss": 2.6212,
      "theoretical_loss": 3.623676645794596,
      "tokens_seen": 1077064704
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00034024072216649946,
      "loss": 2.7761,
      "theoretical_loss": 3.6236559502787165,
      "tokens_seen": 1077130240
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003402306920762287,
      "loss": 2.7202,
      "theoretical_loss": 3.6236352563745244,
      "tokens_seen": 1077195776
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003402206619859579,
      "loss": 2.6571,
      "theoretical_loss": 3.623614564081797,
      "tokens_seen": 1077261312
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00034021063189568706,
      "loss": 2.8752,
      "theoretical_loss": 3.6235938734003117,
      "tokens_seen": 1077326848
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00034020060180541624,
      "loss": 2.6293,
      "theoretical_loss": 3.6235731843298433,
      "tokens_seen": 1077392384
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003401905717151454,
      "loss": 2.7159,
      "theoretical_loss": 3.623552496870169,
      "tokens_seen": 1077457920
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003401805416248746,
      "loss": 2.7924,
      "theoretical_loss": 3.6235318110210657,
      "tokens_seen": 1077523456
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00034017051153460384,
      "loss": 2.7635,
      "theoretical_loss": 3.62351112678231,
      "tokens_seen": 1077588992
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00034016048144433296,
      "loss": 2.9351,
      "theoretical_loss": 3.6234904441536786,
      "tokens_seen": 1077654528
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003401504513540622,
      "loss": 2.9664,
      "theoretical_loss": 3.623469763134948,
      "tokens_seen": 1077720064
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003401404212637914,
      "loss": 2.7644,
      "theoretical_loss": 3.6234490837258955,
      "tokens_seen": 1077785600
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00034013039117352056,
      "loss": 2.8821,
      "theoretical_loss": 3.6234284059262976,
      "tokens_seen": 1077851136
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00034012036108324974,
      "loss": 2.9963,
      "theoretical_loss": 3.6234077297359315,
      "tokens_seen": 1077916672
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003401103309929789,
      "loss": 2.5783,
      "theoretical_loss": 3.6233870551545735,
      "tokens_seen": 1077982208
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003401003009027081,
      "loss": 2.8951,
      "theoretical_loss": 3.6233663821820015,
      "tokens_seen": 1078047744
    },
    {
      "epoch": 3.06,
      "objective/train/docs_used": 1210610,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8817059993743896,
      "objective/train/theoretical_loss": 3.6233612141902043,
      "objective/train/tokens_used": 1098524128,
      "theoretical_loss": 3.6233612141902043,
      "tokens_seen": 1078064128
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00034009027081243734,
      "loss": 2.9285,
      "theoretical_loss": 3.623345710817991,
      "tokens_seen": 1078113280
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003400802407221665,
      "loss": 2.929,
      "theoretical_loss": 3.623325041062321,
      "tokens_seen": 1078178816
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003400702106318957,
      "loss": 2.7234,
      "theoretical_loss": 3.623304372914767,
      "tokens_seen": 1078244352
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003400601805416249,
      "loss": 2.6798,
      "theoretical_loss": 3.6232837063751075,
      "tokens_seen": 1078309888
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00034005015045135406,
      "loss": 2.9209,
      "theoretical_loss": 3.6232630414431184,
      "tokens_seen": 1078375424
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003400401203610833,
      "loss": 2.576,
      "theoretical_loss": 3.6232423781185776,
      "tokens_seen": 1078440960
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003400300902708124,
      "loss": 2.7724,
      "theoretical_loss": 3.623221716401263,
      "tokens_seen": 1078506496
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00034002006018054166,
      "loss": 2.7263,
      "theoretical_loss": 3.6232010562909505,
      "tokens_seen": 1078572032
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00034001003009027084,
      "loss": 2.8475,
      "theoretical_loss": 3.623180397787419,
      "tokens_seen": 1078637568
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00034,
      "loss": 2.863,
      "theoretical_loss": 3.623159740890445,
      "tokens_seen": 1078703104
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003399899699097292,
      "loss": 2.8764,
      "theoretical_loss": 3.6231390855998056,
      "tokens_seen": 1078768640
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003399799398194584,
      "loss": 2.6731,
      "theoretical_loss": 3.6231184319152794,
      "tokens_seen": 1078834176
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033996990972918757,
      "loss": 2.9074,
      "theoretical_loss": 3.6230977798366437,
      "tokens_seen": 1078899712
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003399598796389168,
      "loss": 2.6713,
      "theoretical_loss": 3.6230771293636757,
      "tokens_seen": 1078965248
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033994984954864593,
      "loss": 2.5203,
      "theoretical_loss": 3.623056480496153,
      "tokens_seen": 1079030784
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033993981945837517,
      "loss": 2.8662,
      "theoretical_loss": 3.6230358332338533,
      "tokens_seen": 1079096320
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003399297893681043,
      "loss": 2.7636,
      "theoretical_loss": 3.623015187576555,
      "tokens_seen": 1079161856
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033991975927783353,
      "loss": 2.7715,
      "theoretical_loss": 3.622994543524035,
      "tokens_seen": 1079227392
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003399097291875627,
      "loss": 2.8445,
      "theoretical_loss": 3.6229739010760715,
      "tokens_seen": 1079292928
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003398996990972919,
      "loss": 2.9655,
      "theoretical_loss": 3.6229532602324426,
      "tokens_seen": 1079358464
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033988966900702107,
      "loss": 2.6079,
      "theoretical_loss": 3.622932620992926,
      "tokens_seen": 1079424000
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033987963891675025,
      "loss": 2.6488,
      "theoretical_loss": 3.6229119833572994,
      "tokens_seen": 1079489536
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033986960882647943,
      "loss": 2.5739,
      "theoretical_loss": 3.622891347325341,
      "tokens_seen": 1079555072
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033985957873620867,
      "loss": 2.8467,
      "theoretical_loss": 3.6228707128968294,
      "tokens_seen": 1079620608
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003398495486459378,
      "loss": 2.8359,
      "theoretical_loss": 3.622850080071542,
      "tokens_seen": 1079686144
    },
    {
      "epoch": 3.06,
      "objective/train/docs_used": 1215587,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9586844444274902,
      "objective/train/theoretical_loss": 3.6228449221156978,
      "objective/train/tokens_used": 1100162528,
      "theoretical_loss": 3.6228449221156978,
      "tokens_seen": 1079702528
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033983951855566703,
      "loss": 2.838,
      "theoretical_loss": 3.6228294488492567,
      "tokens_seen": 1079751680
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003398294884653962,
      "loss": 2.6845,
      "theoretical_loss": 3.6228088192297525,
      "tokens_seen": 1079817216
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003398194583751254,
      "loss": 2.6935,
      "theoretical_loss": 3.622788191212807,
      "tokens_seen": 1079882752
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003398094282848546,
      "loss": 2.8,
      "theoretical_loss": 3.622767564798199,
      "tokens_seen": 1079948288
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033979939819458376,
      "loss": 2.8069,
      "theoretical_loss": 3.622746939985706,
      "tokens_seen": 1080013824
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033978936810431294,
      "loss": 2.7078,
      "theoretical_loss": 3.622726316775107,
      "tokens_seen": 1080079360
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033977933801404217,
      "loss": 2.7325,
      "theoretical_loss": 3.6227056951661805,
      "tokens_seen": 1080144896
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003397693079237713,
      "loss": 2.7843,
      "theoretical_loss": 3.6226850751587043,
      "tokens_seen": 1080210432
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033975927783350053,
      "loss": 2.8463,
      "theoretical_loss": 3.6226644567524575,
      "tokens_seen": 1080275968
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033974924774322966,
      "loss": 2.7421,
      "theoretical_loss": 3.622643839947218,
      "tokens_seen": 1080341504
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003397392176529589,
      "loss": 2.8555,
      "theoretical_loss": 3.622623224742765,
      "tokens_seen": 1080407040
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003397291875626881,
      "loss": 2.8457,
      "theoretical_loss": 3.622602611138877,
      "tokens_seen": 1080472576
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033971915747241726,
      "loss": 2.8681,
      "theoretical_loss": 3.6225819991353325,
      "tokens_seen": 1080538112
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033970912738214644,
      "loss": 2.5826,
      "theoretical_loss": 3.6225613887319104,
      "tokens_seen": 1080603648
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003396990972918756,
      "loss": 2.5601,
      "theoretical_loss": 3.622540779928389,
      "tokens_seen": 1080669184
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003396890672016048,
      "loss": 2.7977,
      "theoretical_loss": 3.6225201727245473,
      "tokens_seen": 1080734720
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033967903711133404,
      "loss": 2.592,
      "theoretical_loss": 3.6224995671201645,
      "tokens_seen": 1080800256
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033966900702106316,
      "loss": 2.6062,
      "theoretical_loss": 3.622478963115019,
      "tokens_seen": 1080865792
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003396589769307924,
      "loss": 2.6532,
      "theoretical_loss": 3.6224583607088903,
      "tokens_seen": 1080931328
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003396489468405216,
      "loss": 2.7851,
      "theoretical_loss": 3.6224377599015565,
      "tokens_seen": 1080996864
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033963891675025076,
      "loss": 2.8381,
      "theoretical_loss": 3.6224171606927973,
      "tokens_seen": 1081062400
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033962888665997994,
      "loss": 2.8092,
      "theoretical_loss": 3.6223965630823916,
      "tokens_seen": 1081127936
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003396188565697091,
      "loss": 2.791,
      "theoretical_loss": 3.6223759670701186,
      "tokens_seen": 1081193472
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003396088264794383,
      "loss": 2.8194,
      "theoretical_loss": 3.622355372655757,
      "tokens_seen": 1081259008
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033959879638916754,
      "loss": 3.0145,
      "theoretical_loss": 3.622334779839087,
      "tokens_seen": 1081324544
    },
    {
      "epoch": 3.06,
      "objective/train/docs_used": 1220679,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.306006908416748,
      "objective/train/theoretical_loss": 3.6223296318845324,
      "objective/train/tokens_used": 1101800928,
      "theoretical_loss": 3.6223296318845324,
      "tokens_seen": 1081340928
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033958876629889667,
      "loss": 2.6374,
      "theoretical_loss": 3.622314188619886,
      "tokens_seen": 1081390080
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003395787362086259,
      "loss": 2.7159,
      "theoretical_loss": 3.6222935989979352,
      "tokens_seen": 1081455616
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033956870611835503,
      "loss": 2.764,
      "theoretical_loss": 3.622273010973013,
      "tokens_seen": 1081521152
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033955867602808426,
      "loss": 2.8377,
      "theoretical_loss": 3.622252424544899,
      "tokens_seen": 1081586688
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033954864593781345,
      "loss": 2.7197,
      "theoretical_loss": 3.622231839713372,
      "tokens_seen": 1081652224
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003395386158475426,
      "loss": 2.7794,
      "theoretical_loss": 3.6222112564782125,
      "tokens_seen": 1081717760
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003395285857572718,
      "loss": 2.7446,
      "theoretical_loss": 3.6221906748391994,
      "tokens_seen": 1081783296
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033951855566700104,
      "loss": 2.8116,
      "theoretical_loss": 3.622170094796112,
      "tokens_seen": 1081848832
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033950852557673017,
      "loss": 2.8156,
      "theoretical_loss": 3.6221495163487303,
      "tokens_seen": 1081914368
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003394984954864594,
      "loss": 2.8413,
      "theoretical_loss": 3.622128939496834,
      "tokens_seen": 1081979904
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033948846539618853,
      "loss": 2.7511,
      "theoretical_loss": 3.6221083642402023,
      "tokens_seen": 1082045440
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033947843530591777,
      "loss": 2.7974,
      "theoretical_loss": 3.6220877905786155,
      "tokens_seen": 1082110976
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033946840521564695,
      "loss": 2.7031,
      "theoretical_loss": 3.6220672185118525,
      "tokens_seen": 1082176512
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033945837512537613,
      "loss": 2.4711,
      "theoretical_loss": 3.622046648039694,
      "tokens_seen": 1082242048
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003394483450351053,
      "loss": 2.6463,
      "theoretical_loss": 3.6220260791619197,
      "tokens_seen": 1082307584
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003394383149448345,
      "loss": 2.7821,
      "theoretical_loss": 3.622005511878309,
      "tokens_seen": 1082373120
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003394282848545637,
      "loss": 2.7633,
      "theoretical_loss": 3.6219849461886424,
      "tokens_seen": 1082438656
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003394182547642929,
      "loss": 2.7907,
      "theoretical_loss": 3.621964382092699,
      "tokens_seen": 1082504192
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033940822467402204,
      "loss": 2.5404,
      "theoretical_loss": 3.62194381959026,
      "tokens_seen": 1082569728
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033939819458375127,
      "loss": 2.6696,
      "theoretical_loss": 3.6219232586811048,
      "tokens_seen": 1082635264
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003393881644934804,
      "loss": 2.72,
      "theoretical_loss": 3.6219026993650134,
      "tokens_seen": 1082700800
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033937813440320963,
      "loss": 2.7538,
      "theoretical_loss": 3.621882141641766,
      "tokens_seen": 1082766336
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003393681043129388,
      "loss": 2.7411,
      "theoretical_loss": 3.6218615855111436,
      "tokens_seen": 1082831872
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.000339358074222668,
      "loss": 2.7286,
      "theoretical_loss": 3.6218410309729254,
      "tokens_seen": 1082897408
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003393480441323972,
      "loss": 3.0193,
      "theoretical_loss": 3.6218204780268914,
      "tokens_seen": 1082962944
    },
    {
      "epoch": 3.06,
      "objective/train/docs_used": 1225756,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9983677864074707,
      "objective/train/theoretical_loss": 3.621815340039136,
      "objective/train/tokens_used": 1103439328,
      "theoretical_loss": 3.621815340039136,
      "tokens_seen": 1082979328
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003393380140421264,
      "loss": 2.7483,
      "theoretical_loss": 3.6217999266728236,
      "tokens_seen": 1083028480
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003393279839518556,
      "loss": 2.7398,
      "theoretical_loss": 3.621779376910501,
      "tokens_seen": 1083094016
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003393179538615848,
      "loss": 2.7203,
      "theoretical_loss": 3.6217588287397042,
      "tokens_seen": 1083159552
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033930792377131396,
      "loss": 2.6191,
      "theoretical_loss": 3.621738282160214,
      "tokens_seen": 1083225088
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033929789368104314,
      "loss": 2.97,
      "theoretical_loss": 3.6217177371718106,
      "tokens_seen": 1083290624
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033928786359077237,
      "loss": 2.6384,
      "theoretical_loss": 3.621697193774275,
      "tokens_seen": 1083356160
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003392778335005015,
      "loss": 2.5806,
      "theoretical_loss": 3.6216766519673866,
      "tokens_seen": 1083421696
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033926780341023073,
      "loss": 2.8141,
      "theoretical_loss": 3.621656111750928,
      "tokens_seen": 1083487232
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033925777331995986,
      "loss": 2.7713,
      "theoretical_loss": 3.621635573124678,
      "tokens_seen": 1083552768
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003392477432296891,
      "loss": 2.7386,
      "theoretical_loss": 3.6216150360884183,
      "tokens_seen": 1083618304
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003392377131394183,
      "loss": 2.8081,
      "theoretical_loss": 3.6215945006419297,
      "tokens_seen": 1083683840
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033922768304914746,
      "loss": 2.8118,
      "theoretical_loss": 3.621573966784992,
      "tokens_seen": 1083749376
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033921765295887664,
      "loss": 3.0737,
      "theoretical_loss": 3.621553434517388,
      "tokens_seen": 1083814912
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003392076228686058,
      "loss": 2.7613,
      "theoretical_loss": 3.6215329038388964,
      "tokens_seen": 1083880448
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.000339197592778335,
      "loss": 2.6075,
      "theoretical_loss": 3.621512374749299,
      "tokens_seen": 1083945984
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033918756268806424,
      "loss": 2.7083,
      "theoretical_loss": 3.621491847248378,
      "tokens_seen": 1084011520
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033917753259779336,
      "loss": 2.7446,
      "theoretical_loss": 3.621471321335912,
      "tokens_seen": 1084077056
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003391675025075226,
      "loss": 2.9731,
      "theoretical_loss": 3.6214507970116845,
      "tokens_seen": 1084142592
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003391574724172518,
      "loss": 2.687,
      "theoretical_loss": 3.6214302742754754,
      "tokens_seen": 1084208128
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033914744232698096,
      "loss": 2.7052,
      "theoretical_loss": 3.6214097531270655,
      "tokens_seen": 1084273664
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033913741223671014,
      "loss": 2.7163,
      "theoretical_loss": 3.6213892335662363,
      "tokens_seen": 1084339200
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003391273821464393,
      "loss": 2.7634,
      "theoretical_loss": 3.6213687155927694,
      "tokens_seen": 1084404736
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003391173520561685,
      "loss": 2.651,
      "theoretical_loss": 3.6213481992064462,
      "tokens_seen": 1084470272
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033910732196589774,
      "loss": 2.7872,
      "theoretical_loss": 3.6213276844070474,
      "tokens_seen": 1084535808
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033909729187562687,
      "loss": 2.7807,
      "theoretical_loss": 3.6213071711943545,
      "tokens_seen": 1084601344
    },
    {
      "epoch": 3.06,
      "objective/train/docs_used": 1230848,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.071620225906372,
      "objective/train/theoretical_loss": 3.621302043139079,
      "objective/train/tokens_used": 1105077728,
      "theoretical_loss": 3.621302043139079,
      "tokens_seen": 1084617728
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003390872617853561,
      "loss": 2.2767,
      "theoretical_loss": 3.6212866595681494,
      "tokens_seen": 1084666880
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033907723169508523,
      "loss": 2.7164,
      "theoretical_loss": 3.6212661495282132,
      "tokens_seen": 1084732416
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033906720160481446,
      "loss": 2.5967,
      "theoretical_loss": 3.621245641074328,
      "tokens_seen": 1084797952
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033905717151454365,
      "loss": 2.7439,
      "theoretical_loss": 3.6212251342062736,
      "tokens_seen": 1084863488
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033904714142427283,
      "loss": 2.8463,
      "theoretical_loss": 3.6212046289238335,
      "tokens_seen": 1084929024
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.000339037111334002,
      "loss": 2.9693,
      "theoretical_loss": 3.6211841252267885,
      "tokens_seen": 1084994560
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033902708124373124,
      "loss": 2.7778,
      "theoretical_loss": 3.6211636231149202,
      "tokens_seen": 1085060096
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033901705115346037,
      "loss": 2.9191,
      "theoretical_loss": 3.621143122588011,
      "tokens_seen": 1085125632
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003390070210631896,
      "loss": 2.738,
      "theoretical_loss": 3.6211226236458414,
      "tokens_seen": 1085191168
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033899699097291873,
      "loss": 2.6476,
      "theoretical_loss": 3.6211021262881946,
      "tokens_seen": 1085256704
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033898696088264797,
      "loss": 2.8807,
      "theoretical_loss": 3.621081630514851,
      "tokens_seen": 1085322240
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033897693079237715,
      "loss": 2.5783,
      "theoretical_loss": 3.6210611363255945,
      "tokens_seen": 1085387776
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033896690070210633,
      "loss": 2.8129,
      "theoretical_loss": 3.621040643720205,
      "tokens_seen": 1085453312
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003389568706118355,
      "loss": 2.9223,
      "theoretical_loss": 3.621020152698465,
      "tokens_seen": 1085518848
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003389468405215647,
      "loss": 2.8439,
      "theoretical_loss": 3.6209996632601573,
      "tokens_seen": 1085584384
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003389368104312939,
      "loss": 2.4224,
      "theoretical_loss": 3.620979175405064,
      "tokens_seen": 1085649920
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003389267803410231,
      "loss": 2.8556,
      "theoretical_loss": 3.6209586891329657,
      "tokens_seen": 1085715456
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033891675025075224,
      "loss": 2.9056,
      "theoretical_loss": 3.620938204443646,
      "tokens_seen": 1085780992
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033890672016048147,
      "loss": 2.9773,
      "theoretical_loss": 3.6209177213368866,
      "tokens_seen": 1085846528
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003388966900702106,
      "loss": 2.5649,
      "theoretical_loss": 3.6208972398124697,
      "tokens_seen": 1085912064
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033888665997993983,
      "loss": 2.7481,
      "theoretical_loss": 3.6208767598701774,
      "tokens_seen": 1085977600
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.000338876629889669,
      "loss": 2.7252,
      "theoretical_loss": 3.6208562815097927,
      "tokens_seen": 1086043136
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003388665997993982,
      "loss": 2.7418,
      "theoretical_loss": 3.6208358047310973,
      "tokens_seen": 1086108672
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003388565697091274,
      "loss": 2.7919,
      "theoretical_loss": 3.6208153295338734,
      "tokens_seen": 1086174208
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003388465396188566,
      "loss": 2.8255,
      "theoretical_loss": 3.6207948559179046,
      "tokens_seen": 1086239744
    },
    {
      "epoch": 3.06,
      "objective/train/docs_used": 1232129,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.1969611644744873,
      "objective/train/theoretical_loss": 3.6207897377609575,
      "objective/train/tokens_used": 1106716128,
      "theoretical_loss": 3.6207897377609575,
      "tokens_seen": 1086256128
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033883650952858574,
      "loss": 2.8786,
      "theoretical_loss": 3.6207743838829725,
      "tokens_seen": 1086305280
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.000338826479438315,
      "loss": 2.6704,
      "theoretical_loss": 3.6207539134288593,
      "tokens_seen": 1086370816
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003388164493480441,
      "loss": 2.6908,
      "theoretical_loss": 3.6207334445553485,
      "tokens_seen": 1086436352
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033880641925777334,
      "loss": 2.5391,
      "theoretical_loss": 3.6207129772622224,
      "tokens_seen": 1086501888
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003387963891675025,
      "loss": 2.8369,
      "theoretical_loss": 3.6206925115492634,
      "tokens_seen": 1086567424
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003387863590772317,
      "loss": 2.7543,
      "theoretical_loss": 3.620672047416254,
      "tokens_seen": 1086632960
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003387763289869609,
      "loss": 2.7815,
      "theoretical_loss": 3.6206515848629777,
      "tokens_seen": 1086698496
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033876629889669006,
      "loss": 2.6934,
      "theoretical_loss": 3.6206311238892166,
      "tokens_seen": 1086764032
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033875626880641924,
      "loss": 2.7152,
      "theoretical_loss": 3.6206106644947544,
      "tokens_seen": 1086829568
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003387462387161485,
      "loss": 3.0106,
      "theoretical_loss": 3.620590206679373,
      "tokens_seen": 1086895104
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003387362086258776,
      "loss": 2.9409,
      "theoretical_loss": 3.620569750442856,
      "tokens_seen": 1086960640
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033872617853560684,
      "loss": 2.7743,
      "theoretical_loss": 3.6205492957849863,
      "tokens_seen": 1087026176
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033871614844533597,
      "loss": 2.5937,
      "theoretical_loss": 3.620528842705546,
      "tokens_seen": 1087091712
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003387061183550652,
      "loss": 2.9943,
      "theoretical_loss": 3.6205083912043197,
      "tokens_seen": 1087157248
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003386960882647944,
      "loss": 2.6698,
      "theoretical_loss": 3.620487941281089,
      "tokens_seen": 1087222784
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033868605817452356,
      "loss": 2.5315,
      "theoretical_loss": 3.6204674929356377,
      "tokens_seen": 1087288320
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033867602808425275,
      "loss": 2.8809,
      "theoretical_loss": 3.620447046167749,
      "tokens_seen": 1087353856
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.000338665997993982,
      "loss": 2.6834,
      "theoretical_loss": 3.620426600977207,
      "tokens_seen": 1087419392
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003386559679037111,
      "loss": 2.7171,
      "theoretical_loss": 3.620406157363793,
      "tokens_seen": 1087484928
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033864593781344034,
      "loss": 2.8595,
      "theoretical_loss": 3.6203857153272923,
      "tokens_seen": 1087550464
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033863590772316947,
      "loss": 2.8387,
      "theoretical_loss": 3.6203652748674866,
      "tokens_seen": 1087616000
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003386258776328987,
      "loss": 2.9063,
      "theoretical_loss": 3.6203448359841603,
      "tokens_seen": 1087681536
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003386158475426279,
      "loss": 2.7254,
      "theoretical_loss": 3.6203243986770968,
      "tokens_seen": 1087747072
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033860581745235707,
      "loss": 2.8621,
      "theoretical_loss": 3.620303962946079,
      "tokens_seen": 1087812608
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033859578736208625,
      "loss": 2.8689,
      "theoretical_loss": 3.6202835287908908,
      "tokens_seen": 1087878144
    },
    {
      "epoch": 3.06,
      "objective/train/docs_used": 1232782,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.946256637573242,
      "objective/train/theoretical_loss": 3.620278420498292,
      "objective/train/tokens_used": 1108354528,
      "theoretical_loss": 3.620278420498292,
      "tokens_seen": 1087894528
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033858575727181543,
      "loss": 2.9374,
      "theoretical_loss": 3.620263096211316,
      "tokens_seen": 1087943680
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033857572718154467,
      "loss": 2.9483,
      "theoretical_loss": 3.6202426652071376,
      "tokens_seen": 1088009216
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033856569709127385,
      "loss": 2.8696,
      "theoretical_loss": 3.6202222357781397,
      "tokens_seen": 1088074752
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033855566700100303,
      "loss": 2.6028,
      "theoretical_loss": 3.620201807924106,
      "tokens_seen": 1088140288
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003385456369107322,
      "loss": 2.7729,
      "theoretical_loss": 3.62018138164482,
      "tokens_seen": 1088205824
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033853560682046144,
      "loss": 3.0473,
      "theoretical_loss": 3.6201609569400657,
      "tokens_seen": 1088271360
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033852557673019057,
      "loss": 2.5592,
      "theoretical_loss": 3.620140533809627,
      "tokens_seen": 1088336896
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003385155466399198,
      "loss": 2.6389,
      "theoretical_loss": 3.620120112253287,
      "tokens_seen": 1088402432
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033850551654964893,
      "loss": 2.6717,
      "theoretical_loss": 3.6200996922708306,
      "tokens_seen": 1088467968
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033849548645937817,
      "loss": 2.7892,
      "theoretical_loss": 3.620079273862041,
      "tokens_seen": 1088533504
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033848545636910735,
      "loss": 2.7012,
      "theoretical_loss": 3.620058857026703,
      "tokens_seen": 1088599040
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033847542627883653,
      "loss": 2.4863,
      "theoretical_loss": 3.6200384417646,
      "tokens_seen": 1088664576
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003384653961885657,
      "loss": 2.754,
      "theoretical_loss": 3.6200180280755165,
      "tokens_seen": 1088730112
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003384553660982949,
      "loss": 2.6993,
      "theoretical_loss": 3.619997615959236,
      "tokens_seen": 1088795648
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003384453360080241,
      "loss": 2.6705,
      "theoretical_loss": 3.6199772054155432,
      "tokens_seen": 1088861184
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003384353059177533,
      "loss": 2.8679,
      "theoretical_loss": 3.619956796444222,
      "tokens_seen": 1088926720
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033842527582748244,
      "loss": 2.7584,
      "theoretical_loss": 3.619936389045057,
      "tokens_seen": 1088992256
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033841524573721167,
      "loss": 2.7861,
      "theoretical_loss": 3.6199159832178323,
      "tokens_seen": 1089057792
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003384052156469408,
      "loss": 2.8264,
      "theoretical_loss": 3.619895578962332,
      "tokens_seen": 1089123328
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033839518555667003,
      "loss": 2.8653,
      "theoretical_loss": 3.6198751762783408,
      "tokens_seen": 1089188864
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003383851554663992,
      "loss": 2.864,
      "theoretical_loss": 3.619854775165643,
      "tokens_seen": 1089254400
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003383751253761284,
      "loss": 2.793,
      "theoretical_loss": 3.6198343756240225,
      "tokens_seen": 1089319936
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003383650952858576,
      "loss": 2.7411,
      "theoretical_loss": 3.6198139776532656,
      "tokens_seen": 1089385472
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003383550651955868,
      "loss": 2.8531,
      "theoretical_loss": 3.6197935812531545,
      "tokens_seen": 1089451008
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033834503510531594,
      "loss": 2.6371,
      "theoretical_loss": 3.6197731864234752,
      "tokens_seen": 1089516544
    },
    {
      "epoch": 3.06,
      "objective/train/docs_used": 1234372,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.325249433517456,
      "objective/train/theoretical_loss": 3.6197680879614103,
      "objective/train/tokens_used": 1109992928,
      "theoretical_loss": 3.6197680879614103,
      "tokens_seen": 1089532928
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003383350050150452,
      "loss": 2.4403,
      "theoretical_loss": 3.6197527931640123,
      "tokens_seen": 1089582080
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003383249749247743,
      "loss": 2.7791,
      "theoretical_loss": 3.6197324014745504,
      "tokens_seen": 1089647616
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033831494483450354,
      "loss": 2.9964,
      "theoretical_loss": 3.6197120113548733,
      "tokens_seen": 1089713152
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003383049147442327,
      "loss": 2.7922,
      "theoretical_loss": 3.619691622804767,
      "tokens_seen": 1089778688
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003382948846539619,
      "loss": 2.7788,
      "theoretical_loss": 3.6196712358240157,
      "tokens_seen": 1089844224
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003382848545636911,
      "loss": 2.8688,
      "theoretical_loss": 3.619650850412404,
      "tokens_seen": 1089909760
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033827482447342026,
      "loss": 2.7936,
      "theoretical_loss": 3.6196304665697174,
      "tokens_seen": 1089975296
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033826479438314944,
      "loss": 2.8542,
      "theoretical_loss": 3.619610084295741,
      "tokens_seen": 1090040832
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003382547642928787,
      "loss": 2.6332,
      "theoretical_loss": 3.6195897035902584,
      "tokens_seen": 1090106368
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003382447342026078,
      "loss": 2.6517,
      "theoretical_loss": 3.619569324453056,
      "tokens_seen": 1090171904
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033823470411233704,
      "loss": 2.6189,
      "theoretical_loss": 3.6195489468839184,
      "tokens_seen": 1090237440
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033822467402206617,
      "loss": 2.8172,
      "theoretical_loss": 3.619528570882631,
      "tokens_seen": 1090302976
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003382146439317954,
      "loss": 2.8496,
      "theoretical_loss": 3.6195081964489777,
      "tokens_seen": 1090368512
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003382046138415246,
      "loss": 2.9001,
      "theoretical_loss": 3.6194878235827455,
      "tokens_seen": 1090434048
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033819458375125376,
      "loss": 3.004,
      "theoretical_loss": 3.6194674522837182,
      "tokens_seen": 1090499584
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033818455366098295,
      "loss": 2.5051,
      "theoretical_loss": 3.6194470825516816,
      "tokens_seen": 1090565120
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003381745235707122,
      "loss": 2.6448,
      "theoretical_loss": 3.619426714386421,
      "tokens_seen": 1090630656
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003381644934804413,
      "loss": 2.7133,
      "theoretical_loss": 3.619406347787722,
      "tokens_seen": 1090696192
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033815446339017054,
      "loss": 2.8944,
      "theoretical_loss": 3.61938598275537,
      "tokens_seen": 1090761728
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033814443329989967,
      "loss": 2.7101,
      "theoretical_loss": 3.6193656192891495,
      "tokens_seen": 1090827264
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003381344032096289,
      "loss": 2.7511,
      "theoretical_loss": 3.619345257388847,
      "tokens_seen": 1090892800
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003381243731193581,
      "loss": 2.7677,
      "theoretical_loss": 3.6193248970542475,
      "tokens_seen": 1090958336
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033811434302908727,
      "loss": 2.7238,
      "theoretical_loss": 3.619304538285137,
      "tokens_seen": 1091023872
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033810431293881645,
      "loss": 2.8521,
      "theoretical_loss": 3.6192841810813006,
      "tokens_seen": 1091089408
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033809428284854563,
      "loss": 2.5637,
      "theoretical_loss": 3.6192638254425242,
      "tokens_seen": 1091154944
    },
    {
      "epoch": 3.06,
      "objective/train/docs_used": 1235119,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5149593353271484,
      "objective/train/theoretical_loss": 3.619258736777346,
      "objective/train/tokens_used": 1111631328,
      "theoretical_loss": 3.619258736777346,
      "tokens_seen": 1091171328
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003380842527582748,
      "loss": 2.4153,
      "theoretical_loss": 3.619243471368594,
      "tokens_seen": 1091220480
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033807422266800405,
      "loss": 2.6798,
      "theoretical_loss": 3.6192231188592947,
      "tokens_seen": 1091286016
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003380641925777332,
      "loss": 2.8856,
      "theoretical_loss": 3.6192027679144125,
      "tokens_seen": 1091351552
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003380541624874624,
      "loss": 2.6799,
      "theoretical_loss": 3.619182418533734,
      "tokens_seen": 1091417088
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033804413239719154,
      "loss": 2.7602,
      "theoretical_loss": 3.619162070717044,
      "tokens_seen": 1091482624
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033803410230692077,
      "loss": 2.7477,
      "theoretical_loss": 3.6191417244641286,
      "tokens_seen": 1091548160
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033802407221664995,
      "loss": 2.757,
      "theoretical_loss": 3.619121379774774,
      "tokens_seen": 1091613696
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033801404212637913,
      "loss": 2.7874,
      "theoretical_loss": 3.6191010366487664,
      "tokens_seen": 1091679232
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003380040120361083,
      "loss": 2.8289,
      "theoretical_loss": 3.619080695085891,
      "tokens_seen": 1091744768
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033799398194583755,
      "loss": 2.8967,
      "theoretical_loss": 3.619060355085934,
      "tokens_seen": 1091810304
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003379839518555667,
      "loss": 2.6525,
      "theoretical_loss": 3.6190400166486834,
      "tokens_seen": 1091875840
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003379739217652959,
      "loss": 2.4763,
      "theoretical_loss": 3.6190196797739227,
      "tokens_seen": 1091941376
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033796389167502504,
      "loss": 2.6815,
      "theoretical_loss": 3.6189993444614394,
      "tokens_seen": 1092006912
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003379538615847543,
      "loss": 2.6668,
      "theoretical_loss": 3.61897901071102,
      "tokens_seen": 1092072448
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033794383149448346,
      "loss": 2.7857,
      "theoretical_loss": 3.61895867852245,
      "tokens_seen": 1092137984
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033793380140421264,
      "loss": 2.7561,
      "theoretical_loss": 3.6189383478955164,
      "tokens_seen": 1092203520
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003379237713139418,
      "loss": 2.8636,
      "theoretical_loss": 3.618918018830005,
      "tokens_seen": 1092269056
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.000337913741223671,
      "loss": 2.7024,
      "theoretical_loss": 3.6188976913257025,
      "tokens_seen": 1092334592
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003379037111334002,
      "loss": 2.6201,
      "theoretical_loss": 3.618877365382396,
      "tokens_seen": 1092400128
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003378936810431294,
      "loss": 2.7568,
      "theoretical_loss": 3.6188570409998704,
      "tokens_seen": 1092465664
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033788365095285854,
      "loss": 2.8051,
      "theoretical_loss": 3.618836718177913,
      "tokens_seen": 1092531200
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003378736208625878,
      "loss": 2.7044,
      "theoretical_loss": 3.618816396916311,
      "tokens_seen": 1092596736
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003378635907723169,
      "loss": 2.8435,
      "theoretical_loss": 3.6187960772148506,
      "tokens_seen": 1092662272
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033785356068204614,
      "loss": 2.6148,
      "theoretical_loss": 3.6187757590733183,
      "tokens_seen": 1092727808
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003378435305917753,
      "loss": 2.8435,
      "theoretical_loss": 3.6187554424915005,
      "tokens_seen": 1092793344
    },
    {
      "epoch": 3.06,
      "objective/train/docs_used": 1236502,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.975677490234375,
      "objective/train/theoretical_loss": 3.6187503635897267,
      "objective/train/tokens_used": 1113269728,
      "theoretical_loss": 3.6187503635897267,
      "tokens_seen": 1092809728
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003378335005015045,
      "loss": 2.84,
      "theoretical_loss": 3.6187351274691846,
      "tokens_seen": 1092858880
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033782347041123374,
      "loss": 2.7963,
      "theoretical_loss": 3.6187148140061574,
      "tokens_seen": 1092924416
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003378134403209629,
      "loss": 2.7715,
      "theoretical_loss": 3.618694502102205,
      "tokens_seen": 1092989952
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003378034102306921,
      "loss": 2.7411,
      "theoretical_loss": 3.618674191757115,
      "tokens_seen": 1093055488
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003377933801404213,
      "loss": 2.8463,
      "theoretical_loss": 3.618653882970674,
      "tokens_seen": 1093121024
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033778335005015046,
      "loss": 2.7457,
      "theoretical_loss": 3.6186335757426686,
      "tokens_seen": 1093186560
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033777331995987964,
      "loss": 2.7479,
      "theoretical_loss": 3.6186132700728866,
      "tokens_seen": 1093252096
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003377632898696089,
      "loss": 2.8089,
      "theoretical_loss": 3.6185929659611142,
      "tokens_seen": 1093317632
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.000337753259779338,
      "loss": 2.8787,
      "theoretical_loss": 3.6185726634071393,
      "tokens_seen": 1093383168
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033774322968906724,
      "loss": 2.6973,
      "theoretical_loss": 3.6185523624107487,
      "tokens_seen": 1093448704
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033773319959879637,
      "loss": 2.8744,
      "theoretical_loss": 3.6185320629717292,
      "tokens_seen": 1093514240
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003377231695085256,
      "loss": 2.5628,
      "theoretical_loss": 3.6185117650898677,
      "tokens_seen": 1093579776
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003377131394182548,
      "loss": 2.6778,
      "theoretical_loss": 3.6184914687649528,
      "tokens_seen": 1093645312
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033770310932798396,
      "loss": 2.8106,
      "theoretical_loss": 3.6184711739967708,
      "tokens_seen": 1093710848
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033769307923771315,
      "loss": 2.6888,
      "theoretical_loss": 3.618450880785109,
      "tokens_seen": 1093776384
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003376830491474424,
      "loss": 2.8349,
      "theoretical_loss": 3.618430589129755,
      "tokens_seen": 1093841920
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003376730190571715,
      "loss": 2.8202,
      "theoretical_loss": 3.6184102990304963,
      "tokens_seen": 1093907456
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033766298896690074,
      "loss": 2.6311,
      "theoretical_loss": 3.6183900104871203,
      "tokens_seen": 1093972992
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033765295887662987,
      "loss": 2.8867,
      "theoretical_loss": 3.6183697234994145,
      "tokens_seen": 1094038528
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003376429287863591,
      "loss": 2.5946,
      "theoretical_loss": 3.6183494380671664,
      "tokens_seen": 1094104064
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003376328986960883,
      "loss": 2.5993,
      "theoretical_loss": 3.618329154190163,
      "tokens_seen": 1094169600
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033762286860581747,
      "loss": 2.6903,
      "theoretical_loss": 3.618308871868193,
      "tokens_seen": 1094235136
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033761283851554665,
      "loss": 2.8344,
      "theoretical_loss": 3.6182885911010434,
      "tokens_seen": 1094300672
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033760280842527583,
      "loss": 2.6523,
      "theoretical_loss": 3.618268311888502,
      "tokens_seen": 1094366208
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.000337592778335005,
      "loss": 2.6933,
      "theoretical_loss": 3.6182480342303567,
      "tokens_seen": 1094431744
    },
    {
      "epoch": 3.06,
      "objective/train/docs_used": 1237024,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6148250102996826,
      "objective/train/theoretical_loss": 3.6182429650586703,
      "objective/train/tokens_used": 1114908128,
      "theoretical_loss": 3.6182429650586703,
      "tokens_seen": 1094448128
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033758274824473425,
      "loss": 2.6435,
      "theoretical_loss": 3.6182277581263955,
      "tokens_seen": 1094497280
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003375727181544634,
      "loss": 2.6805,
      "theoretical_loss": 3.6182074835764055,
      "tokens_seen": 1094562816
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003375626880641926,
      "loss": 2.5208,
      "theoretical_loss": 3.6181872105801753,
      "tokens_seen": 1094628352
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033755265797392174,
      "loss": 2.7542,
      "theoretical_loss": 3.618166939137492,
      "tokens_seen": 1094693888
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033754262788365097,
      "loss": 2.6371,
      "theoretical_loss": 3.6181466692481443,
      "tokens_seen": 1094759424
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033753259779338015,
      "loss": 2.8191,
      "theoretical_loss": 3.61812640091192,
      "tokens_seen": 1094824960
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033752256770310933,
      "loss": 2.7634,
      "theoretical_loss": 3.6181061341286074,
      "tokens_seen": 1094890496
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003375125376128385,
      "loss": 2.7474,
      "theoretical_loss": 3.6180858688979938,
      "tokens_seen": 1094956032
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033750250752256775,
      "loss": 2.5913,
      "theoretical_loss": 3.6180656052198676,
      "tokens_seen": 1095021568
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003374924774322969,
      "loss": 2.8893,
      "theoretical_loss": 3.6180453430940176,
      "tokens_seen": 1095087104
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003374824473420261,
      "loss": 2.675,
      "theoretical_loss": 3.618025082520232,
      "tokens_seen": 1095152640
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033747241725175524,
      "loss": 2.7979,
      "theoretical_loss": 3.618004823498297,
      "tokens_seen": 1095218176
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003374623871614845,
      "loss": 2.8325,
      "theoretical_loss": 3.617984566028004,
      "tokens_seen": 1095283712
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033745235707121366,
      "loss": 2.6937,
      "theoretical_loss": 3.617964310109139,
      "tokens_seen": 1095349248
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033744232698094284,
      "loss": 2.777,
      "theoretical_loss": 3.617944055741492,
      "tokens_seen": 1095414784
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.000337432296890672,
      "loss": 2.5714,
      "theoretical_loss": 3.6179238029248495,
      "tokens_seen": 1095480320
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003374222668004012,
      "loss": 2.7678,
      "theoretical_loss": 3.6179035516590017,
      "tokens_seen": 1095545856
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003374122367101304,
      "loss": 2.6847,
      "theoretical_loss": 3.617883301943736,
      "tokens_seen": 1095611392
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003374022066198596,
      "loss": 2.733,
      "theoretical_loss": 3.6178630537788417,
      "tokens_seen": 1095676928
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033739217652958874,
      "loss": 2.5637,
      "theoretical_loss": 3.6178428071641076,
      "tokens_seen": 1095742464
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.000337382146439318,
      "loss": 2.716,
      "theoretical_loss": 3.6178225620993207,
      "tokens_seen": 1095808000
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003373721163490471,
      "loss": 2.7915,
      "theoretical_loss": 3.617802318584271,
      "tokens_seen": 1095873536
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033736208625877634,
      "loss": 2.8944,
      "theoretical_loss": 3.6177820766187474,
      "tokens_seen": 1095939072
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003373520561685055,
      "loss": 2.7548,
      "theoretical_loss": 3.617761836202538,
      "tokens_seen": 1096004608
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003373420260782347,
      "loss": 2.6252,
      "theoretical_loss": 3.6177415973354314,
      "tokens_seen": 1096070144
    },
    {
      "epoch": 3.06,
      "objective/train/docs_used": 1237677,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6507112979888916,
      "objective/train/theoretical_loss": 3.6177365378606776,
      "objective/train/tokens_used": 1116546528,
      "theoretical_loss": 3.6177365378606776,
      "tokens_seen": 1096086528
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003373319959879639,
      "loss": 2.7237,
      "theoretical_loss": 3.6177213600172173,
      "tokens_seen": 1096135680
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003373219658976931,
      "loss": 2.6584,
      "theoretical_loss": 3.6177011242476835,
      "tokens_seen": 1096201216
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033731193580742225,
      "loss": 2.6034,
      "theoretical_loss": 3.6176808900266204,
      "tokens_seen": 1096266752
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003373019057171515,
      "loss": 2.7871,
      "theoretical_loss": 3.6176606573538153,
      "tokens_seen": 1096332288
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003372918756268806,
      "loss": 2.7767,
      "theoretical_loss": 3.617640426229058,
      "tokens_seen": 1096397824
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033728184553660984,
      "loss": 2.5391,
      "theoretical_loss": 3.6176201966521373,
      "tokens_seen": 1096463360
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.000337271815446339,
      "loss": 2.7906,
      "theoretical_loss": 3.6175999686228426,
      "tokens_seen": 1096528896
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003372617853560682,
      "loss": 2.8622,
      "theoretical_loss": 3.6175797421409626,
      "tokens_seen": 1096594432
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003372517552657974,
      "loss": 2.6529,
      "theoretical_loss": 3.6175595172062867,
      "tokens_seen": 1096659968
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033724172517552657,
      "loss": 2.5864,
      "theoretical_loss": 3.617539293818605,
      "tokens_seen": 1096725504
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033723169508525575,
      "loss": 2.6748,
      "theoretical_loss": 3.6175190719777044,
      "tokens_seen": 1096791040
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.000337221664994985,
      "loss": 2.7057,
      "theoretical_loss": 3.6174988516833766,
      "tokens_seen": 1096856576
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003372116349047141,
      "loss": 2.7422,
      "theoretical_loss": 3.617478632935409,
      "tokens_seen": 1096922112
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033720160481444335,
      "loss": 2.813,
      "theoretical_loss": 3.617458415733593,
      "tokens_seen": 1096987648
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003371915747241725,
      "loss": 2.785,
      "theoretical_loss": 3.617438200077716,
      "tokens_seen": 1097053184
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003371815446339017,
      "loss": 2.5565,
      "theoretical_loss": 3.617417985967569,
      "tokens_seen": 1097118720
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003371715145436309,
      "loss": 2.7586,
      "theoretical_loss": 3.6173977734029403,
      "tokens_seen": 1097184256
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033716148445336007,
      "loss": 2.7171,
      "theoretical_loss": 3.6173775623836204,
      "tokens_seen": 1097249792
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033715145436308925,
      "loss": 2.8573,
      "theoretical_loss": 3.6173573529093983,
      "tokens_seen": 1097315328
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003371414242728185,
      "loss": 2.4732,
      "theoretical_loss": 3.6173371449800635,
      "tokens_seen": 1097380864
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003371313941825476,
      "loss": 2.6875,
      "theoretical_loss": 3.6173169385954065,
      "tokens_seen": 1097446400
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033712136409227685,
      "loss": 2.6837,
      "theoretical_loss": 3.6172967337552153,
      "tokens_seen": 1097511936
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.000337111334002006,
      "loss": 2.6919,
      "theoretical_loss": 3.6172765304592813,
      "tokens_seen": 1097577472
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003371013039117352,
      "loss": 2.5633,
      "theoretical_loss": 3.617256328707394,
      "tokens_seen": 1097643008
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003370912738214644,
      "loss": 2.4857,
      "theoretical_loss": 3.617236128499343,
      "tokens_seen": 1097708544
    },
    {
      "epoch": 3.06,
      "objective/train/docs_used": 1238996,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.87542986869812,
      "objective/train/theoretical_loss": 3.6172310786885298,
      "objective/train/tokens_used": 1118184928,
      "theoretical_loss": 3.6172310786885298,
      "tokens_seen": 1097724928
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003370812437311936,
      "loss": 2.9281,
      "theoretical_loss": 3.6172159298349174,
      "tokens_seen": 1097774080
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003370712136409228,
      "loss": 2.849,
      "theoretical_loss": 3.617195732713908,
      "tokens_seen": 1097839616
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033706118355065194,
      "loss": 2.582,
      "theoretical_loss": 3.617175537136105,
      "tokens_seen": 1097905152
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033705115346038117,
      "loss": 2.6484,
      "theoretical_loss": 3.617155343101297,
      "tokens_seen": 1097970688
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033704112337011035,
      "loss": 2.8651,
      "theoretical_loss": 3.6171351506092755,
      "tokens_seen": 1098036224
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033703109327983953,
      "loss": 2.8084,
      "theoretical_loss": 3.61711495965983,
      "tokens_seen": 1098101760
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003370210631895687,
      "loss": 2.72,
      "theoretical_loss": 3.6170947702527503,
      "tokens_seen": 1098167296
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033701103309929795,
      "loss": 2.5411,
      "theoretical_loss": 3.6170745823878274,
      "tokens_seen": 1098232832
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003370010030090271,
      "loss": 2.634,
      "theoretical_loss": 3.6170543960648507,
      "tokens_seen": 1098298368
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003369909729187563,
      "loss": 2.6711,
      "theoretical_loss": 3.6170342112836105,
      "tokens_seen": 1098363904
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033698094282848544,
      "loss": 2.7764,
      "theoretical_loss": 3.6170140280438967,
      "tokens_seen": 1098429440
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003369709127382147,
      "loss": 2.6749,
      "theoretical_loss": 3.6169938463455003,
      "tokens_seen": 1098494976
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033696088264794386,
      "loss": 2.9666,
      "theoretical_loss": 3.616973666188212,
      "tokens_seen": 1098560512
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033695085255767304,
      "loss": 2.7339,
      "theoretical_loss": 3.6169534875718217,
      "tokens_seen": 1098626048
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003369408224674022,
      "loss": 2.6009,
      "theoretical_loss": 3.616933310496119,
      "tokens_seen": 1098691584
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003369307923771314,
      "loss": 2.4912,
      "theoretical_loss": 3.616913134960896,
      "tokens_seen": 1098757120
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003369207622868606,
      "loss": 2.8986,
      "theoretical_loss": 3.6168929609659424,
      "tokens_seen": 1098822656
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003369107321965898,
      "loss": 2.7956,
      "theoretical_loss": 3.6168727885110483,
      "tokens_seen": 1098888192
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033690070210631894,
      "loss": 2.6026,
      "theoretical_loss": 3.6168526175960047,
      "tokens_seen": 1098953728
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003368906720160482,
      "loss": 3.0197,
      "theoretical_loss": 3.6168324482206025,
      "tokens_seen": 1099019264
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003368806419257773,
      "loss": 2.8678,
      "theoretical_loss": 3.6168122803846323,
      "tokens_seen": 1099084800
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033687061183550654,
      "loss": 2.5931,
      "theoretical_loss": 3.6167921140878843,
      "tokens_seen": 1099150336
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003368605817452357,
      "loss": 2.5982,
      "theoretical_loss": 3.6167719493301496,
      "tokens_seen": 1099215872
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003368505516549649,
      "loss": 2.6875,
      "theoretical_loss": 3.6167517861112195,
      "tokens_seen": 1099281408
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003368405215646941,
      "loss": 2.6566,
      "theoretical_loss": 3.6167316244308845,
      "tokens_seen": 1099346944
    },
    {
      "epoch": 3.06,
      "objective/train/docs_used": 1239758,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.1782515048980713,
      "objective/train/theoretical_loss": 3.6167265842511815,
      "objective/train/tokens_used": 1119823328,
      "theoretical_loss": 3.6167265842511815,
      "tokens_seen": 1099363328
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003368304914744233,
      "loss": 2.6187,
      "theoretical_loss": 3.616711464288935,
      "tokens_seen": 1099412480
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033682046138415245,
      "loss": 2.8087,
      "theoretical_loss": 3.616691305685162,
      "tokens_seen": 1099478016
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003368104312938817,
      "loss": 2.7666,
      "theoretical_loss": 3.616671148619357,
      "tokens_seen": 1099543552
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003368004012036108,
      "loss": 2.8175,
      "theoretical_loss": 3.6166509930913113,
      "tokens_seen": 1099609088
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033679037111334004,
      "loss": 2.6504,
      "theoretical_loss": 3.6166308391008153,
      "tokens_seen": 1099674624
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003367803410230692,
      "loss": 2.7364,
      "theoretical_loss": 3.6166106866476593,
      "tokens_seen": 1099740160
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003367703109327984,
      "loss": 2.6164,
      "theoretical_loss": 3.6165905357316364,
      "tokens_seen": 1099805696
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003367602808425276,
      "loss": 2.7391,
      "theoretical_loss": 3.616570386352536,
      "tokens_seen": 1099871232
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033675025075225677,
      "loss": 2.539,
      "theoretical_loss": 3.6165502385101505,
      "tokens_seen": 1099936768
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033674022066198595,
      "loss": 2.6233,
      "theoretical_loss": 3.61653009220427,
      "tokens_seen": 1100002304
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003367301905717152,
      "loss": 2.7204,
      "theoretical_loss": 3.616509947434687,
      "tokens_seen": 1100067840
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003367201604814443,
      "loss": 2.6064,
      "theoretical_loss": 3.6164898042011924,
      "tokens_seen": 1100133376
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033671013039117355,
      "loss": 2.7523,
      "theoretical_loss": 3.616469662503577,
      "tokens_seen": 1100198912
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033670010030090273,
      "loss": 2.7206,
      "theoretical_loss": 3.616449522341633,
      "tokens_seen": 1100264448
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003366900702106319,
      "loss": 2.7963,
      "theoretical_loss": 3.6164293837151513,
      "tokens_seen": 1100329984
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003366800401203611,
      "loss": 2.5174,
      "theoretical_loss": 3.616409246623924,
      "tokens_seen": 1100395520
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033667001003009027,
      "loss": 2.8013,
      "theoretical_loss": 3.6163891110677415,
      "tokens_seen": 1100461056
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033665997993981945,
      "loss": 2.6539,
      "theoretical_loss": 3.6163689770463967,
      "tokens_seen": 1100526592
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003366499498495487,
      "loss": 2.6839,
      "theoretical_loss": 3.6163488445596808,
      "tokens_seen": 1100592128
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003366399197592778,
      "loss": 2.7934,
      "theoretical_loss": 3.616328713607385,
      "tokens_seen": 1100657664
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033662988966900705,
      "loss": 2.7716,
      "theoretical_loss": 3.6163085841893015,
      "tokens_seen": 1100723200
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003366198595787362,
      "loss": 2.7189,
      "theoretical_loss": 3.616288456305222,
      "tokens_seen": 1100788736
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003366098294884654,
      "loss": 2.5594,
      "theoretical_loss": 3.616268329954938,
      "tokens_seen": 1100854272
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003365997993981946,
      "loss": 2.6865,
      "theoretical_loss": 3.616248205138241,
      "tokens_seen": 1100919808
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003365897693079238,
      "loss": 2.812,
      "theoretical_loss": 3.616228081854924,
      "tokens_seen": 1100985344
    },
    {
      "epoch": 3.06,
      "objective/train/docs_used": 1241137,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8605268001556396,
      "objective/train/theoretical_loss": 3.6162230512736606,
      "objective/train/tokens_used": 1121461728,
      "theoretical_loss": 3.6162230512736606,
      "tokens_seen": 1101001728
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033657973921765295,
      "loss": 2.9682,
      "theoretical_loss": 3.616207960104778,
      "tokens_seen": 1101050880
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033656970912738214,
      "loss": 2.8834,
      "theoretical_loss": 3.616187839887595,
      "tokens_seen": 1101116416
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003365596790371113,
      "loss": 2.8656,
      "theoretical_loss": 3.6161677212031673,
      "tokens_seen": 1101181952
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033654964894684055,
      "loss": 2.6308,
      "theoretical_loss": 3.616147604051287,
      "tokens_seen": 1101247488
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003365396188565697,
      "loss": 2.7918,
      "theoretical_loss": 3.6161274884317454,
      "tokens_seen": 1101313024
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003365295887662989,
      "loss": 2.6168,
      "theoretical_loss": 3.6161073743443355,
      "tokens_seen": 1101378560
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003365195586760281,
      "loss": 2.7806,
      "theoretical_loss": 3.616087261788849,
      "tokens_seen": 1101444096
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003365095285857573,
      "loss": 2.6078,
      "theoretical_loss": 3.6160671507650783,
      "tokens_seen": 1101509632
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033649949849548646,
      "loss": 2.5558,
      "theoretical_loss": 3.6160470412728154,
      "tokens_seen": 1101575168
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033648946840521564,
      "loss": 3.0053,
      "theoretical_loss": 3.616026933311853,
      "tokens_seen": 1101640704
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003364794383149448,
      "loss": 2.5638,
      "theoretical_loss": 3.6160068268819825,
      "tokens_seen": 1101706240
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033646940822467406,
      "loss": 2.4866,
      "theoretical_loss": 3.6159867219829973,
      "tokens_seen": 1101771776
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003364593781344032,
      "loss": 2.546,
      "theoretical_loss": 3.615966618614689,
      "tokens_seen": 1101837312
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003364493480441324,
      "loss": 2.6268,
      "theoretical_loss": 3.6159465167768507,
      "tokens_seen": 1101902848
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033643931795386154,
      "loss": 2.7514,
      "theoretical_loss": 3.615926416469274,
      "tokens_seen": 1101968384
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003364292878635908,
      "loss": 2.7597,
      "theoretical_loss": 3.6159063176917527,
      "tokens_seen": 1102033920
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033641925777331996,
      "loss": 2.7634,
      "theoretical_loss": 3.615886220444078,
      "tokens_seen": 1102099456
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033640922768304914,
      "loss": 2.8582,
      "theoretical_loss": 3.6158661247260433,
      "tokens_seen": 1102164992
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003363991975927783,
      "loss": 2.7127,
      "theoretical_loss": 3.6158460305374405,
      "tokens_seen": 1102230528
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003363891675025075,
      "loss": 2.6674,
      "theoretical_loss": 3.6158259378780633,
      "tokens_seen": 1102296064
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003363791374122367,
      "loss": 2.9135,
      "theoretical_loss": 3.615805846747704,
      "tokens_seen": 1102361600
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003363691073219659,
      "loss": 2.7931,
      "theoretical_loss": 3.6157857571461554,
      "tokens_seen": 1102427136
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033635907723169505,
      "loss": 2.7545,
      "theoretical_loss": 3.61576566907321,
      "tokens_seen": 1102492672
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003363490471414243,
      "loss": 2.4555,
      "theoretical_loss": 3.6157455825286604,
      "tokens_seen": 1102558208
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033633901705115346,
      "loss": 2.8385,
      "theoretical_loss": 3.6157254975123,
      "tokens_seen": 1102623744
    },
    {
      "epoch": 3.06,
      "objective/train/docs_used": 1241790,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5041520595550537,
      "objective/train/theoretical_loss": 3.615720476496965,
      "objective/train/tokens_used": 1123100128,
      "theoretical_loss": 3.615720476496965,
      "tokens_seen": 1102640128
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033632898696088265,
      "loss": 2.6422,
      "theoretical_loss": 3.6157054140239215,
      "tokens_seen": 1102689280
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003363189568706119,
      "loss": 2.8461,
      "theoretical_loss": 3.615685332063318,
      "tokens_seen": 1102754816
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.000336308926780341,
      "loss": 2.71,
      "theoretical_loss": 3.6156652516302827,
      "tokens_seen": 1102820352
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033629889669007024,
      "loss": 2.789,
      "theoretical_loss": 3.615645172724608,
      "tokens_seen": 1102885888
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003362888665997994,
      "loss": 2.747,
      "theoretical_loss": 3.6156250953460876,
      "tokens_seen": 1102951424
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003362788365095286,
      "loss": 2.6397,
      "theoretical_loss": 3.615605019494514,
      "tokens_seen": 1103016960
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003362688064192578,
      "loss": 2.6001,
      "theoretical_loss": 3.6155849451696813,
      "tokens_seen": 1103082496
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033625877632898697,
      "loss": 2.4701,
      "theoretical_loss": 3.615564872371382,
      "tokens_seen": 1103148032
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033624874623871615,
      "loss": 2.6813,
      "theoretical_loss": 3.615544801099409,
      "tokens_seen": 1103213568
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003362387161484454,
      "loss": 2.7879,
      "theoretical_loss": 3.615524731353556,
      "tokens_seen": 1103279104
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003362286860581745,
      "loss": 2.6216,
      "theoretical_loss": 3.615504663133617,
      "tokens_seen": 1103344640
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033621865596790375,
      "loss": 2.383,
      "theoretical_loss": 3.615484596439384,
      "tokens_seen": 1103410176
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033620862587763293,
      "loss": 2.4897,
      "theoretical_loss": 3.6154645312706513,
      "tokens_seen": 1103475712
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003361985957873621,
      "loss": 2.7389,
      "theoretical_loss": 3.615444467627212,
      "tokens_seen": 1103541248
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003361885656970913,
      "loss": 2.7377,
      "theoretical_loss": 3.6154244055088602,
      "tokens_seen": 1103606784
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033617853560682047,
      "loss": 2.8105,
      "theoretical_loss": 3.6154043449153885,
      "tokens_seen": 1103672320
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033616850551654965,
      "loss": 2.6996,
      "theoretical_loss": 3.615384285846591,
      "tokens_seen": 1103737856
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003361584754262789,
      "loss": 2.4126,
      "theoretical_loss": 3.6153642283022616,
      "tokens_seen": 1103803392
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.000336148445336008,
      "loss": 2.6139,
      "theoretical_loss": 3.6153441722821933,
      "tokens_seen": 1103868928
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033613841524573725,
      "loss": 2.6361,
      "theoretical_loss": 3.6153241177861797,
      "tokens_seen": 1103934464
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003361283851554664,
      "loss": 2.4851,
      "theoretical_loss": 3.615304064814015,
      "tokens_seen": 1104000000
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003361183550651956,
      "loss": 2.6684,
      "theoretical_loss": 3.615284013365493,
      "tokens_seen": 1104065536
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003361083249749248,
      "loss": 2.7077,
      "theoretical_loss": 3.615263963440407,
      "tokens_seen": 1104131072
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.000336098294884654,
      "loss": 2.7662,
      "theoretical_loss": 3.615243915038551,
      "tokens_seen": 1104196608
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033608826479438316,
      "loss": 2.6451,
      "theoretical_loss": 3.61522386815972,
      "tokens_seen": 1104262144
    },
    {
      "epoch": 3.06,
      "objective/train/docs_used": 1242899,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7701282501220703,
      "objective/train/theoretical_loss": 3.61521885667796,
      "objective/train/tokens_used": 1124738528,
      "theoretical_loss": 3.61521885667796,
      "tokens_seen": 1104278528
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033607823470411234,
      "loss": 2.5712,
      "theoretical_loss": 3.6152038228037062,
      "tokens_seen": 1104327680
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003360682046138415,
      "loss": 2.7366,
      "theoretical_loss": 3.6151837789703043,
      "tokens_seen": 1104393216
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033605817452357075,
      "loss": 2.6724,
      "theoretical_loss": 3.6151637366593086,
      "tokens_seen": 1104458752
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003360481444332999,
      "loss": 2.6507,
      "theoretical_loss": 3.615143695870512,
      "tokens_seen": 1104524288
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003360381143430291,
      "loss": 2.8124,
      "theoretical_loss": 3.6151236566037106,
      "tokens_seen": 1104589824
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003360280842527583,
      "loss": 2.6883,
      "theoretical_loss": 3.6151036188586967,
      "tokens_seen": 1104655360
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.0003360180541624875,
      "loss": 2.6271,
      "theoretical_loss": 3.6150835826352656,
      "tokens_seen": 1104720896
    },
    {
      "epoch": 3.06,
      "learning_rate": 0.00033600802407221666,
      "loss": 2.679,
      "theoretical_loss": 3.6150635479332105,
      "tokens_seen": 1104786432
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033599799398194584,
      "loss": 2.6898,
      "theoretical_loss": 3.615043514752326,
      "tokens_seen": 1104851968
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.000335987963891675,
      "loss": 2.6468,
      "theoretical_loss": 3.615023483092407,
      "tokens_seen": 1104917504
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033597793380140426,
      "loss": 2.7967,
      "theoretical_loss": 3.6150034529532475,
      "tokens_seen": 1104983040
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003359679037111334,
      "loss": 2.8978,
      "theoretical_loss": 3.6149834243346417,
      "tokens_seen": 1105048576
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003359578736208626,
      "loss": 2.6337,
      "theoretical_loss": 3.6149633972363837,
      "tokens_seen": 1105114112
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033594784353059174,
      "loss": 2.6396,
      "theoretical_loss": 3.614943371658268,
      "tokens_seen": 1105179648
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.000335937813440321,
      "loss": 2.8301,
      "theoretical_loss": 3.61492334760009,
      "tokens_seen": 1105245184
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033592778335005016,
      "loss": 3.037,
      "theoretical_loss": 3.6149033250616434,
      "tokens_seen": 1105310720
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033591775325977934,
      "loss": 2.6608,
      "theoretical_loss": 3.6148833040427224,
      "tokens_seen": 1105376256
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003359077231695085,
      "loss": 2.8969,
      "theoretical_loss": 3.614863284543123,
      "tokens_seen": 1105441792
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003358976930792377,
      "loss": 2.7132,
      "theoretical_loss": 3.6148432665626387,
      "tokens_seen": 1105507328
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003358876629889669,
      "loss": 2.4359,
      "theoretical_loss": 3.6148232501010638,
      "tokens_seen": 1105572864
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003358776328986961,
      "loss": 2.9279,
      "theoretical_loss": 3.6148032351581945,
      "tokens_seen": 1105638400
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033586760280842525,
      "loss": 2.7633,
      "theoretical_loss": 3.6147832217338243,
      "tokens_seen": 1105703936
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003358575727181545,
      "loss": 2.7452,
      "theoretical_loss": 3.614763209827749,
      "tokens_seen": 1105769472
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033584754262788366,
      "loss": 2.6413,
      "theoretical_loss": 3.6147431994397623,
      "tokens_seen": 1105835008
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033583751253761285,
      "loss": 2.5344,
      "theoretical_loss": 3.61472319056966,
      "tokens_seen": 1105900544
    },
    {
      "epoch": 3.07,
      "objective/train/docs_used": 1243613,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9853389263153076,
      "objective/train/theoretical_loss": 3.6147181885892796,
      "objective/train/tokens_used": 1126376928,
      "theoretical_loss": 3.6147181885892796,
      "tokens_seen": 1105916928
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.000335827482447342,
      "loss": 2.7355,
      "theoretical_loss": 3.6147031832172365,
      "tokens_seen": 1105966080
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003358174523570712,
      "loss": 2.6175,
      "theoretical_loss": 3.614683177382287,
      "tokens_seen": 1106031616
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003358074222668004,
      "loss": 2.9476,
      "theoretical_loss": 3.6146631730646064,
      "tokens_seen": 1106097152
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003357973921765296,
      "loss": 2.7897,
      "theoretical_loss": 3.6146431702639896,
      "tokens_seen": 1106162688
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033578736208625875,
      "loss": 2.6465,
      "theoretical_loss": 3.6146231689802324,
      "tokens_seen": 1106228224
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.000335777331995988,
      "loss": 2.6784,
      "theoretical_loss": 3.6146031692131295,
      "tokens_seen": 1106293760
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003357673019057171,
      "loss": 2.7595,
      "theoretical_loss": 3.614583170962475,
      "tokens_seen": 1106359296
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033575727181544635,
      "loss": 2.5581,
      "theoretical_loss": 3.614563174228066,
      "tokens_seen": 1106424832
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033574724172517553,
      "loss": 2.7814,
      "theoretical_loss": 3.6145431790096962,
      "tokens_seen": 1106490368
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003357372116349047,
      "loss": 2.6407,
      "theoretical_loss": 3.6145231853071618,
      "tokens_seen": 1106555904
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003357271815446339,
      "loss": 2.7503,
      "theoretical_loss": 3.6145031931202576,
      "tokens_seen": 1106621440
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033571715145436313,
      "loss": 2.794,
      "theoretical_loss": 3.6144832024487794,
      "tokens_seen": 1106686976
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033570712136409225,
      "loss": 2.765,
      "theoretical_loss": 3.614463213292522,
      "tokens_seen": 1106752512
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003356970912738215,
      "loss": 2.7039,
      "theoretical_loss": 3.6144432256512817,
      "tokens_seen": 1106818048
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003356870611835506,
      "loss": 2.5687,
      "theoretical_loss": 3.6144232395248532,
      "tokens_seen": 1106883584
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033567703109327985,
      "loss": 2.7355,
      "theoretical_loss": 3.6144032549130323,
      "tokens_seen": 1106949120
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033566700100300903,
      "loss": 2.7997,
      "theoretical_loss": 3.6143832718156146,
      "tokens_seen": 1107014656
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003356569709127382,
      "loss": 2.6971,
      "theoretical_loss": 3.6143632902323954,
      "tokens_seen": 1107080192
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003356469408224674,
      "loss": 2.8226,
      "theoretical_loss": 3.6143433101631706,
      "tokens_seen": 1107145728
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003356369107321966,
      "loss": 2.458,
      "theoretical_loss": 3.614323331607736,
      "tokens_seen": 1107211264
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033562688064192576,
      "loss": 2.6713,
      "theoretical_loss": 3.6143033545658874,
      "tokens_seen": 1107276800
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.000335616850551655,
      "loss": 2.5748,
      "theoretical_loss": 3.61428337903742,
      "tokens_seen": 1107342336
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003356068204613841,
      "loss": 2.8478,
      "theoretical_loss": 3.61426340502213,
      "tokens_seen": 1107407872
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033559679037111336,
      "loss": 2.8482,
      "theoretical_loss": 3.614243432519813,
      "tokens_seen": 1107473408
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003355867602808425,
      "loss": 2.7905,
      "theoretical_loss": 3.6142234615302655,
      "tokens_seen": 1107538944
    },
    {
      "epoch": 3.07,
      "objective/train/docs_used": 1244863,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.85904860496521,
      "objective/train/theoretical_loss": 3.6142184690192245,
      "objective/train/tokens_used": 1128015328,
      "theoretical_loss": 3.6142184690192245,
      "tokens_seen": 1107555328
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003355767301905717,
      "loss": 2.8847,
      "theoretical_loss": 3.6142034920532824,
      "tokens_seen": 1107604480
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033556670010030095,
      "loss": 2.7116,
      "theoretical_loss": 3.6141835240886606,
      "tokens_seen": 1107670016
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003355566700100301,
      "loss": 2.6555,
      "theoretical_loss": 3.6141635576361955,
      "tokens_seen": 1107735552
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003355466399197593,
      "loss": 2.5725,
      "theoretical_loss": 3.6141435926956835,
      "tokens_seen": 1107801088
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003355366098294885,
      "loss": 2.6107,
      "theoretical_loss": 3.6141236292669205,
      "tokens_seen": 1107866624
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003355265797392177,
      "loss": 2.4844,
      "theoretical_loss": 3.6141036673497027,
      "tokens_seen": 1107932160
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033551654964894686,
      "loss": 2.7763,
      "theoretical_loss": 3.6140837069438256,
      "tokens_seen": 1107997696
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033550651955867604,
      "loss": 2.7287,
      "theoretical_loss": 3.6140637480490865,
      "tokens_seen": 1108063232
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003354964894684052,
      "loss": 2.6421,
      "theoretical_loss": 3.6140437906652805,
      "tokens_seen": 1108128768
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033548645937813446,
      "loss": 2.595,
      "theoretical_loss": 3.614023834792205,
      "tokens_seen": 1108194304
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003354764292878636,
      "loss": 2.7723,
      "theoretical_loss": 3.6140038804296557,
      "tokens_seen": 1108259840
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003354663991975928,
      "loss": 2.8769,
      "theoretical_loss": 3.6139839275774293,
      "tokens_seen": 1108325376
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033545636910732195,
      "loss": 2.587,
      "theoretical_loss": 3.613963976235321,
      "tokens_seen": 1108390912
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003354463390170512,
      "loss": 2.7087,
      "theoretical_loss": 3.613944026403129,
      "tokens_seen": 1108456448
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033543630892678036,
      "loss": 2.8548,
      "theoretical_loss": 3.6139240780806485,
      "tokens_seen": 1108521984
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033542627883650954,
      "loss": 2.5704,
      "theoretical_loss": 3.6139041312676765,
      "tokens_seen": 1108587520
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003354162487462387,
      "loss": 2.7445,
      "theoretical_loss": 3.61388418596401,
      "tokens_seen": 1108653056
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003354062186559679,
      "loss": 2.8618,
      "theoretical_loss": 3.613864242169444,
      "tokens_seen": 1108718592
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003353961885656971,
      "loss": 2.6855,
      "theoretical_loss": 3.6138442998837768,
      "tokens_seen": 1108784128
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003353861584754263,
      "loss": 2.51,
      "theoretical_loss": 3.613824359106805,
      "tokens_seen": 1108849664
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033537612838515545,
      "loss": 2.6772,
      "theoretical_loss": 3.6138044198383237,
      "tokens_seen": 1108915200
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003353660982948847,
      "loss": 2.6674,
      "theoretical_loss": 3.613784482078131,
      "tokens_seen": 1108980736
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033535606820461386,
      "loss": 2.6406,
      "theoretical_loss": 3.6137645458260237,
      "tokens_seen": 1109046272
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033534603811434305,
      "loss": 2.678,
      "theoretical_loss": 3.613744611081798,
      "tokens_seen": 1109111808
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003353360080240722,
      "loss": 2.7416,
      "theoretical_loss": 3.613724677845251,
      "tokens_seen": 1109177344
    },
    {
      "epoch": 3.07,
      "objective/train/docs_used": 1245226,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.714931011199951,
      "objective/train/theoretical_loss": 3.6137196947716657,
      "objective/train/tokens_used": 1129653728,
      "theoretical_loss": 3.6137196947716657,
      "tokens_seen": 1109193728
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003353259779338014,
      "loss": 2.8707,
      "theoretical_loss": 3.6137047461161798,
      "tokens_seen": 1109242880
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003353159478435306,
      "loss": 2.908,
      "theoretical_loss": 3.6136848158943815,
      "tokens_seen": 1109308416
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003353059177532598,
      "loss": 2.5868,
      "theoretical_loss": 3.6136648871796524,
      "tokens_seen": 1109373952
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033529588766298895,
      "loss": 2.7557,
      "theoretical_loss": 3.61364495997179,
      "tokens_seen": 1109439488
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003352858575727182,
      "loss": 2.4417,
      "theoretical_loss": 3.6136250342705907,
      "tokens_seen": 1109505024
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003352758274824473,
      "loss": 2.693,
      "theoretical_loss": 3.613605110075853,
      "tokens_seen": 1109570560
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033526579739217655,
      "loss": 2.9798,
      "theoretical_loss": 3.6135851873873728,
      "tokens_seen": 1109636096
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033525576730190573,
      "loss": 2.6654,
      "theoretical_loss": 3.613565266204948,
      "tokens_seen": 1109701632
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003352457372116349,
      "loss": 2.5048,
      "theoretical_loss": 3.613545346528375,
      "tokens_seen": 1109767168
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003352357071213641,
      "loss": 2.8293,
      "theoretical_loss": 3.613525428357452,
      "tokens_seen": 1109832704
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033522567703109333,
      "loss": 2.658,
      "theoretical_loss": 3.6135055116919754,
      "tokens_seen": 1109898240
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033521564694082245,
      "loss": 2.6935,
      "theoretical_loss": 3.6134855965317434,
      "tokens_seen": 1109963776
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003352056168505517,
      "loss": 2.8863,
      "theoretical_loss": 3.6134656828765523,
      "tokens_seen": 1110029312
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003351955867602808,
      "loss": 2.6411,
      "theoretical_loss": 3.613445770726201,
      "tokens_seen": 1110094848
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033518555667001005,
      "loss": 2.7521,
      "theoretical_loss": 3.613425860080486,
      "tokens_seen": 1110160384
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033517552657973923,
      "loss": 2.4854,
      "theoretical_loss": 3.6134059509392045,
      "tokens_seen": 1110225920
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003351654964894684,
      "loss": 2.8907,
      "theoretical_loss": 3.6133860433021545,
      "tokens_seen": 1110291456
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003351554663991976,
      "loss": 2.616,
      "theoretical_loss": 3.613366137169134,
      "tokens_seen": 1110356992
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003351454363089268,
      "loss": 2.7054,
      "theoretical_loss": 3.6133462325399393,
      "tokens_seen": 1110422528
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033513540621865596,
      "loss": 2.7419,
      "theoretical_loss": 3.6133263294143694,
      "tokens_seen": 1110488064
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003351253761283852,
      "loss": 2.721,
      "theoretical_loss": 3.613306427792221,
      "tokens_seen": 1110553600
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003351153460381143,
      "loss": 2.791,
      "theoretical_loss": 3.6132865276732926,
      "tokens_seen": 1110619136
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033510531594784356,
      "loss": 2.5685,
      "theoretical_loss": 3.6132666290573816,
      "tokens_seen": 1110684672
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003350952858575727,
      "loss": 2.8731,
      "theoretical_loss": 3.613246731944286,
      "tokens_seen": 1110750208
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003350852557673019,
      "loss": 2.5971,
      "theoretical_loss": 3.613226836333803,
      "tokens_seen": 1110815744
    },
    {
      "epoch": 3.07,
      "objective/train/docs_used": 1246492,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.347496747970581,
      "objective/train/theoretical_loss": 3.6132218626659425,
      "objective/train/tokens_used": 1131292128,
      "theoretical_loss": 3.6132218626659425,
      "tokens_seen": 1110832128
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003350752256770311,
      "loss": 2.4939,
      "theoretical_loss": 3.6132069422257316,
      "tokens_seen": 1110881280
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003350651955867603,
      "loss": 2.3556,
      "theoretical_loss": 3.613187049619869,
      "tokens_seen": 1110946816
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033505516549648946,
      "loss": 2.7816,
      "theoretical_loss": 3.6131671585160126,
      "tokens_seen": 1111012352
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003350451354062187,
      "loss": 2.3472,
      "theoretical_loss": 3.6131472689139614,
      "tokens_seen": 1111077888
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003350351053159478,
      "loss": 2.8442,
      "theoretical_loss": 3.6131273808135136,
      "tokens_seen": 1111143424
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033502507522567706,
      "loss": 2.6176,
      "theoretical_loss": 3.613107494214466,
      "tokens_seen": 1111208960
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003350150451354062,
      "loss": 2.7964,
      "theoretical_loss": 3.6130876091166177,
      "tokens_seen": 1111274496
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003350050150451354,
      "loss": 2.5386,
      "theoretical_loss": 3.613067725519767,
      "tokens_seen": 1111340032
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003349949849548646,
      "loss": 2.7889,
      "theoretical_loss": 3.613047843423711,
      "tokens_seen": 1111405568
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003349849548645938,
      "loss": 2.5481,
      "theoretical_loss": 3.61302796282825,
      "tokens_seen": 1111471104
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033497492477432296,
      "loss": 2.804,
      "theoretical_loss": 3.6130080837331797,
      "tokens_seen": 1111536640
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033496489468405215,
      "loss": 2.6183,
      "theoretical_loss": 3.6129882061383,
      "tokens_seen": 1111602176
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003349548645937813,
      "loss": 2.6518,
      "theoretical_loss": 3.612968330043409,
      "tokens_seen": 1111667712
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033494483450351056,
      "loss": 2.731,
      "theoretical_loss": 3.612948455448305,
      "tokens_seen": 1111733248
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003349348044132397,
      "loss": 2.8229,
      "theoretical_loss": 3.6129285823527866,
      "tokens_seen": 1111798784
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003349247743229689,
      "loss": 2.8859,
      "theoretical_loss": 3.6129087107566518,
      "tokens_seen": 1111864320
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033491474423269805,
      "loss": 2.5248,
      "theoretical_loss": 3.6128888406597,
      "tokens_seen": 1111929856
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003349047141424273,
      "loss": 2.869,
      "theoretical_loss": 3.612868972061728,
      "tokens_seen": 1111995392
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033489468405215647,
      "loss": 2.6972,
      "theoretical_loss": 3.6128491049625366,
      "tokens_seen": 1112060928
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033488465396188565,
      "loss": 2.6138,
      "theoretical_loss": 3.612829239361923,
      "tokens_seen": 1112126464
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033487462387161483,
      "loss": 2.9363,
      "theoretical_loss": 3.612809375259686,
      "tokens_seen": 1112192000
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033486459378134406,
      "loss": 2.9163,
      "theoretical_loss": 3.612789512655625,
      "tokens_seen": 1112257536
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003348545636910732,
      "loss": 2.9406,
      "theoretical_loss": 3.612769651549538,
      "tokens_seen": 1112323072
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003348445336008024,
      "loss": 2.6882,
      "theoretical_loss": 3.6127497919412237,
      "tokens_seen": 1112388608
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003348345035105316,
      "loss": 2.8227,
      "theoretical_loss": 3.6127299338304812,
      "tokens_seen": 1112454144
    },
    {
      "epoch": 3.07,
      "objective/train/docs_used": 1247248,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 1.9692270755767822,
      "objective/train/theoretical_loss": 3.6127249695367682,
      "objective/train/tokens_used": 1132930528,
      "theoretical_loss": 3.6127249695367682,
      "tokens_seen": 1112470528
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003348244734202608,
      "loss": 2.3856,
      "theoretical_loss": 3.61271007721711,
      "tokens_seen": 1112519680
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033481444332999,
      "loss": 2.7777,
      "theoretical_loss": 3.612690222100908,
      "tokens_seen": 1112585216
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033480441323971915,
      "loss": 2.5555,
      "theoretical_loss": 3.612670368481675,
      "tokens_seen": 1112650752
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003347943831494484,
      "loss": 2.7394,
      "theoretical_loss": 3.612650516359209,
      "tokens_seen": 1112716288
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003347843530591775,
      "loss": 2.5448,
      "theoretical_loss": 3.6126306657333105,
      "tokens_seen": 1112781824
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033477432296890675,
      "loss": 2.6602,
      "theoretical_loss": 3.6126108166037767,
      "tokens_seen": 1112847360
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033476429287863593,
      "loss": 2.5948,
      "theoretical_loss": 3.612590968970408,
      "tokens_seen": 1112912896
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003347542627883651,
      "loss": 2.8068,
      "theoretical_loss": 3.6125711228330024,
      "tokens_seen": 1112978432
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003347442326980943,
      "loss": 2.6638,
      "theoretical_loss": 3.6125512781913605,
      "tokens_seen": 1113043968
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033473420260782353,
      "loss": 2.8068,
      "theoretical_loss": 3.6125314350452804,
      "tokens_seen": 1113109504
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033472417251755265,
      "loss": 2.7284,
      "theoretical_loss": 3.6125115933945624,
      "tokens_seen": 1113175040
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003347141424272819,
      "loss": 2.6465,
      "theoretical_loss": 3.612491753239004,
      "tokens_seen": 1113240576
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.000334704112337011,
      "loss": 2.5866,
      "theoretical_loss": 3.6124719145784066,
      "tokens_seen": 1113306112
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033469408224674025,
      "loss": 2.776,
      "theoretical_loss": 3.612452077412568,
      "tokens_seen": 1113371648
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033468405215646943,
      "loss": 2.5994,
      "theoretical_loss": 3.612432241741289,
      "tokens_seen": 1113437184
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003346740220661986,
      "loss": 2.7386,
      "theoretical_loss": 3.612412407564367,
      "tokens_seen": 1113502720
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003346639919759278,
      "loss": 2.4634,
      "theoretical_loss": 3.6123925748816035,
      "tokens_seen": 1113568256
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.000334653961885657,
      "loss": 2.6224,
      "theoretical_loss": 3.6123727436927977,
      "tokens_seen": 1113633792
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033464393179538616,
      "loss": 2.8138,
      "theoretical_loss": 3.612352913997748,
      "tokens_seen": 1113699328
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003346339017051154,
      "loss": 2.7124,
      "theoretical_loss": 3.612333085796255,
      "tokens_seen": 1113764864
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003346238716148445,
      "loss": 2.6895,
      "theoretical_loss": 3.6123132590881175,
      "tokens_seen": 1113830400
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033461384152457376,
      "loss": 2.6007,
      "theoretical_loss": 3.612293433873136,
      "tokens_seen": 1113895936
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003346038114343029,
      "loss": 2.5594,
      "theoretical_loss": 3.61227361015111,
      "tokens_seen": 1113961472
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003345937813440321,
      "loss": 2.8955,
      "theoretical_loss": 3.612253787921839,
      "tokens_seen": 1114027008
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003345837512537613,
      "loss": 2.637,
      "theoretical_loss": 3.612233967185123,
      "tokens_seen": 1114092544
    },
    {
      "epoch": 3.07,
      "objective/train/docs_used": 1248654,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.329355001449585,
      "objective/train/theoretical_loss": 3.612229012234132,
      "objective/train/tokens_used": 1134568928,
      "theoretical_loss": 3.612229012234132,
      "tokens_seen": 1114108928
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003345737211634905,
      "loss": 2.763,
      "theoretical_loss": 3.6122141479407617,
      "tokens_seen": 1114158080
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033456369107321966,
      "loss": 2.599,
      "theoretical_loss": 3.6121943301885553,
      "tokens_seen": 1114223616
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003345536609829489,
      "loss": 2.896,
      "theoretical_loss": 3.612174513928303,
      "tokens_seen": 1114289152
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.000334543630892678,
      "loss": 2.5523,
      "theoretical_loss": 3.612154699159806,
      "tokens_seen": 1114354688
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033453360080240726,
      "loss": 2.5513,
      "theoretical_loss": 3.6121348858828624,
      "tokens_seen": 1114420224
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003345235707121364,
      "loss": 2.6675,
      "theoretical_loss": 3.612115074097274,
      "tokens_seen": 1114485760
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003345135406218656,
      "loss": 2.7643,
      "theoretical_loss": 3.6120952638028396,
      "tokens_seen": 1114551296
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003345035105315948,
      "loss": 2.8072,
      "theoretical_loss": 3.61207545499936,
      "tokens_seen": 1114616832
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.000334493480441324,
      "loss": 2.7499,
      "theoretical_loss": 3.6120556476866352,
      "tokens_seen": 1114682368
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033448345035105316,
      "loss": 2.7357,
      "theoretical_loss": 3.612035841864466,
      "tokens_seen": 1114747904
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033447342026078235,
      "loss": 2.9083,
      "theoretical_loss": 3.612016037532651,
      "tokens_seen": 1114813440
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003344633901705115,
      "loss": 2.9005,
      "theoretical_loss": 3.6119962346909915,
      "tokens_seen": 1114878976
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033445336008024076,
      "loss": 2.7063,
      "theoretical_loss": 3.6119764333392883,
      "tokens_seen": 1114944512
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003344433299899699,
      "loss": 2.5994,
      "theoretical_loss": 3.61195663347734,
      "tokens_seen": 1115010048
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003344332998996991,
      "loss": 2.6802,
      "theoretical_loss": 3.6119368351049492,
      "tokens_seen": 1115075584
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033442326980942825,
      "loss": 2.5207,
      "theoretical_loss": 3.6119170382219146,
      "tokens_seen": 1115141120
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003344132397191575,
      "loss": 2.4613,
      "theoretical_loss": 3.6118972428280376,
      "tokens_seen": 1115206656
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033440320962888667,
      "loss": 2.5894,
      "theoretical_loss": 3.611877448923118,
      "tokens_seen": 1115272192
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033439317953861585,
      "loss": 2.5237,
      "theoretical_loss": 3.611857656506957,
      "tokens_seen": 1115337728
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033438314944834503,
      "loss": 2.6439,
      "theoretical_loss": 3.6118378655793544,
      "tokens_seen": 1115403264
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033437311935807427,
      "loss": 2.4564,
      "theoretical_loss": 3.611818076140111,
      "tokens_seen": 1115468800
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003343630892678034,
      "loss": 2.7816,
      "theoretical_loss": 3.611798288189028,
      "tokens_seen": 1115534336
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033435305917753263,
      "loss": 2.8158,
      "theoretical_loss": 3.6117785017259054,
      "tokens_seen": 1115599872
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033434302908726175,
      "loss": 2.7997,
      "theoretical_loss": 3.6117587167505447,
      "tokens_seen": 1115665408
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.000334332998996991,
      "loss": 2.7841,
      "theoretical_loss": 3.611738933262746,
      "tokens_seen": 1115730944
    },
    {
      "epoch": 3.07,
      "objective/train/docs_used": 1249313,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.326357126235962,
      "objective/train/theoretical_loss": 3.6117339876232046,
      "objective/train/tokens_used": 1136207328,
      "theoretical_loss": 3.6117339876232046,
      "tokens_seen": 1115747328
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033432296890672017,
      "loss": 2.5598,
      "theoretical_loss": 3.6117191512623097,
      "tokens_seen": 1115796480
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033431293881644935,
      "loss": 2.6363,
      "theoretical_loss": 3.6116993707490375,
      "tokens_seen": 1115862016
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033430290872617853,
      "loss": 2.5902,
      "theoretical_loss": 3.61167959172273,
      "tokens_seen": 1115927552
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003342928786359077,
      "loss": 2.859,
      "theoretical_loss": 3.611659814183188,
      "tokens_seen": 1115993088
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003342828485456369,
      "loss": 2.7009,
      "theoretical_loss": 3.6116400381302123,
      "tokens_seen": 1116058624
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033427281845536613,
      "loss": 2.8771,
      "theoretical_loss": 3.611620263563605,
      "tokens_seen": 1116124160
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033426278836509526,
      "loss": 2.7368,
      "theoretical_loss": 3.611600490483165,
      "tokens_seen": 1116189696
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003342527582748245,
      "loss": 2.8553,
      "theoretical_loss": 3.611580718888695,
      "tokens_seen": 1116255232
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003342427281845536,
      "loss": 2.6817,
      "theoretical_loss": 3.611560948779996,
      "tokens_seen": 1116320768
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033423269809428285,
      "loss": 2.788,
      "theoretical_loss": 3.611541180156868,
      "tokens_seen": 1116386304
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033422266800401204,
      "loss": 2.6987,
      "theoretical_loss": 3.611521413019113,
      "tokens_seen": 1116451840
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003342126379137412,
      "loss": 2.5525,
      "theoretical_loss": 3.6115016473665325,
      "tokens_seen": 1116517376
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003342026078234704,
      "loss": 2.5741,
      "theoretical_loss": 3.6114818831989277,
      "tokens_seen": 1116582912
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033419257773319963,
      "loss": 2.6247,
      "theoretical_loss": 3.611462120516099,
      "tokens_seen": 1116648448
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033418254764292876,
      "loss": 2.6781,
      "theoretical_loss": 3.6114423593178486,
      "tokens_seen": 1116713984
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.000334172517552658,
      "loss": 2.8599,
      "theoretical_loss": 3.6114225996039773,
      "tokens_seen": 1116779520
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003341624874623871,
      "loss": 2.5877,
      "theoretical_loss": 3.6114028413742867,
      "tokens_seen": 1116845056
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033415245737211636,
      "loss": 2.6957,
      "theoretical_loss": 3.611383084628579,
      "tokens_seen": 1116910592
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033414242728184554,
      "loss": 2.8918,
      "theoretical_loss": 3.6113633293666547,
      "tokens_seen": 1116976128
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003341323971915747,
      "loss": 2.6015,
      "theoretical_loss": 3.6113435755883154,
      "tokens_seen": 1117041664
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003341223671013039,
      "loss": 2.7051,
      "theoretical_loss": 3.611323823293363,
      "tokens_seen": 1117107200
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003341123370110331,
      "loss": 2.6845,
      "theoretical_loss": 3.611304072481598,
      "tokens_seen": 1117172736
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033410230692076226,
      "loss": 2.5767,
      "theoretical_loss": 3.6112843231528244,
      "tokens_seen": 1117238272
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003340922768304915,
      "loss": 2.4481,
      "theoretical_loss": 3.6112645753068415,
      "tokens_seen": 1117303808
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003340822467402207,
      "loss": 2.6844,
      "theoretical_loss": 3.6112448289434527,
      "tokens_seen": 1117369344
    },
    {
      "epoch": 3.07,
      "objective/train/docs_used": 1250355,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9746341705322266,
      "objective/train/theoretical_loss": 3.6112398925842375,
      "objective/train/tokens_used": 1137845728,
      "theoretical_loss": 3.6112398925842375,
      "tokens_seen": 1117385728
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033407221664994986,
      "loss": 2.6491,
      "theoretical_loss": 3.6112250840624585,
      "tokens_seen": 1117434880
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003340621865596791,
      "loss": 2.6187,
      "theoretical_loss": 3.6112053406636617,
      "tokens_seen": 1117500416
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003340521564694082,
      "loss": 2.5762,
      "theoretical_loss": 3.611185598746863,
      "tokens_seen": 1117565952
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033404212637913746,
      "loss": 2.8125,
      "theoretical_loss": 3.611165858311865,
      "tokens_seen": 1117631488
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003340320962888666,
      "loss": 2.8009,
      "theoretical_loss": 3.6111461193584695,
      "tokens_seen": 1117697024
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003340220661985958,
      "loss": 2.7583,
      "theoretical_loss": 3.6111263818864785,
      "tokens_seen": 1117762560
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.000334012036108325,
      "loss": 2.6939,
      "theoretical_loss": 3.6111066458956937,
      "tokens_seen": 1117828096
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003340020060180542,
      "loss": 2.9579,
      "theoretical_loss": 3.6110869113859176,
      "tokens_seen": 1117893632
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033399197592778336,
      "loss": 2.9776,
      "theoretical_loss": 3.611067178356952,
      "tokens_seen": 1117959168
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033398194583751255,
      "loss": 2.8083,
      "theoretical_loss": 3.6110474468085987,
      "tokens_seen": 1118024704
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003339719157472417,
      "loss": 2.5772,
      "theoretical_loss": 3.61102771674066,
      "tokens_seen": 1118090240
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033396188565697096,
      "loss": 2.6275,
      "theoretical_loss": 3.611007988152938,
      "tokens_seen": 1118155776
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003339518555667001,
      "loss": 2.4687,
      "theoretical_loss": 3.6109882610452355,
      "tokens_seen": 1118221312
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003339418254764293,
      "loss": 2.7554,
      "theoretical_loss": 3.6109685354173537,
      "tokens_seen": 1118286848
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033393179538615845,
      "loss": 2.6071,
      "theoretical_loss": 3.610948811269096,
      "tokens_seen": 1118352384
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003339217652958877,
      "loss": 2.705,
      "theoretical_loss": 3.6109290886002636,
      "tokens_seen": 1118417920
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033391173520561687,
      "loss": 2.7084,
      "theoretical_loss": 3.6109093674106596,
      "tokens_seen": 1118483456
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033390170511534605,
      "loss": 2.9054,
      "theoretical_loss": 3.6108896477000862,
      "tokens_seen": 1118548992
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033389167502507523,
      "loss": 2.7773,
      "theoretical_loss": 3.610869929468346,
      "tokens_seen": 1118614528
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033388164493480447,
      "loss": 2.8347,
      "theoretical_loss": 3.610850212715241,
      "tokens_seen": 1118680064
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003338716148445336,
      "loss": 2.7266,
      "theoretical_loss": 3.610830497440574,
      "tokens_seen": 1118745600
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033386158475426283,
      "loss": 2.4787,
      "theoretical_loss": 3.610810783644148,
      "tokens_seen": 1118811136
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033385155466399195,
      "loss": 2.643,
      "theoretical_loss": 3.6107910713257647,
      "tokens_seen": 1118876672
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003338415245737212,
      "loss": 2.7258,
      "theoretical_loss": 3.610771360485227,
      "tokens_seen": 1118942208
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033383149448345037,
      "loss": 2.9903,
      "theoretical_loss": 3.6107516511223383,
      "tokens_seen": 1119007744
    },
    {
      "epoch": 3.07,
      "objective/train/docs_used": 1250984,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.452301502227783,
      "objective/train/theoretical_loss": 3.6107467240124755,
      "objective/train/tokens_used": 1139484128,
      "theoretical_loss": 3.6107467240124755,
      "tokens_seen": 1119024128
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033382146439317955,
      "loss": 2.5613,
      "theoretical_loss": 3.6107319432369005,
      "tokens_seen": 1119073280
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033381143430290873,
      "loss": 2.445,
      "theoretical_loss": 3.610712236828716,
      "tokens_seen": 1119138816
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003338014042126379,
      "loss": 2.8158,
      "theoretical_loss": 3.610692531897589,
      "tokens_seen": 1119204352
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003337913741223671,
      "loss": 2.8361,
      "theoretical_loss": 3.610672828443321,
      "tokens_seen": 1119269888
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033378134403209633,
      "loss": 2.8296,
      "theoretical_loss": 3.6106531264657153,
      "tokens_seen": 1119335424
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033377131394182546,
      "loss": 2.545,
      "theoretical_loss": 3.6106334259645747,
      "tokens_seen": 1119400960
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003337612838515547,
      "loss": 2.8547,
      "theoretical_loss": 3.610613726939702,
      "tokens_seen": 1119466496
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003337512537612838,
      "loss": 2.8374,
      "theoretical_loss": 3.6105940293909002,
      "tokens_seen": 1119532032
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033374122367101306,
      "loss": 2.6208,
      "theoretical_loss": 3.610574333317973,
      "tokens_seen": 1119597568
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033373119358074224,
      "loss": 2.6924,
      "theoretical_loss": 3.610554638720722,
      "tokens_seen": 1119663104
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003337211634904714,
      "loss": 2.8094,
      "theoretical_loss": 3.610534945598952,
      "tokens_seen": 1119728640
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003337111334002006,
      "loss": 2.7446,
      "theoretical_loss": 3.6105152539524648,
      "tokens_seen": 1119794176
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033370110330992983,
      "loss": 2.81,
      "theoretical_loss": 3.6104955637810643,
      "tokens_seen": 1119859712
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033369107321965896,
      "loss": 2.6061,
      "theoretical_loss": 3.6104758750845534,
      "tokens_seen": 1119925248
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003336810431293882,
      "loss": 2.8181,
      "theoretical_loss": 3.6104561878627353,
      "tokens_seen": 1119990784
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003336710130391173,
      "loss": 2.9375,
      "theoretical_loss": 3.6104365021154132,
      "tokens_seen": 1120056320
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033366098294884656,
      "loss": 2.6833,
      "theoretical_loss": 3.61041681784239,
      "tokens_seen": 1120121856
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033365095285857574,
      "loss": 2.9931,
      "theoretical_loss": 3.61039713504347,
      "tokens_seen": 1120187392
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003336409227683049,
      "loss": 2.5654,
      "theoretical_loss": 3.6103774537184563,
      "tokens_seen": 1120252928
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003336308926780341,
      "loss": 2.7439,
      "theoretical_loss": 3.6103577738671513,
      "tokens_seen": 1120318464
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003336208625877633,
      "loss": 2.6787,
      "theoretical_loss": 3.61033809548936,
      "tokens_seen": 1120384000
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033361083249749246,
      "loss": 2.7574,
      "theoretical_loss": 3.6103184185848853,
      "tokens_seen": 1120449536
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003336008024072217,
      "loss": 2.5471,
      "theoretical_loss": 3.61029874315353,
      "tokens_seen": 1120515072
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003335907723169508,
      "loss": 2.8783,
      "theoretical_loss": 3.6102790691950988,
      "tokens_seen": 1120580608
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033358074222668006,
      "loss": 2.819,
      "theoretical_loss": 3.6102593967093943,
      "tokens_seen": 1120646144
    },
    {
      "epoch": 3.07,
      "objective/train/docs_used": 1252222,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9513473510742188,
      "objective/train/theoretical_loss": 3.610254478818059,
      "objective/train/tokens_used": 1141122528,
      "theoretical_loss": 3.610254478818059,
      "tokens_seen": 1120662528
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033357071213640924,
      "loss": 2.7699,
      "theoretical_loss": 3.6102397256962204,
      "tokens_seen": 1120711680
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003335606820461384,
      "loss": 2.6794,
      "theoretical_loss": 3.6102200561553817,
      "tokens_seen": 1120777216
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003335506519558676,
      "loss": 2.694,
      "theoretical_loss": 3.6102003880866804,
      "tokens_seen": 1120842752
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003335406218655968,
      "loss": 2.8837,
      "theoretical_loss": 3.6101807214899218,
      "tokens_seen": 1120908288
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033353059177532597,
      "loss": 2.7733,
      "theoretical_loss": 3.610161056364908,
      "tokens_seen": 1120973824
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003335205616850552,
      "loss": 2.7387,
      "theoretical_loss": 3.6101413927114443,
      "tokens_seen": 1121039360
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033351053159478433,
      "loss": 2.7936,
      "theoretical_loss": 3.6101217305293343,
      "tokens_seen": 1121104896
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033350050150451356,
      "loss": 2.9508,
      "theoretical_loss": 3.610102069818381,
      "tokens_seen": 1121170432
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003334904714142427,
      "loss": 2.6335,
      "theoretical_loss": 3.6100824105783893,
      "tokens_seen": 1121235968
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003334804413239719,
      "loss": 2.6319,
      "theoretical_loss": 3.610062752809163,
      "tokens_seen": 1121301504
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003334704112337011,
      "loss": 2.7402,
      "theoretical_loss": 3.6100430965105055,
      "tokens_seen": 1121367040
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003334603811434303,
      "loss": 2.7282,
      "theoretical_loss": 3.6100234416822223,
      "tokens_seen": 1121432576
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033345035105315947,
      "loss": 2.7094,
      "theoretical_loss": 3.6100037883241156,
      "tokens_seen": 1121498112
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033344032096288865,
      "loss": 2.8534,
      "theoretical_loss": 3.6099841364359904,
      "tokens_seen": 1121563648
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033343029087261783,
      "loss": 2.909,
      "theoretical_loss": 3.609964486017651,
      "tokens_seen": 1121629184
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033342026078234707,
      "loss": 2.9188,
      "theoretical_loss": 3.609944837068902,
      "tokens_seen": 1121694720
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003334102306920762,
      "loss": 2.8513,
      "theoretical_loss": 3.609925189589547,
      "tokens_seen": 1121760256
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033340020060180543,
      "loss": 2.5853,
      "theoretical_loss": 3.60990554357939,
      "tokens_seen": 1121825792
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003333901705115346,
      "loss": 2.6128,
      "theoretical_loss": 3.609885899038236,
      "tokens_seen": 1121891328
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003333801404212638,
      "loss": 2.6936,
      "theoretical_loss": 3.6098662559658887,
      "tokens_seen": 1121956864
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.000333370110330993,
      "loss": 2.5918,
      "theoretical_loss": 3.6098466143621533,
      "tokens_seen": 1122022400
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033336008024072215,
      "loss": 2.6725,
      "theoretical_loss": 3.609826974226834,
      "tokens_seen": 1122087936
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033335005015045134,
      "loss": 2.7519,
      "theoretical_loss": 3.6098073355597347,
      "tokens_seen": 1122153472
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033334002006018057,
      "loss": 2.6089,
      "theoretical_loss": 3.6097876983606607,
      "tokens_seen": 1122219008
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033332998996990975,
      "loss": 2.4769,
      "theoretical_loss": 3.6097680626294157,
      "tokens_seen": 1122284544
    },
    {
      "epoch": 3.07,
      "objective/train/docs_used": 1252729,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7386367321014404,
      "objective/train/theoretical_loss": 3.60976315392593,
      "objective/train/tokens_used": 1142760928,
      "theoretical_loss": 3.60976315392593,
      "tokens_seen": 1122300928
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033331995987963893,
      "loss": 2.7276,
      "theoretical_loss": 3.609748428365805,
      "tokens_seen": 1122350080
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003333099297893681,
      "loss": 2.7054,
      "theoretical_loss": 3.609728795569633,
      "tokens_seen": 1122415616
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003332998996990973,
      "loss": 2.8766,
      "theoretical_loss": 3.6097091642407038,
      "tokens_seen": 1122481152
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033328986960882653,
      "loss": 2.7833,
      "theoretical_loss": 3.6096895343788225,
      "tokens_seen": 1122546688
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033327983951855566,
      "loss": 2.7231,
      "theoretical_loss": 3.6096699059837944,
      "tokens_seen": 1122612224
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003332698094282849,
      "loss": 3.13,
      "theoretical_loss": 3.6096502790554235,
      "tokens_seen": 1122677760
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.000333259779338014,
      "loss": 2.7866,
      "theoretical_loss": 3.6096306535935154,
      "tokens_seen": 1122743296
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033324974924774326,
      "loss": 2.7094,
      "theoretical_loss": 3.6096110295978736,
      "tokens_seen": 1122808832
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033323971915747244,
      "loss": 2.699,
      "theoretical_loss": 3.6095914070683044,
      "tokens_seen": 1122874368
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003332296890672016,
      "loss": 2.8131,
      "theoretical_loss": 3.609571786004612,
      "tokens_seen": 1122939904
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003332196589769308,
      "loss": 2.669,
      "theoretical_loss": 3.6095521664066013,
      "tokens_seen": 1123005440
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033320962888666003,
      "loss": 2.7178,
      "theoretical_loss": 3.6095325482740774,
      "tokens_seen": 1123070976
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033319959879638916,
      "loss": 2.848,
      "theoretical_loss": 3.6095129316068455,
      "tokens_seen": 1123136512
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003331895687061184,
      "loss": 2.7948,
      "theoretical_loss": 3.609493316404711,
      "tokens_seen": 1123202048
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003331795386158475,
      "loss": 2.5098,
      "theoretical_loss": 3.6094737026674775,
      "tokens_seen": 1123267584
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033316950852557676,
      "loss": 2.8703,
      "theoretical_loss": 3.609454090394952,
      "tokens_seen": 1123333120
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033315947843530594,
      "loss": 2.6075,
      "theoretical_loss": 3.6094344795869384,
      "tokens_seen": 1123398656
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003331494483450351,
      "loss": 2.7532,
      "theoretical_loss": 3.609414870243243,
      "tokens_seen": 1123464192
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003331394182547643,
      "loss": 2.7733,
      "theoretical_loss": 3.6093952623636696,
      "tokens_seen": 1123529728
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003331293881644935,
      "loss": 2.8531,
      "theoretical_loss": 3.609375655948025,
      "tokens_seen": 1123595264
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033311935807422266,
      "loss": 2.5311,
      "theoretical_loss": 3.609356050996113,
      "tokens_seen": 1123660800
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003331093279839519,
      "loss": 2.9696,
      "theoretical_loss": 3.609336447507741,
      "tokens_seen": 1123726336
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.000333099297893681,
      "loss": 2.9202,
      "theoretical_loss": 3.609316845482712,
      "tokens_seen": 1123791872
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033308926780341026,
      "loss": 2.9049,
      "theoretical_loss": 3.6092972449208336,
      "tokens_seen": 1123857408
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033307923771313944,
      "loss": 2.842,
      "theoretical_loss": 3.60927764582191,
      "tokens_seen": 1123922944
    },
    {
      "epoch": 3.07,
      "objective/train/docs_used": 1254028,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9574615955352783,
      "objective/train/theoretical_loss": 3.609272746275743,
      "objective/train/tokens_used": 1144399328,
      "theoretical_loss": 3.609272746275743,
      "tokens_seen": 1123939328
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003330692076228686,
      "loss": 2.911,
      "theoretical_loss": 3.6092580481857466,
      "tokens_seen": 1123988480
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003330591775325978,
      "loss": 2.7521,
      "theoretical_loss": 3.60923845201215,
      "tokens_seen": 1124054016
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.000333049147442327,
      "loss": 2.762,
      "theoretical_loss": 3.6092188573009247,
      "tokens_seen": 1124119552
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033303911735205617,
      "loss": 3.0036,
      "theoretical_loss": 3.6091992640518766,
      "tokens_seen": 1124185088
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003330290872617854,
      "loss": 2.87,
      "theoretical_loss": 3.6091796722648124,
      "tokens_seen": 1124250624
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033301905717151453,
      "loss": 2.833,
      "theoretical_loss": 3.609160081939536,
      "tokens_seen": 1124316160
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033300902708124376,
      "loss": 3.1339,
      "theoretical_loss": 3.609140493075855,
      "tokens_seen": 1124381696
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003329989969909729,
      "loss": 2.846,
      "theoretical_loss": 3.6091209056735734,
      "tokens_seen": 1124447232
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003329889669007021,
      "loss": 2.7986,
      "theoretical_loss": 3.609101319732498,
      "tokens_seen": 1124512768
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003329789368104313,
      "loss": 2.8447,
      "theoretical_loss": 3.6090817352524347,
      "tokens_seen": 1124578304
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003329689067201605,
      "loss": 2.6167,
      "theoretical_loss": 3.6090621522331894,
      "tokens_seen": 1124643840
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033295887662988967,
      "loss": 2.6057,
      "theoretical_loss": 3.6090425706745677,
      "tokens_seen": 1124709376
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033294884653961885,
      "loss": 2.8129,
      "theoretical_loss": 3.609022990576376,
      "tokens_seen": 1124774912
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033293881644934803,
      "loss": 2.5224,
      "theoretical_loss": 3.609003411938419,
      "tokens_seen": 1124840448
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033292878635907727,
      "loss": 2.826,
      "theoretical_loss": 3.6089838347605045,
      "tokens_seen": 1124905984
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003329187562688064,
      "loss": 2.4935,
      "theoretical_loss": 3.6089642590424376,
      "tokens_seen": 1124971520
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033290872617853563,
      "loss": 2.5188,
      "theoretical_loss": 3.6089446847840243,
      "tokens_seen": 1125037056
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003328986960882648,
      "loss": 2.6992,
      "theoretical_loss": 3.608925111985071,
      "tokens_seen": 1125102592
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.000332888665997994,
      "loss": 2.7602,
      "theoretical_loss": 3.6089055406453845,
      "tokens_seen": 1125168128
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003328786359077232,
      "loss": 2.6183,
      "theoretical_loss": 3.60888597076477,
      "tokens_seen": 1125233664
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033286860581745235,
      "loss": 2.7598,
      "theoretical_loss": 3.608866402343034,
      "tokens_seen": 1125299200
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033285857572718154,
      "loss": 2.5389,
      "theoretical_loss": 3.608846835379983,
      "tokens_seen": 1125364736
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033284854563691077,
      "loss": 2.928,
      "theoretical_loss": 3.6088272698754236,
      "tokens_seen": 1125430272
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003328385155466399,
      "loss": 2.8564,
      "theoretical_loss": 3.6088077058291614,
      "tokens_seen": 1125495808
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033282848545636913,
      "loss": 2.6543,
      "theoretical_loss": 3.6087881432410036,
      "tokens_seen": 1125561344
    },
    {
      "epoch": 3.07,
      "objective/train/docs_used": 1254775,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.805901527404785,
      "objective/train/theoretical_loss": 3.60878325282177,
      "objective/train/tokens_used": 1146037728,
      "theoretical_loss": 3.60878325282177,
      "tokens_seen": 1125577728
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033281845536609826,
      "loss": 2.7767,
      "theoretical_loss": 3.608768582110756,
      "tokens_seen": 1125626880
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003328084252758275,
      "loss": 2.6417,
      "theoretical_loss": 3.608749022438226,
      "tokens_seen": 1125692416
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003327983951855567,
      "loss": 2.6314,
      "theoretical_loss": 3.6087294642232184,
      "tokens_seen": 1125757952
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033278836509528586,
      "loss": 2.4157,
      "theoretical_loss": 3.6087099074655415,
      "tokens_seen": 1125823488
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033277833500501504,
      "loss": 2.8866,
      "theoretical_loss": 3.6086903521650013,
      "tokens_seen": 1125889024
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003327683049147442,
      "loss": 2.9665,
      "theoretical_loss": 3.608670798321404,
      "tokens_seen": 1125954560
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003327582748244734,
      "loss": 2.6735,
      "theoretical_loss": 3.608651245934557,
      "tokens_seen": 1126020096
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033274824473420264,
      "loss": 2.656,
      "theoretical_loss": 3.6086316950042665,
      "tokens_seen": 1126085632
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033273821464393176,
      "loss": 2.5484,
      "theoretical_loss": 3.608612145530339,
      "tokens_seen": 1126151168
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.000332728184553661,
      "loss": 2.5752,
      "theoretical_loss": 3.608592597512582,
      "tokens_seen": 1126216704
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003327181544633902,
      "loss": 2.5263,
      "theoretical_loss": 3.6085730509508016,
      "tokens_seen": 1126282240
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033270812437311936,
      "loss": 2.968,
      "theoretical_loss": 3.608553505844805,
      "tokens_seen": 1126347776
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033269809428284854,
      "loss": 2.5893,
      "theoretical_loss": 3.6085339621943993,
      "tokens_seen": 1126413312
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003326880641925777,
      "loss": 2.4944,
      "theoretical_loss": 3.6085144199993904,
      "tokens_seen": 1126478848
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003326780341023069,
      "loss": 2.5355,
      "theoretical_loss": 3.608494879259587,
      "tokens_seen": 1126544384
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033266800401203614,
      "loss": 2.702,
      "theoretical_loss": 3.6084753399747944,
      "tokens_seen": 1126609920
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033265797392176527,
      "loss": 2.6876,
      "theoretical_loss": 3.6084558021448205,
      "tokens_seen": 1126675456
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003326479438314945,
      "loss": 2.7097,
      "theoretical_loss": 3.6084362657694724,
      "tokens_seen": 1126740992
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033263791374122363,
      "loss": 2.8215,
      "theoretical_loss": 3.608416730848557,
      "tokens_seen": 1126806528
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033262788365095286,
      "loss": 2.7713,
      "theoretical_loss": 3.608397197381881,
      "tokens_seen": 1126872064
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033261785356068205,
      "loss": 2.6377,
      "theoretical_loss": 3.6083776653692525,
      "tokens_seen": 1126937600
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003326078234704112,
      "loss": 2.8388,
      "theoretical_loss": 3.6083581348104783,
      "tokens_seen": 1127003136
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003325977933801404,
      "loss": 2.8913,
      "theoretical_loss": 3.608338605705365,
      "tokens_seen": 1127068672
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033258776328986964,
      "loss": 2.7082,
      "theoretical_loss": 3.6083190780537207,
      "tokens_seen": 1127134208
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003325777331995988,
      "loss": 2.6834,
      "theoretical_loss": 3.608299551855353,
      "tokens_seen": 1127199744
    },
    {
      "epoch": 3.07,
      "objective/train/docs_used": 1256132,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.0457611083984375,
      "objective/train/theoretical_loss": 3.6082946705328123,
      "objective/train/tokens_used": 1147676128,
      "theoretical_loss": 3.6082946705328123,
      "tokens_seen": 1127216128
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.000332567703109328,
      "loss": 3.1307,
      "theoretical_loss": 3.608280027110068,
      "tokens_seen": 1127265280
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003325576730190572,
      "loss": 2.5617,
      "theoretical_loss": 3.6082605038176743,
      "tokens_seen": 1127330816
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033254764292878637,
      "loss": 2.6656,
      "theoretical_loss": 3.6082409819779784,
      "tokens_seen": 1127396352
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003325376128385156,
      "loss": 2.9161,
      "theoretical_loss": 3.608221461590789,
      "tokens_seen": 1127461888
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033252758274824473,
      "loss": 2.7213,
      "theoretical_loss": 3.6082019426559127,
      "tokens_seen": 1127527424
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033251755265797396,
      "loss": 2.6276,
      "theoretical_loss": 3.6081824251731565,
      "tokens_seen": 1127592960
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003325075225677031,
      "loss": 2.6259,
      "theoretical_loss": 3.6081629091423295,
      "tokens_seen": 1127658496
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033249749247743233,
      "loss": 2.7656,
      "theoretical_loss": 3.608143394563238,
      "tokens_seen": 1127724032
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003324874623871615,
      "loss": 2.7165,
      "theoretical_loss": 3.6081238814356906,
      "tokens_seen": 1127789568
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003324774322968907,
      "loss": 2.8234,
      "theoretical_loss": 3.608104369759494,
      "tokens_seen": 1127855104
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033246740220661987,
      "loss": 2.9237,
      "theoretical_loss": 3.608084859534457,
      "tokens_seen": 1127920640
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033245737211634905,
      "loss": 2.7804,
      "theoretical_loss": 3.6080653507603864,
      "tokens_seen": 1127986176
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033244734202607823,
      "loss": 2.7878,
      "theoretical_loss": 3.6080458434370906,
      "tokens_seen": 1128051712
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033243731193580747,
      "loss": 2.8638,
      "theoretical_loss": 3.608026337564377,
      "tokens_seen": 1128117248
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003324272818455366,
      "loss": 2.5568,
      "theoretical_loss": 3.6080068331420545,
      "tokens_seen": 1128182784
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033241725175526583,
      "loss": 2.7804,
      "theoretical_loss": 3.6079873301699292,
      "tokens_seen": 1128248320
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.000332407221664995,
      "loss": 2.6939,
      "theoretical_loss": 3.607967828647811,
      "tokens_seen": 1128313856
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003323971915747242,
      "loss": 2.9341,
      "theoretical_loss": 3.6079483285755067,
      "tokens_seen": 1128379392
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003323871614844534,
      "loss": 2.8499,
      "theoretical_loss": 3.6079288299528245,
      "tokens_seen": 1128444928
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033237713139418255,
      "loss": 2.6364,
      "theoretical_loss": 3.6079093327795726,
      "tokens_seen": 1128510464
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033236710130391174,
      "loss": 2.692,
      "theoretical_loss": 3.607889837055559,
      "tokens_seen": 1128576000
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033235707121364097,
      "loss": 2.7181,
      "theoretical_loss": 3.607870342780592,
      "tokens_seen": 1128641536
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003323470411233701,
      "loss": 2.6819,
      "theoretical_loss": 3.6078508499544797,
      "tokens_seen": 1128707072
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033233701103309933,
      "loss": 2.7742,
      "theoretical_loss": 3.6078313585770303,
      "tokens_seen": 1128772608
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033232698094282846,
      "loss": 2.8938,
      "theoretical_loss": 3.607811868648052,
      "tokens_seen": 1128838144
    },
    {
      "epoch": 3.07,
      "objective/train/docs_used": 1256724,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9342129230499268,
      "objective/train/theoretical_loss": 3.6078069963921084,
      "objective/train/tokens_used": 1149314528,
      "theoretical_loss": 3.6078069963921084,
      "tokens_seen": 1128854528
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003323169508525577,
      "loss": 2.6303,
      "theoretical_loss": 3.6077923801673526,
      "tokens_seen": 1128903680
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003323069207622869,
      "loss": 2.9345,
      "theoretical_loss": 3.607772893134741,
      "tokens_seen": 1128969216
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033229689067201606,
      "loss": 2.5573,
      "theoretical_loss": 3.6077534075500264,
      "tokens_seen": 1129034752
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033228686058174524,
      "loss": 2.6129,
      "theoretical_loss": 3.607733923413015,
      "tokens_seen": 1129100288
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003322768304914744,
      "loss": 2.7973,
      "theoretical_loss": 3.607714440723517,
      "tokens_seen": 1129165824
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003322668004012036,
      "loss": 2.7817,
      "theoretical_loss": 3.60769495948134,
      "tokens_seen": 1129231360
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033225677031093284,
      "loss": 2.9254,
      "theoretical_loss": 3.6076754796862933,
      "tokens_seen": 1129296896
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033224674022066196,
      "loss": 2.8688,
      "theoretical_loss": 3.6076560013381846,
      "tokens_seen": 1129362432
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003322367101303912,
      "loss": 2.731,
      "theoretical_loss": 3.607636524436823,
      "tokens_seen": 1129427968
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003322266800401204,
      "loss": 2.3886,
      "theoretical_loss": 3.607617048982017,
      "tokens_seen": 1129493504
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033221664994984956,
      "loss": 2.4788,
      "theoretical_loss": 3.6075975749735747,
      "tokens_seen": 1129559040
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033220661985957874,
      "loss": 2.6113,
      "theoretical_loss": 3.6075781024113054,
      "tokens_seen": 1129624576
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003321965897693079,
      "loss": 2.7369,
      "theoretical_loss": 3.607558631295017,
      "tokens_seen": 1129690112
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003321865596790371,
      "loss": 2.7931,
      "theoretical_loss": 3.60753916162452,
      "tokens_seen": 1129755648
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033217652958876634,
      "loss": 2.6653,
      "theoretical_loss": 3.607519693399621,
      "tokens_seen": 1129821184
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033216649949849547,
      "loss": 2.771,
      "theoretical_loss": 3.6075002266201306,
      "tokens_seen": 1129886720
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003321564694082247,
      "loss": 2.8361,
      "theoretical_loss": 3.6074807612858573,
      "tokens_seen": 1129952256
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033214643931795383,
      "loss": 2.4755,
      "theoretical_loss": 3.6074612973966085,
      "tokens_seen": 1130017792
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033213640922768306,
      "loss": 2.882,
      "theoretical_loss": 3.607441834952195,
      "tokens_seen": 1130083328
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033212637913741225,
      "loss": 2.6837,
      "theoretical_loss": 3.607422373952425,
      "tokens_seen": 1130148864
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003321163490471414,
      "loss": 2.825,
      "theoretical_loss": 3.607402914397108,
      "tokens_seen": 1130214400
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003321063189568706,
      "loss": 2.7678,
      "theoretical_loss": 3.607383456286051,
      "tokens_seen": 1130279936
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033209628886659984,
      "loss": 2.7812,
      "theoretical_loss": 3.607363999619066,
      "tokens_seen": 1130345472
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033208625877632897,
      "loss": 2.7462,
      "theoretical_loss": 3.6073445443959606,
      "tokens_seen": 1130411008
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003320762286860582,
      "loss": 2.9633,
      "theoretical_loss": 3.607325090616544,
      "tokens_seen": 1130476544
    },
    {
      "epoch": 3.07,
      "objective/train/docs_used": 1258334,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.567551851272583,
      "objective/train/theoretical_loss": 3.607320227397244,
      "objective/train/tokens_used": 1150952928,
      "theoretical_loss": 3.607320227397244,
      "tokens_seen": 1130492928
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033206619859578733,
      "loss": 2.753,
      "theoretical_loss": 3.607305638280626,
      "tokens_seen": 1130542080
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033205616850551657,
      "loss": 2.5546,
      "theoretical_loss": 3.6072861873880147,
      "tokens_seen": 1130607616
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033204613841524575,
      "loss": 2.8395,
      "theoretical_loss": 3.6072667379385197,
      "tokens_seen": 1130673152
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033203610832497493,
      "loss": 2.8606,
      "theoretical_loss": 3.607247289931951,
      "tokens_seen": 1130738688
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003320260782347041,
      "loss": 2.7365,
      "theoretical_loss": 3.6072278433681175,
      "tokens_seen": 1130804224
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003320160481444333,
      "loss": 2.7856,
      "theoretical_loss": 3.6072083982468284,
      "tokens_seen": 1130869760
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003320060180541625,
      "loss": 2.5667,
      "theoretical_loss": 3.607188954567894,
      "tokens_seen": 1130935296
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003319959879638917,
      "loss": 2.629,
      "theoretical_loss": 3.6071695123311223,
      "tokens_seen": 1131000832
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033198595787362084,
      "loss": 2.7256,
      "theoretical_loss": 3.6071500715363234,
      "tokens_seen": 1131066368
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033197592778335007,
      "loss": 3.0258,
      "theoretical_loss": 3.607130632183307,
      "tokens_seen": 1131131904
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003319658976930792,
      "loss": 2.8214,
      "theoretical_loss": 3.6071111942718836,
      "tokens_seen": 1131197440
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033195586760280843,
      "loss": 2.6264,
      "theoretical_loss": 3.607091757801861,
      "tokens_seen": 1131262976
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003319458375125376,
      "loss": 2.6839,
      "theoretical_loss": 3.607072322773049,
      "tokens_seen": 1131328512
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003319358074222668,
      "loss": 2.9318,
      "theoretical_loss": 3.6070528891852582,
      "tokens_seen": 1131394048
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.000331925777331996,
      "loss": 2.7424,
      "theoretical_loss": 3.6070334570382983,
      "tokens_seen": 1131459584
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003319157472417252,
      "loss": 2.7137,
      "theoretical_loss": 3.6070140263319783,
      "tokens_seen": 1131525120
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033190571715145434,
      "loss": 2.8819,
      "theoretical_loss": 3.6069945970661084,
      "tokens_seen": 1131590656
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003318956870611836,
      "loss": 2.8069,
      "theoretical_loss": 3.6069751692404983,
      "tokens_seen": 1131656192
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003318856569709127,
      "loss": 2.7572,
      "theoretical_loss": 3.606955742854958,
      "tokens_seen": 1131721728
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033187562688064194,
      "loss": 2.646,
      "theoretical_loss": 3.606936317909297,
      "tokens_seen": 1131787264
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003318655967903711,
      "loss": 2.8115,
      "theoretical_loss": 3.6069168944033256,
      "tokens_seen": 1131852800
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003318555667001003,
      "loss": 2.629,
      "theoretical_loss": 3.606897472336853,
      "tokens_seen": 1131918336
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003318455366098295,
      "loss": 2.5971,
      "theoretical_loss": 3.6068780517096903,
      "tokens_seen": 1131983872
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033183550651955866,
      "loss": 2.7799,
      "theoretical_loss": 3.606858632521647,
      "tokens_seen": 1132049408
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003318254764292879,
      "loss": 2.7061,
      "theoretical_loss": 3.606839214772533,
      "tokens_seen": 1132114944
    },
    {
      "epoch": 3.07,
      "objective/train/docs_used": 1259128,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.0284805297851562,
      "objective/train/theoretical_loss": 3.606834360560065,
      "objective/train/tokens_used": 1152591328,
      "theoretical_loss": 3.606834360560065,
      "tokens_seen": 1132131328
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003318154463390171,
      "loss": 2.8331,
      "theoretical_loss": 3.6068197984621584,
      "tokens_seen": 1132180480
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033180541624874626,
      "loss": 2.4103,
      "theoretical_loss": 3.6068003835903335,
      "tokens_seen": 1132246016
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033179538615847544,
      "loss": 2.5474,
      "theoretical_loss": 3.6067809701568687,
      "tokens_seen": 1132311552
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003317853560682046,
      "loss": 2.5511,
      "theoretical_loss": 3.6067615581615735,
      "tokens_seen": 1132377088
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003317753259779338,
      "loss": 2.7727,
      "theoretical_loss": 3.606742147604259,
      "tokens_seen": 1132442624
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033176529588766304,
      "loss": 2.8224,
      "theoretical_loss": 3.6067227384847347,
      "tokens_seen": 1132508160
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033175526579739216,
      "loss": 2.6972,
      "theoretical_loss": 3.6067033308028114,
      "tokens_seen": 1132573696
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003317452357071214,
      "loss": 2.626,
      "theoretical_loss": 3.6066839245582987,
      "tokens_seen": 1132639232
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003317352056168506,
      "loss": 2.6497,
      "theoretical_loss": 3.6066645197510083,
      "tokens_seen": 1132704768
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033172517552657976,
      "loss": 2.6526,
      "theoretical_loss": 3.6066451163807494,
      "tokens_seen": 1132770304
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033171514543630894,
      "loss": 2.8228,
      "theoretical_loss": 3.606625714447333,
      "tokens_seen": 1132835840
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003317051153460381,
      "loss": 2.9014,
      "theoretical_loss": 3.6066063139505697,
      "tokens_seen": 1132901376
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003316950852557673,
      "loss": 2.4233,
      "theoretical_loss": 3.6065869148902694,
      "tokens_seen": 1132966912
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033168505516549654,
      "loss": 2.7184,
      "theoretical_loss": 3.6065675172662433,
      "tokens_seen": 1133032448
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033167502507522567,
      "loss": 2.7,
      "theoretical_loss": 3.606548121078302,
      "tokens_seen": 1133097984
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003316649949849549,
      "loss": 2.771,
      "theoretical_loss": 3.6065287263262555,
      "tokens_seen": 1133163520
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033165496489468403,
      "loss": 2.7794,
      "theoretical_loss": 3.6065093330099147,
      "tokens_seen": 1133229056
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033164493480441326,
      "loss": 2.5853,
      "theoretical_loss": 3.606489941129091,
      "tokens_seen": 1133294592
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033163490471414245,
      "loss": 2.5687,
      "theoretical_loss": 3.6064705506835946,
      "tokens_seen": 1133360128
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003316248746238716,
      "loss": 2.7562,
      "theoretical_loss": 3.6064511616732355,
      "tokens_seen": 1133425664
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003316148445336008,
      "loss": 2.6475,
      "theoretical_loss": 3.6064317740978264,
      "tokens_seen": 1133491200
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033160481444333004,
      "loss": 2.8928,
      "theoretical_loss": 3.6064123879571763,
      "tokens_seen": 1133556736
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033159478435305917,
      "loss": 2.8312,
      "theoretical_loss": 3.606393003251097,
      "tokens_seen": 1133622272
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003315847542627884,
      "loss": 2.723,
      "theoretical_loss": 3.606373619979399,
      "tokens_seen": 1133687808
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033157472417251753,
      "loss": 2.8412,
      "theoretical_loss": 3.606354238141894,
      "tokens_seen": 1133753344
    },
    {
      "epoch": 3.07,
      "objective/train/docs_used": 1260370,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.1477930545806885,
      "objective/train/theoretical_loss": 3.606349392906588,
      "objective/train/tokens_used": 1154229728,
      "theoretical_loss": 3.606349392906588,
      "tokens_seen": 1133769728
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033156469408224677,
      "loss": 2.5493,
      "theoretical_loss": 3.606334857738392,
      "tokens_seen": 1133818880
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033155466399197595,
      "loss": 2.5485,
      "theoretical_loss": 3.606315478768704,
      "tokens_seen": 1133884416
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033154463390170513,
      "loss": 2.8706,
      "theoretical_loss": 3.6062961012326427,
      "tokens_seen": 1133949952
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003315346038114343,
      "loss": 2.6468,
      "theoretical_loss": 3.6062767251300176,
      "tokens_seen": 1134015488
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003315245737211635,
      "loss": 2.7919,
      "theoretical_loss": 3.6062573504606403,
      "tokens_seen": 1134081024
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003315145436308927,
      "loss": 2.8601,
      "theoretical_loss": 3.6062379772243216,
      "tokens_seen": 1134146560
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003315045135406219,
      "loss": 2.712,
      "theoretical_loss": 3.6062186054208736,
      "tokens_seen": 1134212096
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033149448345035104,
      "loss": 2.5173,
      "theoretical_loss": 3.6061992350501066,
      "tokens_seen": 1134277632
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033148445336008027,
      "loss": 2.4526,
      "theoretical_loss": 3.606179866111832,
      "tokens_seen": 1134343168
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003314744232698094,
      "loss": 2.8311,
      "theoretical_loss": 3.6061604986058615,
      "tokens_seen": 1134408704
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033146439317953863,
      "loss": 2.8495,
      "theoretical_loss": 3.6061411325320067,
      "tokens_seen": 1134474240
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003314543630892678,
      "loss": 2.9322,
      "theoretical_loss": 3.6061217678900785,
      "tokens_seen": 1134539776
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.000331444332998997,
      "loss": 2.5131,
      "theoretical_loss": 3.6061024046798886,
      "tokens_seen": 1134605312
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003314343029087262,
      "loss": 2.8674,
      "theoretical_loss": 3.606083042901248,
      "tokens_seen": 1134670848
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003314242728184554,
      "loss": 2.8922,
      "theoretical_loss": 3.6060636825539687,
      "tokens_seen": 1134736384
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033141424272818454,
      "loss": 2.7273,
      "theoretical_loss": 3.6060443236378616,
      "tokens_seen": 1134801920
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003314042126379138,
      "loss": 2.7511,
      "theoretical_loss": 3.606024966152739,
      "tokens_seen": 1134867456
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003313941825476429,
      "loss": 2.6649,
      "theoretical_loss": 3.606005610098412,
      "tokens_seen": 1134932992
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033138415245737214,
      "loss": 2.8792,
      "theoretical_loss": 3.6059862554746926,
      "tokens_seen": 1134998528
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003313741223671013,
      "loss": 2.6218,
      "theoretical_loss": 3.605966902281392,
      "tokens_seen": 1135064064
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003313640922768305,
      "loss": 2.811,
      "theoretical_loss": 3.605947550518322,
      "tokens_seen": 1135129600
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003313540621865597,
      "loss": 2.5705,
      "theoretical_loss": 3.6059282001852946,
      "tokens_seen": 1135195136
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033134403209628886,
      "loss": 2.9007,
      "theoretical_loss": 3.6059088512821216,
      "tokens_seen": 1135260672
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033133400200601804,
      "loss": 2.7223,
      "theoretical_loss": 3.6058895038086147,
      "tokens_seen": 1135326208
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003313239719157473,
      "loss": 2.8814,
      "theoretical_loss": 3.6058701577645857,
      "tokens_seen": 1135391744
    },
    {
      "epoch": 3.07,
      "objective/train/docs_used": 1261008,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7761924266815186,
      "objective/train/theoretical_loss": 3.6058653214769123,
      "objective/train/tokens_used": 1155868128,
      "theoretical_loss": 3.6058653214769123,
      "tokens_seen": 1135408128
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003313139418254764,
      "loss": 2.8254,
      "theoretical_loss": 3.605850813149847,
      "tokens_seen": 1135457280
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033130391173520564,
      "loss": 2.6401,
      "theoretical_loss": 3.605831469964209,
      "tokens_seen": 1135522816
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033129388164493477,
      "loss": 2.9165,
      "theoretical_loss": 3.6058121282074853,
      "tokens_seen": 1135588352
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.000331283851554664,
      "loss": 2.6665,
      "theoretical_loss": 3.605792787879487,
      "tokens_seen": 1135653888
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003312738214643932,
      "loss": 2.8296,
      "theoretical_loss": 3.6057734489800266,
      "tokens_seen": 1135719424
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033126379137412236,
      "loss": 2.4184,
      "theoretical_loss": 3.605754111508916,
      "tokens_seen": 1135784960
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033125376128385155,
      "loss": 2.8367,
      "theoretical_loss": 3.605734775465967,
      "tokens_seen": 1135850496
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003312437311935808,
      "loss": 2.6468,
      "theoretical_loss": 3.605715440850992,
      "tokens_seen": 1135916032
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003312337011033099,
      "loss": 2.5701,
      "theoretical_loss": 3.6056961076638037,
      "tokens_seen": 1135981568
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033122367101303914,
      "loss": 2.7412,
      "theoretical_loss": 3.6056767759042136,
      "tokens_seen": 1136047104
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033121364092276827,
      "loss": 2.4669,
      "theoretical_loss": 3.6056574455720334,
      "tokens_seen": 1136112640
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003312036108324975,
      "loss": 2.7462,
      "theoretical_loss": 3.605638116667077,
      "tokens_seen": 1136178176
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003311935807422267,
      "loss": 2.8023,
      "theoretical_loss": 3.605618789189155,
      "tokens_seen": 1136243712
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033118355065195587,
      "loss": 2.5532,
      "theoretical_loss": 3.6055994631380806,
      "tokens_seen": 1136309248
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033117352056168505,
      "loss": 2.6124,
      "theoretical_loss": 3.6055801385136665,
      "tokens_seen": 1136374784
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033116349047141423,
      "loss": 2.5696,
      "theoretical_loss": 3.605560815315725,
      "tokens_seen": 1136440320
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003311534603811434,
      "loss": 2.5819,
      "theoretical_loss": 3.605541493544067,
      "tokens_seen": 1136505856
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033114343029087265,
      "loss": 2.8007,
      "theoretical_loss": 3.6055221731985068,
      "tokens_seen": 1136571392
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033113340020060177,
      "loss": 2.6863,
      "theoretical_loss": 3.605502854278857,
      "tokens_seen": 1136636928
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.000331123370110331,
      "loss": 2.8361,
      "theoretical_loss": 3.605483536784929,
      "tokens_seen": 1136702464
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033111334002006013,
      "loss": 2.8174,
      "theoretical_loss": 3.605464220716536,
      "tokens_seen": 1136768000
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033110330992978937,
      "loss": 2.7956,
      "theoretical_loss": 3.6054449060734903,
      "tokens_seen": 1136833536
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033109327983951855,
      "loss": 2.8733,
      "theoretical_loss": 3.605425592855605,
      "tokens_seen": 1136899072
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033108324974924773,
      "loss": 2.6193,
      "theoretical_loss": 3.6054062810626926,
      "tokens_seen": 1136964608
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033107321965897697,
      "loss": 2.8384,
      "theoretical_loss": 3.605386970694566,
      "tokens_seen": 1137030144
    },
    {
      "epoch": 3.07,
      "objective/train/docs_used": 1262017,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.1084911823272705,
      "objective/train/theoretical_loss": 3.6053821433251354,
      "objective/train/tokens_used": 1157506528,
      "theoretical_loss": 3.6053821433251354,
      "tokens_seen": 1137046528
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033106318956870615,
      "loss": 2.7541,
      "theoretical_loss": 3.6053676617510377,
      "tokens_seen": 1137095680
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033105315947843533,
      "loss": 2.5768,
      "theoretical_loss": 3.6053483542319205,
      "tokens_seen": 1137161216
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003310431293881645,
      "loss": 2.6472,
      "theoretical_loss": 3.6053290481370275,
      "tokens_seen": 1137226752
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003310330992978937,
      "loss": 2.4923,
      "theoretical_loss": 3.6053097434661714,
      "tokens_seen": 1137292288
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003310230692076229,
      "loss": 2.7072,
      "theoretical_loss": 3.6052904402191652,
      "tokens_seen": 1137357824
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003310130391173521,
      "loss": 2.7793,
      "theoretical_loss": 3.6052711383958216,
      "tokens_seen": 1137423360
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033100300902708124,
      "loss": 2.8709,
      "theoretical_loss": 3.605251837995954,
      "tokens_seen": 1137488896
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033099297893681047,
      "loss": 2.58,
      "theoretical_loss": 3.605232539019375,
      "tokens_seen": 1137554432
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0003309829488465396,
      "loss": 2.7207,
      "theoretical_loss": 3.605213241465898,
      "tokens_seen": 1137619968
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.00033097291875626883,
      "loss": 2.5586,
      "theoretical_loss": 3.605193945335336,
      "tokens_seen": 1137685504
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.000330962888665998,
      "loss": 2.5498,
      "theoretical_loss": 3.605174650627502,
      "tokens_seen": 1137751040
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003309528585757272,
      "loss": 2.7564,
      "theoretical_loss": 3.6051553573422086,
      "tokens_seen": 1137816576
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003309428284854564,
      "loss": 2.8811,
      "theoretical_loss": 3.60513606547927,
      "tokens_seen": 1137882112
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003309327983951856,
      "loss": 2.7868,
      "theoretical_loss": 3.6051167750384994,
      "tokens_seen": 1137947648
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00033092276830491474,
      "loss": 2.8146,
      "theoretical_loss": 3.6050974860197096,
      "tokens_seen": 1138013184
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.000330912738214644,
      "loss": 2.7553,
      "theoretical_loss": 3.6050781984227136,
      "tokens_seen": 1138078720
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003309027081243731,
      "loss": 2.8369,
      "theoretical_loss": 3.6050589122473253,
      "tokens_seen": 1138144256
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00033089267803410234,
      "loss": 2.5716,
      "theoretical_loss": 3.605039627493358,
      "tokens_seen": 1138209792
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003308826479438315,
      "loss": 2.9157,
      "theoretical_loss": 3.605020344160625,
      "tokens_seen": 1138275328
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003308726178535607,
      "loss": 2.565,
      "theoretical_loss": 3.60500106224894,
      "tokens_seen": 1138340864
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003308625877632899,
      "loss": 2.4975,
      "theoretical_loss": 3.6049817817581156,
      "tokens_seen": 1138406400
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00033085255767301906,
      "loss": 2.6338,
      "theoretical_loss": 3.6049625026879664,
      "tokens_seen": 1138471936
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00033084252758274824,
      "loss": 2.6649,
      "theoretical_loss": 3.604943225038305,
      "tokens_seen": 1138537472
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003308324974924775,
      "loss": 2.669,
      "theoretical_loss": 3.604923948808946,
      "tokens_seen": 1138603008
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003308224674022066,
      "loss": 2.7445,
      "theoretical_loss": 3.604904673999702,
      "tokens_seen": 1138668544
    },
    {
      "epoch": 3.08,
      "objective/train/docs_used": 1262682,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.618945837020874,
      "objective/train/theoretical_loss": 3.604899855519262,
      "objective/train/tokens_used": 1159144928,
      "theoretical_loss": 3.604899855519262,
      "tokens_seen": 1138684928
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00033081243731193584,
      "loss": 2.759,
      "theoretical_loss": 3.604885400610388,
      "tokens_seen": 1138734080
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00033080240722166497,
      "loss": 2.7804,
      "theoretical_loss": 3.6048661286408157,
      "tokens_seen": 1138799616
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003307923771313942,
      "loss": 2.8219,
      "theoretical_loss": 3.6048468580908004,
      "tokens_seen": 1138865152
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003307823470411234,
      "loss": 2.6843,
      "theoretical_loss": 3.6048275889601555,
      "tokens_seen": 1138930688
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00033077231695085256,
      "loss": 2.9408,
      "theoretical_loss": 3.6048083212486945,
      "tokens_seen": 1138996224
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00033076228686058175,
      "loss": 2.6726,
      "theoretical_loss": 3.604789054956232,
      "tokens_seen": 1139061760
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.000330752256770311,
      "loss": 2.8145,
      "theoretical_loss": 3.604769790082581,
      "tokens_seen": 1139127296
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003307422266800401,
      "loss": 2.6603,
      "theoretical_loss": 3.604750526627555,
      "tokens_seen": 1139192832
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00033073219658976934,
      "loss": 2.6343,
      "theoretical_loss": 3.6047312645909697,
      "tokens_seen": 1139258368
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00033072216649949847,
      "loss": 2.5703,
      "theoretical_loss": 3.6047120039726375,
      "tokens_seen": 1139323904
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003307121364092277,
      "loss": 3.0216,
      "theoretical_loss": 3.6046927447723727,
      "tokens_seen": 1139389440
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003307021063189569,
      "loss": 2.7889,
      "theoretical_loss": 3.6046734869899892,
      "tokens_seen": 1139454976
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00033069207622868607,
      "loss": 2.7531,
      "theoretical_loss": 3.604654230625302,
      "tokens_seen": 1139520512
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00033068204613841525,
      "loss": 2.6034,
      "theoretical_loss": 3.6046349756781244,
      "tokens_seen": 1139586048
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00033067201604814443,
      "loss": 2.7754,
      "theoretical_loss": 3.604615722148271,
      "tokens_seen": 1139651584
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003306619859578736,
      "loss": 2.7181,
      "theoretical_loss": 3.6045964700355553,
      "tokens_seen": 1139717120
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00033065195586760285,
      "loss": 2.8112,
      "theoretical_loss": 3.6045772193397916,
      "tokens_seen": 1139782656
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00033064192577733197,
      "loss": 2.7175,
      "theoretical_loss": 3.6045579700607955,
      "tokens_seen": 1139848192
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003306318956870612,
      "loss": 2.7196,
      "theoretical_loss": 3.6045387221983796,
      "tokens_seen": 1139913728
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00033062186559679034,
      "loss": 2.5488,
      "theoretical_loss": 3.604519475752359,
      "tokens_seen": 1139979264
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00033061183550651957,
      "loss": 2.7686,
      "theoretical_loss": 3.604500230722548,
      "tokens_seen": 1140044800
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00033060180541624875,
      "loss": 2.8398,
      "theoretical_loss": 3.6044809871087606,
      "tokens_seen": 1140110336
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00033059177532597793,
      "loss": 2.8059,
      "theoretical_loss": 3.604461744910812,
      "tokens_seen": 1140175872
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003305817452357071,
      "loss": 2.8029,
      "theoretical_loss": 3.6044425041285164,
      "tokens_seen": 1140241408
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00033057171514543635,
      "loss": 2.8484,
      "theoretical_loss": 3.604423264761688,
      "tokens_seen": 1140306944
    },
    {
      "epoch": 3.08,
      "objective/train/docs_used": 1264116,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.787461757659912,
      "objective/train/theoretical_loss": 3.6044184551411256,
      "objective/train/tokens_used": 1160783328,
      "theoretical_loss": 3.6044184551411256,
      "tokens_seen": 1140323328
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003305616850551655,
      "loss": 2.7743,
      "theoretical_loss": 3.604404026810141,
      "tokens_seen": 1140372480
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003305516549648947,
      "loss": 2.7962,
      "theoretical_loss": 3.6043847902736905,
      "tokens_seen": 1140438016
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00033054162487462384,
      "loss": 2.5863,
      "theoretical_loss": 3.6043655551521514,
      "tokens_seen": 1140503552
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003305315947843531,
      "loss": 2.7442,
      "theoretical_loss": 3.6043463214453375,
      "tokens_seen": 1140569088
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00033052156469408225,
      "loss": 2.7425,
      "theoretical_loss": 3.604327089153064,
      "tokens_seen": 1140634624
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00033051153460381144,
      "loss": 2.7151,
      "theoretical_loss": 3.6043078582751455,
      "tokens_seen": 1140700160
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003305015045135406,
      "loss": 2.7677,
      "theoretical_loss": 3.604288628811397,
      "tokens_seen": 1140765696
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003304914744232698,
      "loss": 2.8816,
      "theoretical_loss": 3.6042694007616327,
      "tokens_seen": 1140831232
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.000330481444332999,
      "loss": 2.75,
      "theoretical_loss": 3.604250174125668,
      "tokens_seen": 1140896768
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003304714142427282,
      "loss": 2.8542,
      "theoretical_loss": 3.604230948903317,
      "tokens_seen": 1140962304
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00033046138415245734,
      "loss": 2.7008,
      "theoretical_loss": 3.6042117250943955,
      "tokens_seen": 1141027840
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003304513540621866,
      "loss": 2.8737,
      "theoretical_loss": 3.6041925026987176,
      "tokens_seen": 1141093376
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003304413239719157,
      "loss": 2.6164,
      "theoretical_loss": 3.6041732817160987,
      "tokens_seen": 1141158912
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00033043129388164494,
      "loss": 2.9281,
      "theoretical_loss": 3.604154062146354,
      "tokens_seen": 1141224448
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003304212637913741,
      "loss": 2.714,
      "theoretical_loss": 3.604134843989298,
      "tokens_seen": 1141289984
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003304112337011033,
      "loss": 2.6748,
      "theoretical_loss": 3.604115627244746,
      "tokens_seen": 1141355520
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003304012036108325,
      "loss": 2.5346,
      "theoretical_loss": 3.604096411912513,
      "tokens_seen": 1141421056
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003303911735205617,
      "loss": 2.7443,
      "theoretical_loss": 3.604077197992414,
      "tokens_seen": 1141486592
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00033038114343029084,
      "loss": 2.901,
      "theoretical_loss": 3.604057985484265,
      "tokens_seen": 1141552128
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003303711133400201,
      "loss": 2.9053,
      "theoretical_loss": 3.6040387743878797,
      "tokens_seen": 1141617664
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003303610832497492,
      "loss": 2.9141,
      "theoretical_loss": 3.6040195647030746,
      "tokens_seen": 1141683200
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00033035105315947844,
      "loss": 2.6902,
      "theoretical_loss": 3.604000356429664,
      "tokens_seen": 1141748736
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003303410230692076,
      "loss": 2.7633,
      "theoretical_loss": 3.6039811495674643,
      "tokens_seen": 1141814272
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003303309929789368,
      "loss": 2.9135,
      "theoretical_loss": 3.60396194411629,
      "tokens_seen": 1141879808
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00033032096288866604,
      "loss": 2.5842,
      "theoretical_loss": 3.6039427400759565,
      "tokens_seen": 1141945344
    },
    {
      "epoch": 3.08,
      "objective/train/docs_used": 1264658,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5670065879821777,
      "objective/train/theoretical_loss": 3.603937939286295,
      "objective/train/tokens_used": 1162421728,
      "theoretical_loss": 3.603937939286295,
      "tokens_seen": 1141961728
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00033031093279839517,
      "loss": 2.6734,
      "theoretical_loss": 3.6039235374462795,
      "tokens_seen": 1142010880
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003303009027081244,
      "loss": 2.6518,
      "theoretical_loss": 3.6039043362270746,
      "tokens_seen": 1142076416
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003302908726178536,
      "loss": 2.6918,
      "theoretical_loss": 3.6038851364181568,
      "tokens_seen": 1142141952
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00033028084252758276,
      "loss": 2.8456,
      "theoretical_loss": 3.603865938019341,
      "tokens_seen": 1142207488
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00033027081243731195,
      "loss": 2.6436,
      "theoretical_loss": 3.603846741030445,
      "tokens_seen": 1142273024
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003302607823470412,
      "loss": 2.6678,
      "theoretical_loss": 3.603827545451282,
      "tokens_seen": 1142338560
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003302507522567703,
      "loss": 2.7732,
      "theoretical_loss": 3.6038083512816685,
      "tokens_seen": 1142404096
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00033024072216649954,
      "loss": 2.7161,
      "theoretical_loss": 3.6037891585214203,
      "tokens_seen": 1142469632
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00033023069207622867,
      "loss": 2.7033,
      "theoretical_loss": 3.6037699671703534,
      "tokens_seen": 1142535168
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003302206619859579,
      "loss": 2.7151,
      "theoretical_loss": 3.6037507772282824,
      "tokens_seen": 1142600704
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003302106318956871,
      "loss": 2.7908,
      "theoretical_loss": 3.603731588695024,
      "tokens_seen": 1142666240
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00033020060180541627,
      "loss": 2.8021,
      "theoretical_loss": 3.6037124015703936,
      "tokens_seen": 1142731776
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00033019057171514545,
      "loss": 2.6327,
      "theoretical_loss": 3.603693215854207,
      "tokens_seen": 1142797312
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00033018054162487463,
      "loss": 2.7521,
      "theoretical_loss": 3.6036740315462805,
      "tokens_seen": 1142862848
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003301705115346038,
      "loss": 2.6045,
      "theoretical_loss": 3.6036548486464293,
      "tokens_seen": 1142928384
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00033016048144433305,
      "loss": 2.6104,
      "theoretical_loss": 3.60363566715447,
      "tokens_seen": 1142993920
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003301504513540622,
      "loss": 2.602,
      "theoretical_loss": 3.603616487070218,
      "tokens_seen": 1143059456
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003301404212637914,
      "loss": 2.6504,
      "theoretical_loss": 3.6035973083934896,
      "tokens_seen": 1143124992
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00033013039117352054,
      "loss": 2.6241,
      "theoretical_loss": 3.6035781311241006,
      "tokens_seen": 1143190528
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00033012036108324977,
      "loss": 2.4047,
      "theoretical_loss": 3.6035589552618674,
      "tokens_seen": 1143256064
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00033011033099297895,
      "loss": 2.6619,
      "theoretical_loss": 3.6035397808066056,
      "tokens_seen": 1143321600
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00033010030090270813,
      "loss": 2.7529,
      "theoretical_loss": 3.6035206077581314,
      "tokens_seen": 1143387136
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003300902708124373,
      "loss": 2.5172,
      "theoretical_loss": 3.603501436116262,
      "tokens_seen": 1143452672
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00033008024072216655,
      "loss": 2.3102,
      "theoretical_loss": 3.603482265880812,
      "tokens_seen": 1143518208
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003300702106318957,
      "loss": 2.8067,
      "theoretical_loss": 3.6034630970515984,
      "tokens_seen": 1143583744
    },
    {
      "epoch": 3.08,
      "objective/train/docs_used": 1266087,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.628338098526001,
      "objective/train/theoretical_loss": 3.603458305063998,
      "objective/train/tokens_used": 1164060128,
      "theoretical_loss": 3.603458305063998,
      "tokens_seen": 1143600128
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003300601805416249,
      "loss": 2.8557,
      "theoretical_loss": 3.6034439296284377,
      "tokens_seen": 1143649280
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00033005015045135404,
      "loss": 2.7171,
      "theoretical_loss": 3.6034247636111463,
      "tokens_seen": 1143714816
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003300401203610833,
      "loss": 2.6378,
      "theoretical_loss": 3.6034055989995393,
      "tokens_seen": 1143780352
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00033003009027081245,
      "loss": 2.5898,
      "theoretical_loss": 3.6033864357934347,
      "tokens_seen": 1143845888
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00033002006018054164,
      "loss": 2.6563,
      "theoretical_loss": 3.603367273992648,
      "tokens_seen": 1143911424
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003300100300902708,
      "loss": 2.9048,
      "theoretical_loss": 3.603348113596996,
      "tokens_seen": 1143976960
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00033,
      "loss": 2.9012,
      "theoretical_loss": 3.6033289546062948,
      "tokens_seen": 1144042496
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003299899699097292,
      "loss": 2.437,
      "theoretical_loss": 3.603309797020361,
      "tokens_seen": 1144108032
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003299799398194584,
      "loss": 2.7824,
      "theoretical_loss": 3.6032906408390115,
      "tokens_seen": 1144173568
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032996990972918754,
      "loss": 2.5932,
      "theoretical_loss": 3.603271486062062,
      "tokens_seen": 1144239104
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003299598796389168,
      "loss": 2.8508,
      "theoretical_loss": 3.60325233268933,
      "tokens_seen": 1144304640
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003299498495486459,
      "loss": 2.912,
      "theoretical_loss": 3.6032331807206326,
      "tokens_seen": 1144370176
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032993981945837514,
      "loss": 2.7717,
      "theoretical_loss": 3.603214030155785,
      "tokens_seen": 1144435712
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003299297893681043,
      "loss": 2.4534,
      "theoretical_loss": 3.6031948809946046,
      "tokens_seen": 1144501248
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003299197592778335,
      "loss": 2.6931,
      "theoretical_loss": 3.6031757332369088,
      "tokens_seen": 1144566784
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003299097291875627,
      "loss": 2.7032,
      "theoretical_loss": 3.603156586882513,
      "tokens_seen": 1144632320
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003298996990972919,
      "loss": 2.4714,
      "theoretical_loss": 3.6031374419312354,
      "tokens_seen": 1144697856
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032988966900702104,
      "loss": 2.7777,
      "theoretical_loss": 3.603118298382892,
      "tokens_seen": 1144763392
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003298796389167503,
      "loss": 2.6707,
      "theoretical_loss": 3.6030991562373003,
      "tokens_seen": 1144828928
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003298696088264794,
      "loss": 2.808,
      "theoretical_loss": 3.6030800154942764,
      "tokens_seen": 1144894464
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032985957873620864,
      "loss": 2.8968,
      "theoretical_loss": 3.603060876153638,
      "tokens_seen": 1144960000
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003298495486459378,
      "loss": 2.7766,
      "theoretical_loss": 3.603041738215202,
      "tokens_seen": 1145025536
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.000329839518555667,
      "loss": 2.5747,
      "theoretical_loss": 3.6030226016787847,
      "tokens_seen": 1145091072
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003298294884653962,
      "loss": 2.5625,
      "theoretical_loss": 3.603003466544204,
      "tokens_seen": 1145156608
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032981945837512537,
      "loss": 2.6457,
      "theoretical_loss": 3.6029843328112765,
      "tokens_seen": 1145222144
    },
    {
      "epoch": 3.08,
      "objective/train/docs_used": 1266852,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9177322387695312,
      "objective/train/theoretical_loss": 3.602979549597032,
      "objective/train/tokens_used": 1165698528,
      "theoretical_loss": 3.602979549597032,
      "tokens_seen": 1145238528
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032980942828485455,
      "loss": 2.7197,
      "theoretical_loss": 3.6029652004798196,
      "tokens_seen": 1145287680
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003297993981945838,
      "loss": 2.5158,
      "theoretical_loss": 3.6029460695496507,
      "tokens_seen": 1145353216
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003297893681043129,
      "loss": 2.5897,
      "theoretical_loss": 3.6029269400205868,
      "tokens_seen": 1145418752
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032977933801404215,
      "loss": 2.7719,
      "theoretical_loss": 3.602907811892444,
      "tokens_seen": 1145484288
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003297693079237713,
      "loss": 2.9219,
      "theoretical_loss": 3.6028886851650412,
      "tokens_seen": 1145549824
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003297592778335005,
      "loss": 2.6883,
      "theoretical_loss": 3.6028695598381955,
      "tokens_seen": 1145615360
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003297492477432297,
      "loss": 2.7103,
      "theoretical_loss": 3.6028504359117233,
      "tokens_seen": 1145680896
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032973921765295887,
      "loss": 2.7711,
      "theoretical_loss": 3.6028313133854426,
      "tokens_seen": 1145746432
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032972918756268805,
      "loss": 2.6575,
      "theoretical_loss": 3.6028121922591705,
      "tokens_seen": 1145811968
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003297191574724173,
      "loss": 2.783,
      "theoretical_loss": 3.602793072532725,
      "tokens_seen": 1145877504
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003297091273821464,
      "loss": 2.6168,
      "theoretical_loss": 3.6027739542059223,
      "tokens_seen": 1145943040
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032969909729187565,
      "loss": 2.8734,
      "theoretical_loss": 3.6027548372785816,
      "tokens_seen": 1146008576
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003296890672016048,
      "loss": 2.8027,
      "theoretical_loss": 3.60273572175052,
      "tokens_seen": 1146074112
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.000329679037111334,
      "loss": 2.6679,
      "theoretical_loss": 3.6027166076215535,
      "tokens_seen": 1146139648
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003296690070210632,
      "loss": 2.7966,
      "theoretical_loss": 3.6026974948915016,
      "tokens_seen": 1146205184
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003296589769307924,
      "loss": 2.6869,
      "theoretical_loss": 3.602678383560181,
      "tokens_seen": 1146270720
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032964894684052155,
      "loss": 2.7343,
      "theoretical_loss": 3.6026592736274097,
      "tokens_seen": 1146336256
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032963891675025074,
      "loss": 2.6343,
      "theoretical_loss": 3.602640165093005,
      "tokens_seen": 1146401792
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003296288866599799,
      "loss": 2.9379,
      "theoretical_loss": 3.6026210579567852,
      "tokens_seen": 1146467328
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032961885656970915,
      "loss": 2.5409,
      "theoretical_loss": 3.6026019522185684,
      "tokens_seen": 1146532864
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003296088264794383,
      "loss": 2.711,
      "theoretical_loss": 3.602582847878171,
      "tokens_seen": 1146598400
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003295987963891675,
      "loss": 2.8479,
      "theoretical_loss": 3.6025637449354124,
      "tokens_seen": 1146663936
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003295887662988967,
      "loss": 2.7248,
      "theoretical_loss": 3.6025446433901096,
      "tokens_seen": 1146729472
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003295787362086259,
      "loss": 2.7041,
      "theoretical_loss": 3.6025255432420806,
      "tokens_seen": 1146795008
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003295687061183551,
      "loss": 2.8072,
      "theoretical_loss": 3.6025064444911434,
      "tokens_seen": 1146860544
    },
    {
      "debugging/Self-BLEU-5": 0.4473570585925922,
      "debugging/distinct-1-grams": 0.7971195468618069,
      "debugging/distinct-2-grams": 0.9608189858650149,
      "debugging/entropy-1-grams": 5.677884863520164,
      "debugging/entropy-2-grams": 6.546455298638856,
      "debugging/length": 517.5,
      "debugging/num_segments": 10,
      "debugging/score": 0.004519826938251854,
      "debugging/score_std": 0.004873429920306367,
      "epoch": 3.08,
      "objective/train/docs_used": 1267667,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.091930866241455,
      "objective/train/theoretical_loss": 3.602501670021683,
      "objective/train/tokens_used": 1167336928,
      "theoretical_loss": 3.602501670021683,
      "tokens_seen": 1146876928
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032955867602808424,
      "loss": 2.711,
      "theoretical_loss": 3.6024873471371155,
      "tokens_seen": 1146926080
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003295486459378135,
      "loss": 2.6344,
      "theoretical_loss": 3.602468251179816,
      "tokens_seen": 1146991616
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032953861584754265,
      "loss": 2.6278,
      "theoretical_loss": 3.6024491566190626,
      "tokens_seen": 1147057152
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032952858575727184,
      "loss": 2.5794,
      "theoretical_loss": 3.6024300634546726,
      "tokens_seen": 1147122688
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.000329518555667001,
      "loss": 2.5395,
      "theoretical_loss": 3.6024109716864654,
      "tokens_seen": 1147188224
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003295085255767302,
      "loss": 2.5811,
      "theoretical_loss": 3.602391881314258,
      "tokens_seen": 1147253760
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003294984954864594,
      "loss": 2.8294,
      "theoretical_loss": 3.602372792337869,
      "tokens_seen": 1147319296
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003294884653961886,
      "loss": 2.4544,
      "theoretical_loss": 3.6023537047571166,
      "tokens_seen": 1147384832
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032947843530591774,
      "loss": 2.4459,
      "theoretical_loss": 3.6023346185718195,
      "tokens_seen": 1147450368
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.000329468405215647,
      "loss": 2.7234,
      "theoretical_loss": 3.6023155337817956,
      "tokens_seen": 1147515904
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003294583751253761,
      "loss": 2.7293,
      "theoretical_loss": 3.602296450386863,
      "tokens_seen": 1147581440
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032944834503510534,
      "loss": 2.6907,
      "theoretical_loss": 3.60227736838684,
      "tokens_seen": 1147646976
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003294383149448345,
      "loss": 2.7077,
      "theoretical_loss": 3.602258287781545,
      "tokens_seen": 1147712512
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003294282848545637,
      "loss": 2.979,
      "theoretical_loss": 3.6022392085707975,
      "tokens_seen": 1147778048
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003294182547642929,
      "loss": 2.687,
      "theoretical_loss": 3.602220130754415,
      "tokens_seen": 1147843584
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003294082246740221,
      "loss": 2.7995,
      "theoretical_loss": 3.602201054332216,
      "tokens_seen": 1147909120
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032939819458375124,
      "loss": 2.847,
      "theoretical_loss": 3.602181979304019,
      "tokens_seen": 1147974656
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003293881644934805,
      "loss": 2.6341,
      "theoretical_loss": 3.602162905669643,
      "tokens_seen": 1148040192
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003293781344032096,
      "loss": 2.773,
      "theoretical_loss": 3.602143833428906,
      "tokens_seen": 1148105728
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032936810431293884,
      "loss": 2.7618,
      "theoretical_loss": 3.6021247625816275,
      "tokens_seen": 1148171264
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.000329358074222668,
      "loss": 2.6518,
      "theoretical_loss": 3.6021056931276254,
      "tokens_seen": 1148236800
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003293480441323972,
      "loss": 2.7529,
      "theoretical_loss": 3.6020866250667183,
      "tokens_seen": 1148302336
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003293380140421264,
      "loss": 2.71,
      "theoretical_loss": 3.602067558398726,
      "tokens_seen": 1148367872
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032932798395185557,
      "loss": 2.8273,
      "theoretical_loss": 3.6020484931234655,
      "tokens_seen": 1148433408
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032931795386158475,
      "loss": 2.8462,
      "theoretical_loss": 3.6020294292407575,
      "tokens_seen": 1148498944
    },
    {
      "epoch": 3.08,
      "objective/train/docs_used": 1268910,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8044888973236084,
      "objective/train/theoretical_loss": 3.6020246634876454,
      "objective/train/tokens_used": 1168975328,
      "theoretical_loss": 3.6020246634876454,
      "tokens_seen": 1148515328
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.000329307923771314,
      "loss": 2.5725,
      "theoretical_loss": 3.6020103667504193,
      "tokens_seen": 1148564480
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003292978936810431,
      "loss": 2.7818,
      "theoretical_loss": 3.601991305652271,
      "tokens_seen": 1148630016
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032928786359077235,
      "loss": 2.7852,
      "theoretical_loss": 3.601972245946131,
      "tokens_seen": 1148695552
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003292778335005015,
      "loss": 2.8112,
      "theoretical_loss": 3.6019531876318176,
      "tokens_seen": 1148761088
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003292678034102307,
      "loss": 2.7196,
      "theoretical_loss": 3.60193413070915,
      "tokens_seen": 1148826624
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003292577733199599,
      "loss": 2.822,
      "theoretical_loss": 3.6019150751779483,
      "tokens_seen": 1148892160
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032924774322968907,
      "loss": 2.8823,
      "theoretical_loss": 3.601896021038031,
      "tokens_seen": 1148957696
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032923771313941825,
      "loss": 2.6873,
      "theoretical_loss": 3.6018769682892158,
      "tokens_seen": 1149023232
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003292276830491475,
      "loss": 2.7185,
      "theoretical_loss": 3.6018579169313236,
      "tokens_seen": 1149088768
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003292176529588766,
      "loss": 2.6924,
      "theoretical_loss": 3.601838866964173,
      "tokens_seen": 1149154304
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032920762286860585,
      "loss": 2.6421,
      "theoretical_loss": 3.6018198183875825,
      "tokens_seen": 1149219840
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.000329197592778335,
      "loss": 2.6667,
      "theoretical_loss": 3.6018007712013724,
      "tokens_seen": 1149285376
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003291875626880642,
      "loss": 2.5585,
      "theoretical_loss": 3.6017817254053615,
      "tokens_seen": 1149350912
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003291775325977934,
      "loss": 2.761,
      "theoretical_loss": 3.6017626809993684,
      "tokens_seen": 1149416448
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003291675025075226,
      "loss": 2.5061,
      "theoretical_loss": 3.6017436379832133,
      "tokens_seen": 1149481984
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032915747241725175,
      "loss": 2.7733,
      "theoretical_loss": 3.601724596356715,
      "tokens_seen": 1149547520
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032914744232698094,
      "loss": 2.4415,
      "theoretical_loss": 3.6017055561196933,
      "tokens_seen": 1149613056
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003291374122367101,
      "loss": 2.5899,
      "theoretical_loss": 3.6016865172719674,
      "tokens_seen": 1149678592
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032912738214643935,
      "loss": 2.5564,
      "theoretical_loss": 3.601667479813356,
      "tokens_seen": 1149744128
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003291173520561685,
      "loss": 2.7516,
      "theoretical_loss": 3.60164844374368,
      "tokens_seen": 1149809664
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003291073219658977,
      "loss": 2.7717,
      "theoretical_loss": 3.6016294090627583,
      "tokens_seen": 1149875200
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003290972918756269,
      "loss": 2.81,
      "theoretical_loss": 3.6016103757704103,
      "tokens_seen": 1149940736
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003290872617853561,
      "loss": 2.7642,
      "theoretical_loss": 3.6015913438664553,
      "tokens_seen": 1150006272
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032907723169508526,
      "loss": 2.5971,
      "theoretical_loss": 3.6015723133507134,
      "tokens_seen": 1150071808
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032906720160481444,
      "loss": 2.6199,
      "theoretical_loss": 3.6015532842230042,
      "tokens_seen": 1150137344
    },
    {
      "epoch": 3.08,
      "objective/train/docs_used": 1269509,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.0626049041748047,
      "objective/train/theoretical_loss": 3.601548527157936,
      "objective/train/tokens_used": 1170613728,
      "theoretical_loss": 3.601548527157936,
      "tokens_seen": 1150153728
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003290571715145436,
      "loss": 2.6663,
      "theoretical_loss": 3.6015342564831476,
      "tokens_seen": 1150202880
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032904714142427286,
      "loss": 2.5544,
      "theoretical_loss": 3.6015152301309623,
      "tokens_seen": 1150268416
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.000329037111334002,
      "loss": 2.7881,
      "theoretical_loss": 3.601496205166269,
      "tokens_seen": 1150333952
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003290270812437312,
      "loss": 2.8447,
      "theoretical_loss": 3.601477181588887,
      "tokens_seen": 1150399488
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032901705115346034,
      "loss": 2.8106,
      "theoretical_loss": 3.6014581593986366,
      "tokens_seen": 1150465024
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003290070210631896,
      "loss": 2.6651,
      "theoretical_loss": 3.6014391385953375,
      "tokens_seen": 1150530560
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032899699097291876,
      "loss": 2.6976,
      "theoretical_loss": 3.601420119178809,
      "tokens_seen": 1150596096
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032898696088264794,
      "loss": 2.6538,
      "theoretical_loss": 3.601401101148872,
      "tokens_seen": 1150661632
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003289769307923771,
      "loss": 2.5883,
      "theoretical_loss": 3.6013820845053455,
      "tokens_seen": 1150727168
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003289669007021063,
      "loss": 2.7749,
      "theoretical_loss": 3.60136306924805,
      "tokens_seen": 1150792704
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003289568706118355,
      "loss": 2.7311,
      "theoretical_loss": 3.6013440553768055,
      "tokens_seen": 1150858240
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003289468405215647,
      "loss": 2.7103,
      "theoretical_loss": 3.601325042891432,
      "tokens_seen": 1150923776
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032893681043129385,
      "loss": 2.8096,
      "theoretical_loss": 3.601306031791749,
      "tokens_seen": 1150989312
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003289267803410231,
      "loss": 2.6958,
      "theoretical_loss": 3.6012870220775772,
      "tokens_seen": 1151054848
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032891675025075226,
      "loss": 2.6214,
      "theoretical_loss": 3.6012680137487374,
      "tokens_seen": 1151120384
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032890672016048145,
      "loss": 2.5678,
      "theoretical_loss": 3.6012490068050482,
      "tokens_seen": 1151185920
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003288966900702106,
      "loss": 2.8512,
      "theoretical_loss": 3.6012300012463316,
      "tokens_seen": 1151251456
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003288866599799398,
      "loss": 2.8587,
      "theoretical_loss": 3.601210997072406,
      "tokens_seen": 1151316992
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.000328876629889669,
      "loss": 2.6661,
      "theoretical_loss": 3.6011919942830932,
      "tokens_seen": 1151382528
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003288665997993982,
      "loss": 2.7665,
      "theoretical_loss": 3.601172992878213,
      "tokens_seen": 1151448064
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032885656970912735,
      "loss": 2.7371,
      "theoretical_loss": 3.6011539928575855,
      "tokens_seen": 1151513600
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003288465396188566,
      "loss": 2.5777,
      "theoretical_loss": 3.601134994221032,
      "tokens_seen": 1151579136
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003288365095285857,
      "loss": 2.7746,
      "theoretical_loss": 3.601115996968371,
      "tokens_seen": 1151644672
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032882647943831495,
      "loss": 2.6474,
      "theoretical_loss": 3.6010970010994248,
      "tokens_seen": 1151710208
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003288164493480442,
      "loss": 2.5405,
      "theoretical_loss": 3.6010780066140127,
      "tokens_seen": 1151775744
    },
    {
      "epoch": 3.08,
      "objective/train/docs_used": 1270458,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7559614181518555,
      "objective/train/theoretical_loss": 3.601073258208816,
      "objective/train/tokens_used": 1172252128,
      "theoretical_loss": 3.601073258208816,
      "tokens_seen": 1151792128
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003288064192577733,
      "loss": 2.6892,
      "theoretical_loss": 3.6010590135119562,
      "tokens_seen": 1151841280
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032879638916750255,
      "loss": 2.5914,
      "theoretical_loss": 3.601040021793075,
      "tokens_seen": 1151906816
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003287863590772317,
      "loss": 2.6654,
      "theoretical_loss": 3.601021031457191,
      "tokens_seen": 1151972352
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003287763289869609,
      "loss": 2.6697,
      "theoretical_loss": 3.601002042504123,
      "tokens_seen": 1152037888
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003287662988966901,
      "loss": 2.6694,
      "theoretical_loss": 3.600983054933693,
      "tokens_seen": 1152103424
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032875626880641927,
      "loss": 2.8141,
      "theoretical_loss": 3.600964068745721,
      "tokens_seen": 1152168960
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032874623871614845,
      "loss": 2.6888,
      "theoretical_loss": 3.600945083940028,
      "tokens_seen": 1152234496
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003287362086258777,
      "loss": 2.8488,
      "theoretical_loss": 3.6009261005164346,
      "tokens_seen": 1152300032
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003287261785356068,
      "loss": 2.5548,
      "theoretical_loss": 3.600907118474762,
      "tokens_seen": 1152365568
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032871614844533605,
      "loss": 2.746,
      "theoretical_loss": 3.600888137814831,
      "tokens_seen": 1152431104
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003287061183550652,
      "loss": 2.5498,
      "theoretical_loss": 3.6008691585364616,
      "tokens_seen": 1152496640
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003286960882647944,
      "loss": 2.5718,
      "theoretical_loss": 3.6008501806394753,
      "tokens_seen": 1152562176
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003286860581745236,
      "loss": 2.8185,
      "theoretical_loss": 3.600831204123693,
      "tokens_seen": 1152627712
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003286760280842528,
      "loss": 2.7984,
      "theoretical_loss": 3.6008122289889357,
      "tokens_seen": 1152693248
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032866599799398195,
      "loss": 2.6887,
      "theoretical_loss": 3.600793255235024,
      "tokens_seen": 1152758784
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032865596790371114,
      "loss": 2.6701,
      "theoretical_loss": 3.6007742828617797,
      "tokens_seen": 1152824320
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003286459378134403,
      "loss": 2.5288,
      "theoretical_loss": 3.600755311869023,
      "tokens_seen": 1152889856
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032863590772316955,
      "loss": 2.7081,
      "theoretical_loss": 3.6007363422565755,
      "tokens_seen": 1152955392
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003286258776328987,
      "loss": 2.7438,
      "theoretical_loss": 3.600717374024258,
      "tokens_seen": 1153020928
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003286158475426279,
      "loss": 2.6553,
      "theoretical_loss": 3.600698407171892,
      "tokens_seen": 1153086464
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003286058174523571,
      "loss": 2.8296,
      "theoretical_loss": 3.600679441699299,
      "tokens_seen": 1153152000
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003285957873620863,
      "loss": 2.7507,
      "theoretical_loss": 3.6006604776062994,
      "tokens_seen": 1153217536
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032858575727181546,
      "loss": 2.7123,
      "theoretical_loss": 3.6006415148927147,
      "tokens_seen": 1153283072
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032857572718154464,
      "loss": 2.5924,
      "theoretical_loss": 3.6006225535583667,
      "tokens_seen": 1153348608
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003285656970912738,
      "loss": 2.8264,
      "theoretical_loss": 3.6006035936030756,
      "tokens_seen": 1153414144
    },
    {
      "epoch": 3.08,
      "objective/train/docs_used": 1271091,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.858920097351074,
      "objective/train/theoretical_loss": 3.60059885382971,
      "objective/train/tokens_used": 1173890528,
      "theoretical_loss": 3.60059885382971,
      "tokens_seen": 1153430528
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032855566700100306,
      "loss": 3.0625,
      "theoretical_loss": 3.6005846350266637,
      "tokens_seen": 1153479680
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003285456369107322,
      "loss": 2.8801,
      "theoretical_loss": 3.6005656778289525,
      "tokens_seen": 1153545216
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003285356068204614,
      "loss": 2.6872,
      "theoretical_loss": 3.6005467220097627,
      "tokens_seen": 1153610752
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032852557673019054,
      "loss": 2.8549,
      "theoretical_loss": 3.6005277675689165,
      "tokens_seen": 1153676288
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003285155466399198,
      "loss": 2.833,
      "theoretical_loss": 3.6005088145062345,
      "tokens_seen": 1153741824
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032850551654964896,
      "loss": 2.7275,
      "theoretical_loss": 3.600489862821539,
      "tokens_seen": 1153807360
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032849548645937814,
      "loss": 2.7133,
      "theoretical_loss": 3.600470912514651,
      "tokens_seen": 1153872896
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003284854563691073,
      "loss": 2.6046,
      "theoretical_loss": 3.6004519635853924,
      "tokens_seen": 1153938432
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003284754262788365,
      "loss": 2.4431,
      "theoretical_loss": 3.600433016033585,
      "tokens_seen": 1154003968
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003284653961885657,
      "loss": 2.5941,
      "theoretical_loss": 3.6004140698590503,
      "tokens_seen": 1154069504
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003284553660982949,
      "loss": 2.6592,
      "theoretical_loss": 3.6003951250616097,
      "tokens_seen": 1154135040
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032844533600802405,
      "loss": 2.8864,
      "theoretical_loss": 3.600376181641085,
      "tokens_seen": 1154200576
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003284353059177533,
      "loss": 2.7652,
      "theoretical_loss": 3.6003572395972987,
      "tokens_seen": 1154266112
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032842527582748246,
      "loss": 2.7161,
      "theoretical_loss": 3.6003382989300716,
      "tokens_seen": 1154331648
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032841524573721165,
      "loss": 2.7713,
      "theoretical_loss": 3.600319359639226,
      "tokens_seen": 1154397184
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003284052156469408,
      "loss": 2.6318,
      "theoretical_loss": 3.600300421724583,
      "tokens_seen": 1154462720
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032839518555667,
      "loss": 2.5957,
      "theoretical_loss": 3.600281485185966,
      "tokens_seen": 1154528256
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003283851554663992,
      "loss": 2.536,
      "theoretical_loss": 3.6002625500231957,
      "tokens_seen": 1154593792
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003283751253761284,
      "loss": 2.5904,
      "theoretical_loss": 3.6002436162360945,
      "tokens_seen": 1154659328
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032836509528585755,
      "loss": 2.5183,
      "theoretical_loss": 3.600224683824484,
      "tokens_seen": 1154724864
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003283550651955868,
      "loss": 2.7599,
      "theoretical_loss": 3.6002057527881868,
      "tokens_seen": 1154790400
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003283450351053159,
      "loss": 2.7745,
      "theoretical_loss": 3.600186823127024,
      "tokens_seen": 1154855936
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032833500501504515,
      "loss": 2.6968,
      "theoretical_loss": 3.600167894840819,
      "tokens_seen": 1154921472
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032832497492477433,
      "loss": 2.7957,
      "theoretical_loss": 3.6001489679293934,
      "tokens_seen": 1154987008
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003283149448345035,
      "loss": 2.8614,
      "theoretical_loss": 3.6001300423925686,
      "tokens_seen": 1155052544
    },
    {
      "epoch": 3.08,
      "objective/train/docs_used": 1272129,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.853865623474121,
      "objective/train/theoretical_loss": 3.600125311223123,
      "objective/train/tokens_used": 1175528928,
      "theoretical_loss": 3.600125311223123,
      "tokens_seen": 1155068928
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003283049147442327,
      "loss": 2.8174,
      "theoretical_loss": 3.600111118230168,
      "tokens_seen": 1155118080
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003282948846539619,
      "loss": 2.8958,
      "theoretical_loss": 3.600092195442013,
      "tokens_seen": 1155183616
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032828485456369105,
      "loss": 2.7905,
      "theoretical_loss": 3.6000732740279258,
      "tokens_seen": 1155249152
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003282748244734203,
      "loss": 2.7081,
      "theoretical_loss": 3.600054353987729,
      "tokens_seen": 1155314688
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003282647943831494,
      "loss": 2.413,
      "theoretical_loss": 3.600035435321245,
      "tokens_seen": 1155380224
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032825476429287865,
      "loss": 2.695,
      "theoretical_loss": 3.6000165180282955,
      "tokens_seen": 1155445760
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032824473420260783,
      "loss": 2.5683,
      "theoretical_loss": 3.5999976021087035,
      "tokens_seen": 1155511296
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.000328234704112337,
      "loss": 2.5468,
      "theoretical_loss": 3.5999786875622917,
      "tokens_seen": 1155576832
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003282246740220662,
      "loss": 2.413,
      "theoretical_loss": 3.599959774388882,
      "tokens_seen": 1155642368
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003282146439317954,
      "loss": 2.7004,
      "theoretical_loss": 3.599940862588297,
      "tokens_seen": 1155707904
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032820461384152456,
      "loss": 2.6909,
      "theoretical_loss": 3.5999219521603587,
      "tokens_seen": 1155773440
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003281945837512538,
      "loss": 2.6059,
      "theoretical_loss": 3.5999030431048906,
      "tokens_seen": 1155838976
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003281845536609829,
      "loss": 2.6892,
      "theoretical_loss": 3.599884135421715,
      "tokens_seen": 1155904512
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032817452357071215,
      "loss": 2.5906,
      "theoretical_loss": 3.5998652291106543,
      "tokens_seen": 1155970048
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003281644934804413,
      "loss": 2.7893,
      "theoretical_loss": 3.599846324171531,
      "tokens_seen": 1156035584
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003281544633901705,
      "loss": 2.6951,
      "theoretical_loss": 3.5998274206041683,
      "tokens_seen": 1156101120
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003281444332998997,
      "loss": 2.6377,
      "theoretical_loss": 3.5998085184083886,
      "tokens_seen": 1156166656
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003281344032096289,
      "loss": 2.5826,
      "theoretical_loss": 3.5997896175840145,
      "tokens_seen": 1156232192
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032812437311935806,
      "loss": 2.7148,
      "theoretical_loss": 3.5997707181308685,
      "tokens_seen": 1156297728
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003281143430290873,
      "loss": 2.7672,
      "theoretical_loss": 3.599751820048774,
      "tokens_seen": 1156363264
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003281043129388164,
      "loss": 2.8374,
      "theoretical_loss": 3.5997329233375543,
      "tokens_seen": 1156428800
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032809428284854566,
      "loss": 2.6762,
      "theoretical_loss": 3.599714027997031,
      "tokens_seen": 1156494336
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003280842527582748,
      "loss": 2.3282,
      "theoretical_loss": 3.599695134027028,
      "tokens_seen": 1156559872
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.000328074222668004,
      "loss": 2.7864,
      "theoretical_loss": 3.5996762414273675,
      "tokens_seen": 1156625408
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032806419257773326,
      "loss": 2.8162,
      "theoretical_loss": 3.599657350197873,
      "tokens_seen": 1156690944
    },
    {
      "epoch": 3.08,
      "objective/train/docs_used": 1272864,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.16133189201355,
      "objective/train/theoretical_loss": 3.5996526276045673,
      "objective/train/tokens_used": 1177167328,
      "theoretical_loss": 3.5996526276045673,
      "tokens_seen": 1156707328
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003280541624874624,
      "loss": 2.8585,
      "theoretical_loss": 3.599638460338368,
      "tokens_seen": 1156756480
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003280441323971916,
      "loss": 2.5503,
      "theoretical_loss": 3.599619571848674,
      "tokens_seen": 1156822016
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032803410230692074,
      "loss": 2.7527,
      "theoretical_loss": 3.5996006847286157,
      "tokens_seen": 1156887552
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032802407221665,
      "loss": 2.6692,
      "theoretical_loss": 3.599581798978015,
      "tokens_seen": 1156953088
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032801404212637916,
      "loss": 2.475,
      "theoretical_loss": 3.599562914596696,
      "tokens_seen": 1157018624
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032800401203610834,
      "loss": 2.62,
      "theoretical_loss": 3.5995440315844807,
      "tokens_seen": 1157084160
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003279939819458375,
      "loss": 2.8344,
      "theoretical_loss": 3.5995251499411935,
      "tokens_seen": 1157149696
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003279839518555667,
      "loss": 2.6817,
      "theoretical_loss": 3.5995062696666573,
      "tokens_seen": 1157215232
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003279739217652959,
      "loss": 2.7615,
      "theoretical_loss": 3.599487390760695,
      "tokens_seen": 1157280768
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003279638916750251,
      "loss": 2.567,
      "theoretical_loss": 3.5994685132231297,
      "tokens_seen": 1157346304
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032795386158475425,
      "loss": 2.4982,
      "theoretical_loss": 3.5994496370537856,
      "tokens_seen": 1157411840
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003279438314944835,
      "loss": 2.7333,
      "theoretical_loss": 3.599430762252486,
      "tokens_seen": 1157477376
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032793380140421266,
      "loss": 2.8056,
      "theoretical_loss": 3.599411888819054,
      "tokens_seen": 1157542912
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032792377131394185,
      "loss": 2.5637,
      "theoretical_loss": 3.599393016753312,
      "tokens_seen": 1157608448
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.000327913741223671,
      "loss": 2.7559,
      "theoretical_loss": 3.599374146055085,
      "tokens_seen": 1157673984
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003279037111334002,
      "loss": 2.7122,
      "theoretical_loss": 3.599355276724196,
      "tokens_seen": 1157739520
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003278936810431294,
      "loss": 2.7572,
      "theoretical_loss": 3.5993364087604682,
      "tokens_seen": 1157805056
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003278836509528586,
      "loss": 2.5657,
      "theoretical_loss": 3.5993175421637256,
      "tokens_seen": 1157870592
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032787362086258775,
      "loss": 2.7728,
      "theoretical_loss": 3.599298676933792,
      "tokens_seen": 1157936128
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.000327863590772317,
      "loss": 2.2984,
      "theoretical_loss": 3.59927981307049,
      "tokens_seen": 1158001664
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003278535606820461,
      "loss": 2.5368,
      "theoretical_loss": 3.599260950573644,
      "tokens_seen": 1158067200
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032784353059177535,
      "loss": 2.6498,
      "theoretical_loss": 3.599242089443078,
      "tokens_seen": 1158132736
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032783350050150453,
      "loss": 2.5435,
      "theoretical_loss": 3.599223229678615,
      "tokens_seen": 1158198272
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003278234704112337,
      "loss": 2.6249,
      "theoretical_loss": 3.5992043712800794,
      "tokens_seen": 1158263808
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003278134403209629,
      "loss": 2.8027,
      "theoretical_loss": 3.599185514247295,
      "tokens_seen": 1158329344
    },
    {
      "epoch": 3.08,
      "objective/train/docs_used": 1274329,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.622804641723633,
      "objective/train/theoretical_loss": 3.5991808002024763,
      "objective/train/tokens_used": 1178805728,
      "theoretical_loss": 3.5991808002024763,
      "tokens_seen": 1158345728
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032780341023069213,
      "loss": 2.5089,
      "theoretical_loss": 3.5991666585800846,
      "tokens_seen": 1158394880
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032779338014042125,
      "loss": 2.9317,
      "theoretical_loss": 3.599147804278273,
      "tokens_seen": 1158460416
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003277833500501505,
      "loss": 2.7415,
      "theoretical_loss": 3.599128951341684,
      "tokens_seen": 1158525952
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003277733199598796,
      "loss": 2.6007,
      "theoretical_loss": 3.5991100997701415,
      "tokens_seen": 1158591488
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032776328986960885,
      "loss": 2.5627,
      "theoretical_loss": 3.599091249563469,
      "tokens_seen": 1158657024
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032775325977933803,
      "loss": 2.5528,
      "theoretical_loss": 3.5990724007214916,
      "tokens_seen": 1158722560
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003277432296890672,
      "loss": 2.4406,
      "theoretical_loss": 3.5990535532440324,
      "tokens_seen": 1158788096
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003277331995987964,
      "loss": 2.4564,
      "theoretical_loss": 3.5990347071309152,
      "tokens_seen": 1158853632
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003277231695085256,
      "loss": 2.8616,
      "theoretical_loss": 3.5990158623819646,
      "tokens_seen": 1158919168
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032771313941825476,
      "loss": 2.6851,
      "theoretical_loss": 3.5989970189970046,
      "tokens_seen": 1158984704
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.000327703109327984,
      "loss": 2.8071,
      "theoretical_loss": 3.5989781769758595,
      "tokens_seen": 1159050240
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003276930792377131,
      "loss": 2.7617,
      "theoretical_loss": 3.5989593363183534,
      "tokens_seen": 1159115776
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032768304914744235,
      "loss": 2.6882,
      "theoretical_loss": 3.598940497024311,
      "tokens_seen": 1159181312
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003276730190571715,
      "loss": 2.7579,
      "theoretical_loss": 3.5989216590935555,
      "tokens_seen": 1159246848
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003276629889669007,
      "loss": 2.548,
      "theoretical_loss": 3.5989028225259116,
      "tokens_seen": 1159312384
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003276529588766299,
      "loss": 2.6048,
      "theoretical_loss": 3.598883987321204,
      "tokens_seen": 1159377920
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003276429287863591,
      "loss": 2.7181,
      "theoretical_loss": 3.598865153479257,
      "tokens_seen": 1159443456
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032763289869608826,
      "loss": 2.7681,
      "theoretical_loss": 3.5988463209998947,
      "tokens_seen": 1159508992
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003276228686058175,
      "loss": 2.7062,
      "theoretical_loss": 3.5988274898829413,
      "tokens_seen": 1159574528
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003276128385155466,
      "loss": 2.6341,
      "theoretical_loss": 3.598808660128222,
      "tokens_seen": 1159640064
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032760280842527586,
      "loss": 2.8696,
      "theoretical_loss": 3.5987898317355604,
      "tokens_seen": 1159705600
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.000327592778335005,
      "loss": 2.7675,
      "theoretical_loss": 3.5987710047047816,
      "tokens_seen": 1159771136
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003275827482447342,
      "loss": 2.5033,
      "theoretical_loss": 3.59875217903571,
      "tokens_seen": 1159836672
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003275727181544634,
      "loss": 2.7395,
      "theoretical_loss": 3.5987333547281706,
      "tokens_seen": 1159902208
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003275626880641926,
      "loss": 2.6689,
      "theoretical_loss": 3.598714531781987,
      "tokens_seen": 1159967744
    },
    {
      "epoch": 3.08,
      "objective/train/docs_used": 1274819,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.889496326446533,
      "objective/train/theoretical_loss": 3.5987098262581325,
      "objective/train/tokens_used": 1180444128,
      "theoretical_loss": 3.5987098262581325,
      "tokens_seen": 1159984128
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032755265797392176,
      "loss": 2.6646,
      "theoretical_loss": 3.598695710196985,
      "tokens_seen": 1160033280
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032754262788365094,
      "loss": 2.621,
      "theoretical_loss": 3.598676889972988,
      "tokens_seen": 1160098816
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003275325977933801,
      "loss": 2.7977,
      "theoretical_loss": 3.598658071109822,
      "tokens_seen": 1160164352
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032752256770310936,
      "loss": 2.6562,
      "theoretical_loss": 3.5986392536073106,
      "tokens_seen": 1160229888
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003275125376128385,
      "loss": 2.7267,
      "theoretical_loss": 3.5986204374652795,
      "tokens_seen": 1160295424
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003275025075225677,
      "loss": 2.7961,
      "theoretical_loss": 3.598601622683553,
      "tokens_seen": 1160360960
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032749247743229685,
      "loss": 2.5709,
      "theoretical_loss": 3.598582809261956,
      "tokens_seen": 1160426496
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003274824473420261,
      "loss": 2.7136,
      "theoretical_loss": 3.598563997200314,
      "tokens_seen": 1160492032
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032747241725175527,
      "loss": 2.6793,
      "theoretical_loss": 3.59854518649845,
      "tokens_seen": 1160557568
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032746238716148445,
      "loss": 2.7906,
      "theoretical_loss": 3.5985263771561913,
      "tokens_seen": 1160623104
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032745235707121363,
      "loss": 2.5507,
      "theoretical_loss": 3.5985075691733615,
      "tokens_seen": 1160688640
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032744232698094286,
      "loss": 2.6493,
      "theoretical_loss": 3.598488762549786,
      "tokens_seen": 1160754176
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.000327432296890672,
      "loss": 2.5678,
      "theoretical_loss": 3.59846995728529,
      "tokens_seen": 1160819712
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003274222668004012,
      "loss": 2.8937,
      "theoretical_loss": 3.598451153379698,
      "tokens_seen": 1160885248
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032741223671013035,
      "loss": 2.802,
      "theoretical_loss": 3.5984323508328355,
      "tokens_seen": 1160950784
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003274022066198596,
      "loss": 2.7575,
      "theoretical_loss": 3.598413549644528,
      "tokens_seen": 1161016320
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032739217652958877,
      "loss": 2.5724,
      "theoretical_loss": 3.5983947498145996,
      "tokens_seen": 1161081856
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032738214643931795,
      "loss": 2.695,
      "theoretical_loss": 3.5983759513428764,
      "tokens_seen": 1161147392
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032737211634904713,
      "loss": 2.7675,
      "theoretical_loss": 3.5983571542291832,
      "tokens_seen": 1161212928
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003273620862587763,
      "loss": 2.6807,
      "theoretical_loss": 3.5983383584733453,
      "tokens_seen": 1161278464
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003273520561685055,
      "loss": 2.8354,
      "theoretical_loss": 3.5983195640751884,
      "tokens_seen": 1161344000
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032734202607823473,
      "loss": 2.6488,
      "theoretical_loss": 3.598300771034537,
      "tokens_seen": 1161409536
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032733199598796386,
      "loss": 2.4031,
      "theoretical_loss": 3.598281979351217,
      "tokens_seen": 1161475072
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003273219658976931,
      "loss": 2.4916,
      "theoretical_loss": 3.5982631890250545,
      "tokens_seen": 1161540608
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032731193580742233,
      "loss": 2.7317,
      "theoretical_loss": 3.5982444000558735,
      "tokens_seen": 1161606144
    },
    {
      "epoch": 3.08,
      "objective/train/docs_used": 1275406,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6565632820129395,
      "objective/train/theoretical_loss": 3.5982397030255866,
      "objective/train/tokens_used": 1182082528,
      "theoretical_loss": 3.5982397030255866,
      "tokens_seen": 1161622528
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032730190571715145,
      "loss": 2.3889,
      "theoretical_loss": 3.5982256124435006,
      "tokens_seen": 1161671680
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003272918756268807,
      "loss": 2.7799,
      "theoretical_loss": 3.59820682618776,
      "tokens_seen": 1161737216
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003272818455366098,
      "loss": 2.7662,
      "theoretical_loss": 3.598188041288479,
      "tokens_seen": 1161802752
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032727181544633905,
      "loss": 2.337,
      "theoretical_loss": 3.5981692577454814,
      "tokens_seen": 1161868288
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032726178535606823,
      "loss": 2.8643,
      "theoretical_loss": 3.598150475558594,
      "tokens_seen": 1161933824
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003272517552657974,
      "loss": 2.6224,
      "theoretical_loss": 3.5981316947276416,
      "tokens_seen": 1161999360
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003272417251755266,
      "loss": 2.4621,
      "theoretical_loss": 3.5981129152524503,
      "tokens_seen": 1162064896
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003272316950852558,
      "loss": 2.322,
      "theoretical_loss": 3.598094137132846,
      "tokens_seen": 1162130432
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032722166499498496,
      "loss": 2.8936,
      "theoretical_loss": 3.5980753603686537,
      "tokens_seen": 1162195968
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003272116349047142,
      "loss": 2.8057,
      "theoretical_loss": 3.5980565849597,
      "tokens_seen": 1162261504
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003272016048144433,
      "loss": 2.7092,
      "theoretical_loss": 3.5980378109058098,
      "tokens_seen": 1162327040
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032719157472417256,
      "loss": 2.6417,
      "theoretical_loss": 3.598019038206809,
      "tokens_seen": 1162392576
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003271815446339017,
      "loss": 2.4224,
      "theoretical_loss": 3.598000266862525,
      "tokens_seen": 1162458112
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003271715145436309,
      "loss": 2.4214,
      "theoretical_loss": 3.597981496872781,
      "tokens_seen": 1162523648
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003271614844533601,
      "loss": 2.5387,
      "theoretical_loss": 3.5979627282374054,
      "tokens_seen": 1162589184
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003271514543630893,
      "loss": 2.6946,
      "theoretical_loss": 3.5979439609562225,
      "tokens_seen": 1162654720
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032714142427281846,
      "loss": 2.6044,
      "theoretical_loss": 3.5979251950290587,
      "tokens_seen": 1162720256
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003271313941825477,
      "loss": 2.516,
      "theoretical_loss": 3.5979064304557404,
      "tokens_seen": 1162785792
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003271213640922768,
      "loss": 2.5136,
      "theoretical_loss": 3.5978876672360935,
      "tokens_seen": 1162851328
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032711133400200606,
      "loss": 2.625,
      "theoretical_loss": 3.597868905369944,
      "tokens_seen": 1162916864
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003271013039117352,
      "loss": 2.4939,
      "theoretical_loss": 3.5978501448571176,
      "tokens_seen": 1162982400
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003270912738214644,
      "loss": 2.684,
      "theoretical_loss": 3.5978313856974413,
      "tokens_seen": 1163047936
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003270812437311936,
      "loss": 2.8434,
      "theoretical_loss": 3.59781262789074,
      "tokens_seen": 1163113472
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003270712136409228,
      "loss": 2.5827,
      "theoretical_loss": 3.597793871436841,
      "tokens_seen": 1163179008
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032706118355065196,
      "loss": 2.7304,
      "theoretical_loss": 3.59777511633557,
      "tokens_seen": 1163244544
    },
    {
      "epoch": 3.08,
      "objective/train/docs_used": 1276640,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.588357448577881,
      "objective/train/theoretical_loss": 3.5977704277715805,
      "objective/train/tokens_used": 1183720928,
      "theoretical_loss": 3.5977704277715805,
      "tokens_seen": 1163260928
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032705115346038114,
      "loss": 2.678,
      "theoretical_loss": 3.597756362586754,
      "tokens_seen": 1163310080
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003270411233701103,
      "loss": 2.7792,
      "theoretical_loss": 3.597737610190218,
      "tokens_seen": 1163375616
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032703109327983956,
      "loss": 2.67,
      "theoretical_loss": 3.597718859145789,
      "tokens_seen": 1163441152
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003270210631895687,
      "loss": 2.6397,
      "theoretical_loss": 3.5977001094532937,
      "tokens_seen": 1163506688
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003270110330992979,
      "loss": 2.4046,
      "theoretical_loss": 3.597681361112558,
      "tokens_seen": 1163572224
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032700100300902705,
      "loss": 2.55,
      "theoretical_loss": 3.5976626141234087,
      "tokens_seen": 1163637760
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003269909729187563,
      "loss": 2.4804,
      "theoretical_loss": 3.5976438684856715,
      "tokens_seen": 1163703296
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032698094282848547,
      "loss": 2.6268,
      "theoretical_loss": 3.597625124199174,
      "tokens_seen": 1163768832
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032697091273821465,
      "loss": 2.5692,
      "theoretical_loss": 3.5976063812637418,
      "tokens_seen": 1163834368
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032696088264794383,
      "loss": 2.5237,
      "theoretical_loss": 3.5975876396792015,
      "tokens_seen": 1163899904
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032695085255767306,
      "loss": 2.6462,
      "theoretical_loss": 3.5975688994453803,
      "tokens_seen": 1163965440
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003269408224674022,
      "loss": 2.7789,
      "theoretical_loss": 3.597550160562104,
      "tokens_seen": 1164030976
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003269307923771314,
      "loss": 2.5586,
      "theoretical_loss": 3.5975314230292,
      "tokens_seen": 1164096512
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032692076228686055,
      "loss": 2.6079,
      "theoretical_loss": 3.5975126868464953,
      "tokens_seen": 1164162048
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003269107321965898,
      "loss": 2.5372,
      "theoretical_loss": 3.5974939520138154,
      "tokens_seen": 1164227584
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032690070210631897,
      "loss": 2.4518,
      "theoretical_loss": 3.5974752185309873,
      "tokens_seen": 1164293120
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032689067201604815,
      "loss": 2.4755,
      "theoretical_loss": 3.5974564863978387,
      "tokens_seen": 1164358656
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032688064192577733,
      "loss": 2.6264,
      "theoretical_loss": 3.5974377556141954,
      "tokens_seen": 1164424192
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003268706118355065,
      "loss": 2.4069,
      "theoretical_loss": 3.5974190261798844,
      "tokens_seen": 1164489728
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003268605817452357,
      "loss": 2.5928,
      "theoretical_loss": 3.5974002980947333,
      "tokens_seen": 1164555264
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032685055165496493,
      "loss": 2.5782,
      "theoretical_loss": 3.597381571358568,
      "tokens_seen": 1164620800
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032684052156469406,
      "loss": 2.4116,
      "theoretical_loss": 3.597362845971216,
      "tokens_seen": 1164686336
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003268304914744233,
      "loss": 2.8265,
      "theoretical_loss": 3.597344121932504,
      "tokens_seen": 1164751872
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003268204613841524,
      "loss": 2.7104,
      "theoretical_loss": 3.59732539924226,
      "tokens_seen": 1164817408
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032681043129388165,
      "loss": 2.5426,
      "theoretical_loss": 3.5973066779003093,
      "tokens_seen": 1164882944
    },
    {
      "epoch": 3.08,
      "objective/train/docs_used": 1277429,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5294029712677,
      "objective/train/theoretical_loss": 3.5973019977754723,
      "objective/train/tokens_used": 1185359328,
      "theoretical_loss": 3.5973019977754723,
      "tokens_seen": 1164899328
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032680040120361084,
      "loss": 2.485,
      "theoretical_loss": 3.59728795790648,
      "tokens_seen": 1164948480
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032679037111334,
      "loss": 2.661,
      "theoretical_loss": 3.5972692392605996,
      "tokens_seen": 1165014016
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003267803410230692,
      "loss": 2.813,
      "theoretical_loss": 3.5972505219624944,
      "tokens_seen": 1165079552
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032677031093279843,
      "loss": 2.4742,
      "theoretical_loss": 3.597231806011991,
      "tokens_seen": 1165145088
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032676028084252756,
      "loss": 2.6719,
      "theoretical_loss": 3.5972130914089187,
      "tokens_seen": 1165210624
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003267502507522568,
      "loss": 2.5704,
      "theoretical_loss": 3.5971943781531026,
      "tokens_seen": 1165276160
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003267402206619859,
      "loss": 2.6898,
      "theoretical_loss": 3.597175666244371,
      "tokens_seen": 1165341696
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032673019057171516,
      "loss": 2.3992,
      "theoretical_loss": 3.597156955682551,
      "tokens_seen": 1165407232
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032672016048144434,
      "loss": 2.4696,
      "theoretical_loss": 3.5971382464674697,
      "tokens_seen": 1165472768
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003267101303911735,
      "loss": 2.7106,
      "theoretical_loss": 3.5971195385989545,
      "tokens_seen": 1165538304
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003267001003009027,
      "loss": 2.4775,
      "theoretical_loss": 3.597100832076833,
      "tokens_seen": 1165603840
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003266900702106319,
      "loss": 2.4111,
      "theoretical_loss": 3.5970821269009328,
      "tokens_seen": 1165669376
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032668004012036106,
      "loss": 2.6574,
      "theoretical_loss": 3.5970634230710803,
      "tokens_seen": 1165734912
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003266700100300903,
      "loss": 2.5731,
      "theoretical_loss": 3.597044720587104,
      "tokens_seen": 1165800448
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003266599799398194,
      "loss": 2.7557,
      "theoretical_loss": 3.5970260194488315,
      "tokens_seen": 1165865984
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032664994984954866,
      "loss": 2.6017,
      "theoretical_loss": 3.5970073196560897,
      "tokens_seen": 1165931520
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032663991975927784,
      "loss": 2.6644,
      "theoretical_loss": 3.5969886212087063,
      "tokens_seen": 1165997056
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.000326629889669007,
      "loss": 2.6361,
      "theoretical_loss": 3.596969924106509,
      "tokens_seen": 1166062592
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003266198595787362,
      "loss": 2.7494,
      "theoretical_loss": 3.596951228349325,
      "tokens_seen": 1166128128
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003266098294884654,
      "loss": 2.7595,
      "theoretical_loss": 3.5969325339369833,
      "tokens_seen": 1166193664
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032659979939819457,
      "loss": 2.6563,
      "theoretical_loss": 3.59691384086931,
      "tokens_seen": 1166259200
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003265897693079238,
      "loss": 2.7333,
      "theoretical_loss": 3.5968951491461336,
      "tokens_seen": 1166324736
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032657973921765293,
      "loss": 2.7429,
      "theoretical_loss": 3.596876458767282,
      "tokens_seen": 1166390272
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032656970912738216,
      "loss": 2.4125,
      "theoretical_loss": 3.596857769732583,
      "tokens_seen": 1166455808
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032655967903711135,
      "loss": 2.5437,
      "theoretical_loss": 3.596839082041864,
      "tokens_seen": 1166521344
    },
    {
      "epoch": 3.08,
      "objective/train/docs_used": 1282424,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.0159366130828857,
      "objective/train/theoretical_loss": 3.5968344103291607,
      "objective/train/tokens_used": 1186997728,
      "theoretical_loss": 3.5968344103291607,
      "tokens_seen": 1166537728
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003265496489468405,
      "loss": 2.6358,
      "theoretical_loss": 3.596820395694953,
      "tokens_seen": 1166586880
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032653961885656976,
      "loss": 2.3213,
      "theoretical_loss": 3.5968017106916776,
      "tokens_seen": 1166652416
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003265295887662989,
      "loss": 2.6772,
      "theoretical_loss": 3.5967830270318664,
      "tokens_seen": 1166717952
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003265195586760281,
      "loss": 2.7451,
      "theoretical_loss": 3.596764344715347,
      "tokens_seen": 1166783488
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032650952858575725,
      "loss": 2.7173,
      "theoretical_loss": 3.596745663741947,
      "tokens_seen": 1166849024
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003264994984954865,
      "loss": 2.751,
      "theoretical_loss": 3.596726984111495,
      "tokens_seen": 1166914560
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032648946840521567,
      "loss": 2.6119,
      "theoretical_loss": 3.5967083058238187,
      "tokens_seen": 1166980096
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032647943831494485,
      "loss": 2.7055,
      "theoretical_loss": 3.5966896288787464,
      "tokens_seen": 1167045632
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032646940822467403,
      "loss": 2.4983,
      "theoretical_loss": 3.596670953276106,
      "tokens_seen": 1167111168
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032645937813440326,
      "loss": 2.6833,
      "theoretical_loss": 3.596652279015726,
      "tokens_seen": 1167176704
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003264493480441324,
      "loss": 2.5952,
      "theoretical_loss": 3.596633606097434,
      "tokens_seen": 1167242240
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003264393179538616,
      "loss": 2.6169,
      "theoretical_loss": 3.596614934521059,
      "tokens_seen": 1167307776
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032642928786359075,
      "loss": 2.6897,
      "theoretical_loss": 3.596596264286428,
      "tokens_seen": 1167373312
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032641925777332,
      "loss": 2.4475,
      "theoretical_loss": 3.596577595393371,
      "tokens_seen": 1167438848
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032640922768304917,
      "loss": 2.7334,
      "theoretical_loss": 3.5965589278417145,
      "tokens_seen": 1167504384
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032639919759277835,
      "loss": 2.5306,
      "theoretical_loss": 3.596540261631288,
      "tokens_seen": 1167569920
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032638916750250753,
      "loss": 2.5529,
      "theoretical_loss": 3.596521596761919,
      "tokens_seen": 1167635456
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003263791374122367,
      "loss": 2.9282,
      "theoretical_loss": 3.5965029332334373,
      "tokens_seen": 1167700992
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003263691073219659,
      "loss": 2.487,
      "theoretical_loss": 3.5964842710456697,
      "tokens_seen": 1167766528
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032635907723169513,
      "loss": 2.6882,
      "theoretical_loss": 3.5964656101984454,
      "tokens_seen": 1167832064
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032634904714142426,
      "loss": 2.8419,
      "theoretical_loss": 3.5964469506915933,
      "tokens_seen": 1167897600
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003263390170511535,
      "loss": 2.5661,
      "theoretical_loss": 3.5964282925249407,
      "tokens_seen": 1167963136
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003263289869608826,
      "loss": 2.821,
      "theoretical_loss": 3.5964096356983175,
      "tokens_seen": 1168028672
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032631895687061185,
      "loss": 2.7562,
      "theoretical_loss": 3.5963909802115515,
      "tokens_seen": 1168094208
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032630892678034104,
      "loss": 2.3968,
      "theoretical_loss": 3.5963723260644715,
      "tokens_seen": 1168159744
    },
    {
      "epoch": 3.08,
      "objective/train/docs_used": 1282424,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.3768837451934814,
      "objective/train/theoretical_loss": 3.5963676627370074,
      "objective/train/tokens_used": 1187029472,
      "theoretical_loss": 3.5963676627370074,
      "tokens_seen": 1168176128
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003262988966900702,
      "loss": 2.6501,
      "theoretical_loss": 3.596353673256906,
      "tokens_seen": 1168225280
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003262888665997994,
      "loss": 2.4647,
      "theoretical_loss": 3.5963350217886845,
      "tokens_seen": 1168290816
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032627883650952863,
      "loss": 2.7632,
      "theoretical_loss": 3.5963163716596345,
      "tokens_seen": 1168356352
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032626880641925776,
      "loss": 2.441,
      "theoretical_loss": 3.5962977228695854,
      "tokens_seen": 1168421888
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.000326258776328987,
      "loss": 2.7156,
      "theoretical_loss": 3.596279075418366,
      "tokens_seen": 1168487424
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003262487462387161,
      "loss": 2.5267,
      "theoretical_loss": 3.5962604293058043,
      "tokens_seen": 1168552960
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032623871614844536,
      "loss": 2.628,
      "theoretical_loss": 3.5962417845317303,
      "tokens_seen": 1168618496
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032622868605817454,
      "loss": 2.5469,
      "theoretical_loss": 3.596223141095972,
      "tokens_seen": 1168684032
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003262186559679037,
      "loss": 2.6036,
      "theoretical_loss": 3.5962044989983593,
      "tokens_seen": 1168749568
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003262086258776329,
      "loss": 2.3583,
      "theoretical_loss": 3.5961858582387194,
      "tokens_seen": 1168815104
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003261985957873621,
      "loss": 2.6966,
      "theoretical_loss": 3.596167218816883,
      "tokens_seen": 1168880640
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032618856569709126,
      "loss": 2.6841,
      "theoretical_loss": 3.5961485807326783,
      "tokens_seen": 1168946176
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003261785356068205,
      "loss": 2.6477,
      "theoretical_loss": 3.5961299439859347,
      "tokens_seen": 1169011712
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003261685055165496,
      "loss": 2.7623,
      "theoretical_loss": 3.596111308576481,
      "tokens_seen": 1169077248
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032615847542627886,
      "loss": 2.5495,
      "theoretical_loss": 3.5960926745041455,
      "tokens_seen": 1169142784
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032614844533600804,
      "loss": 2.7324,
      "theoretical_loss": 3.596074041768759,
      "tokens_seen": 1169208320
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003261384152457372,
      "loss": 2.5712,
      "theoretical_loss": 3.596055410370149,
      "tokens_seen": 1169273856
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003261283851554664,
      "loss": 2.6998,
      "theoretical_loss": 3.596036780308146,
      "tokens_seen": 1169339392
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003261183550651956,
      "loss": 2.6135,
      "theoretical_loss": 3.5960181515825784,
      "tokens_seen": 1169404928
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032610832497492477,
      "loss": 2.577,
      "theoretical_loss": 3.595999524193276,
      "tokens_seen": 1169470464
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.000326098294884654,
      "loss": 2.8685,
      "theoretical_loss": 3.595980898140067,
      "tokens_seen": 1169536000
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032608826479438313,
      "loss": 2.5749,
      "theoretical_loss": 3.5959622734227823,
      "tokens_seen": 1169601536
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032607823470411236,
      "loss": 2.7083,
      "theoretical_loss": 3.5959436500412494,
      "tokens_seen": 1169667072
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003260682046138415,
      "loss": 2.5683,
      "theoretical_loss": 3.5959250279952997,
      "tokens_seen": 1169732608
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003260581745235707,
      "loss": 2.7942,
      "theoretical_loss": 3.595906407284761,
      "tokens_seen": 1169798144
    },
    {
      "epoch": 3.08,
      "objective/train/docs_used": 1282424,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.560084581375122,
      "objective/train/theoretical_loss": 3.595901752315764,
      "objective/train/tokens_used": 1187029472,
      "theoretical_loss": 3.595901752315764,
      "tokens_seen": 1169814528
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003260481444332999,
      "loss": 2.7187,
      "theoretical_loss": 3.595887787909463,
      "tokens_seen": 1169863680
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003260381143430291,
      "loss": 2.909,
      "theoretical_loss": 3.5958691698692355,
      "tokens_seen": 1169929216
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032602808425275827,
      "loss": 2.7557,
      "theoretical_loss": 3.595850553163908,
      "tokens_seen": 1169994752
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032601805416248745,
      "loss": 2.5457,
      "theoretical_loss": 3.59583193779331,
      "tokens_seen": 1170060288
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032600802407221663,
      "loss": 2.6619,
      "theoretical_loss": 3.5958133237572714,
      "tokens_seen": 1170125824
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032599799398194587,
      "loss": 2.5647,
      "theoretical_loss": 3.595794711055621,
      "tokens_seen": 1170191360
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.000325987963891675,
      "loss": 2.7163,
      "theoretical_loss": 3.595776099688189,
      "tokens_seen": 1170256896
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032597793380140423,
      "loss": 2.4849,
      "theoretical_loss": 3.5957574896548046,
      "tokens_seen": 1170322432
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003259679037111334,
      "loss": 2.8802,
      "theoretical_loss": 3.595738880955298,
      "tokens_seen": 1170387968
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003259578736208626,
      "loss": 2.296,
      "theoretical_loss": 3.5957202735894986,
      "tokens_seen": 1170453504
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003259478435305918,
      "loss": 2.6945,
      "theoretical_loss": 3.595701667557236,
      "tokens_seen": 1170519040
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032593781344032095,
      "loss": 2.6302,
      "theoretical_loss": 3.5956830628583405,
      "tokens_seen": 1170584576
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032592778335005014,
      "loss": 2.4871,
      "theoretical_loss": 3.5956644594926415,
      "tokens_seen": 1170650112
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.00032591775325977937,
      "loss": 2.6558,
      "theoretical_loss": 3.5956458574599686,
      "tokens_seen": 1170715648
    },
    {
      "epoch": 3.08,
      "learning_rate": 0.0003259077231695085,
      "loss": 2.5413,
      "theoretical_loss": 3.595627256760152,
      "tokens_seen": 1170781184
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.00032589769307923773,
      "loss": 2.4552,
      "theoretical_loss": 3.5956086573930217,
      "tokens_seen": 1170846720
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.00032588766298896686,
      "loss": 2.4495,
      "theoretical_loss": 3.5955900593584076,
      "tokens_seen": 1170912256
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.0003258776328986961,
      "loss": 2.7012,
      "theoretical_loss": 3.5955714626561397,
      "tokens_seen": 1170977792
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.0003258676028084253,
      "loss": 2.4193,
      "theoretical_loss": 3.5955528672860475,
      "tokens_seen": 1171043328
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.00032585757271815446,
      "loss": 2.7626,
      "theoretical_loss": 3.5955342732479614,
      "tokens_seen": 1171108864
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.00032584754262788364,
      "loss": 2.6797,
      "theoretical_loss": 3.5955156805417117,
      "tokens_seen": 1171174400
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.0003258375125376128,
      "loss": 2.3862,
      "theoretical_loss": 3.595497089167128,
      "tokens_seen": 1171239936
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.000325827482447342,
      "loss": 2.4888,
      "theoretical_loss": 3.5954784991240407,
      "tokens_seen": 1171305472
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.00032581745235707124,
      "loss": 2.5438,
      "theoretical_loss": 3.5954599104122797,
      "tokens_seen": 1171371008
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.0003258074222668004,
      "loss": 2.3983,
      "theoretical_loss": 3.5954413230316757,
      "tokens_seen": 1171436544
    },
    {
      "epoch": 3.09,
      "objective/train/docs_used": 1282424,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6180059909820557,
      "objective/train/theoretical_loss": 3.595436676394498,
      "objective/train/tokens_used": 1187029472,
      "theoretical_loss": 3.595436676394498,
      "tokens_seen": 1171452928
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.0003257973921765296,
      "loss": 2.658,
      "theoretical_loss": 3.5954227369820586,
      "tokens_seen": 1171502080
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.00032578736208625883,
      "loss": 2.4644,
      "theoretical_loss": 3.5954041522632583,
      "tokens_seen": 1171567616
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.00032577733199598796,
      "loss": 2.6907,
      "theoretical_loss": 3.5953855688751055,
      "tokens_seen": 1171633152
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.0003257673019057172,
      "loss": 2.7324,
      "theoretical_loss": 3.5953669868174307,
      "tokens_seen": 1171698688
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.0003257572718154463,
      "loss": 2.2496,
      "theoretical_loss": 3.595348406090064,
      "tokens_seen": 1171764224
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.00032574724172517556,
      "loss": 2.7043,
      "theoretical_loss": 3.595329826692835,
      "tokens_seen": 1171829760
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.00032573721163490474,
      "loss": 2.5151,
      "theoretical_loss": 3.5953112486255754,
      "tokens_seen": 1171895296
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.0003257271815446339,
      "loss": 2.5465,
      "theoretical_loss": 3.5952926718881146,
      "tokens_seen": 1171960832
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.0003257171514543631,
      "loss": 2.6868,
      "theoretical_loss": 3.595274096480284,
      "tokens_seen": 1172026368
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.0003257071213640923,
      "loss": 2.7925,
      "theoretical_loss": 3.5952555224019136,
      "tokens_seen": 1172091904
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.00032569709127382146,
      "loss": 2.7214,
      "theoretical_loss": 3.595236949652833,
      "tokens_seen": 1172157440
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.0003256870611835507,
      "loss": 2.564,
      "theoretical_loss": 3.5952183782328744,
      "tokens_seen": 1172222976
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.0003256770310932798,
      "loss": 2.4106,
      "theoretical_loss": 3.5951998081418672,
      "tokens_seen": 1172288512
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.00032566700100300906,
      "loss": 2.505,
      "theoretical_loss": 3.5951812393796434,
      "tokens_seen": 1172354048
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.00032565697091273824,
      "loss": 2.5024,
      "theoretical_loss": 3.595162671946032,
      "tokens_seen": 1172419584
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.0003256469408224674,
      "loss": 2.4508,
      "theoretical_loss": 3.5951441058408644,
      "tokens_seen": 1172485120
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.0003256369107321966,
      "loss": 2.7553,
      "theoretical_loss": 3.5951255410639713,
      "tokens_seen": 1172550656
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.0003256268806419258,
      "loss": 2.5269,
      "theoretical_loss": 3.595106977615184,
      "tokens_seen": 1172616192
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.00032561685055165497,
      "loss": 2.6031,
      "theoretical_loss": 3.595088415494332,
      "tokens_seen": 1172681728
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.0003256068204613842,
      "loss": 2.5579,
      "theoretical_loss": 3.5950698547012463,
      "tokens_seen": 1172747264
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.00032559679037111333,
      "loss": 2.374,
      "theoretical_loss": 3.595051295235759,
      "tokens_seen": 1172812800
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.00032558676028084256,
      "loss": 2.6482,
      "theoretical_loss": 3.5950327370977,
      "tokens_seen": 1172878336
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.0003255767301905717,
      "loss": 2.7096,
      "theoretical_loss": 3.595014180286901,
      "tokens_seen": 1172943872
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.0003255667001003009,
      "loss": 2.4358,
      "theoretical_loss": 3.5949956248031913,
      "tokens_seen": 1173009408
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.0003255566700100301,
      "loss": 2.5746,
      "theoretical_loss": 3.5949770706464035,
      "tokens_seen": 1173074944
    },
    {
      "epoch": 3.09,
      "objective/train/docs_used": 1282424,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.576780080795288,
      "objective/train/theoretical_loss": 3.5949724323145182,
      "objective/train/tokens_used": 1187029472,
      "theoretical_loss": 3.5949724323145182,
      "tokens_seen": 1173091328
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.0003255466399197593,
      "loss": 2.6818,
      "theoretical_loss": 3.594958517816368,
      "tokens_seen": 1173140480
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.00032553660982948847,
      "loss": 2.6777,
      "theoretical_loss": 3.5949399663129156,
      "tokens_seen": 1173206016
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.00032552657973921765,
      "loss": 2.7612,
      "theoretical_loss": 3.5949214161358776,
      "tokens_seen": 1173271552
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.00032551654964894683,
      "loss": 2.5333,
      "theoretical_loss": 3.594902867285085,
      "tokens_seen": 1173337088
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.00032550651955867607,
      "loss": 2.6658,
      "theoretical_loss": 3.594884319760369,
      "tokens_seen": 1173402624
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.0003254964894684052,
      "loss": 2.8645,
      "theoretical_loss": 3.59486577356156,
      "tokens_seen": 1173468160
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.00032548645937813443,
      "loss": 2.7192,
      "theoretical_loss": 3.594847228688491,
      "tokens_seen": 1173533696
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.0003254764292878636,
      "loss": 2.7719,
      "theoretical_loss": 3.594828685140991,
      "tokens_seen": 1173599232
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.0003254663991975928,
      "loss": 2.535,
      "theoretical_loss": 3.5948101429188926,
      "tokens_seen": 1173664768
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.000325456369107322,
      "loss": 2.5644,
      "theoretical_loss": 3.5947916020220267,
      "tokens_seen": 1173730304
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.00032544633901705115,
      "loss": 2.6096,
      "theoretical_loss": 3.594773062450225,
      "tokens_seen": 1173795840
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.00032543630892678034,
      "loss": 2.14,
      "theoretical_loss": 3.594754524203318,
      "tokens_seen": 1173861376
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.00032542627883650957,
      "loss": 2.4617,
      "theoretical_loss": 3.5947359872811377,
      "tokens_seen": 1173926912
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.0003254162487462387,
      "loss": 2.5123,
      "theoretical_loss": 3.5947174516835156,
      "tokens_seen": 1173992448
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.00032540621865596793,
      "loss": 2.6792,
      "theoretical_loss": 3.5946989174102826,
      "tokens_seen": 1174057984
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.00032539618856569706,
      "loss": 2.5653,
      "theoretical_loss": 3.5946803844612703,
      "tokens_seen": 1174123520
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.0003253861584754263,
      "loss": 2.5136,
      "theoretical_loss": 3.5946618528363103,
      "tokens_seen": 1174189056
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.0003253761283851555,
      "loss": 2.6376,
      "theoretical_loss": 3.594643322535234,
      "tokens_seen": 1174254592
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.00032536609829488466,
      "loss": 2.4409,
      "theoretical_loss": 3.5946247935578732,
      "tokens_seen": 1174320128
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.00032535606820461384,
      "loss": 2.6197,
      "theoretical_loss": 3.5946062659040594,
      "tokens_seen": 1174385664
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.000325346038114343,
      "loss": 2.6027,
      "theoretical_loss": 3.5945877395736243,
      "tokens_seen": 1174451200
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.0003253360080240722,
      "loss": 2.634,
      "theoretical_loss": 3.5945692145663988,
      "tokens_seen": 1174516736
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.00032532597793380144,
      "loss": 2.8002,
      "theoretical_loss": 3.5945506908822153,
      "tokens_seen": 1174582272
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.00032531594784353056,
      "loss": 2.607,
      "theoretical_loss": 3.5945321685209053,
      "tokens_seen": 1174647808
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.0003253059177532598,
      "loss": 2.6358,
      "theoretical_loss": 3.5945136474823007,
      "tokens_seen": 1174713344
    },
    {
      "epoch": 3.09,
      "objective/train/docs_used": 1282424,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.806351900100708,
      "objective/train/theoretical_loss": 3.5945090174293024,
      "objective/train/tokens_used": 1187029472,
      "theoretical_loss": 3.5945090174293024,
      "tokens_seen": 1174729728
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.000325295887662989,
      "loss": 2.546,
      "theoretical_loss": 3.594495127766233,
      "tokens_seen": 1174778880
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.00032528585757271816,
      "loss": 2.735,
      "theoretical_loss": 3.594476609372534,
      "tokens_seen": 1174844416
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.00032527582748244734,
      "loss": 2.649,
      "theoretical_loss": 3.5944580923010356,
      "tokens_seen": 1174909952
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.0003252657973921765,
      "loss": 2.1022,
      "theoretical_loss": 3.5944395765515695,
      "tokens_seen": 1174975488
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.0003252557673019057,
      "loss": 2.6525,
      "theoretical_loss": 3.594421062123968,
      "tokens_seen": 1175041024
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.00032524573721163494,
      "loss": 2.6477,
      "theoretical_loss": 3.5944025490180627,
      "tokens_seen": 1175106560
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.00032523570712136407,
      "loss": 2.6208,
      "theoretical_loss": 3.5943840372336853,
      "tokens_seen": 1175172096
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.0003252256770310933,
      "loss": 2.603,
      "theoretical_loss": 3.5943655267706687,
      "tokens_seen": 1175237632
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.00032521564694082243,
      "loss": 2.6506,
      "theoretical_loss": 3.594347017628844,
      "tokens_seen": 1175303168
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.00032520561685055166,
      "loss": 2.7226,
      "theoretical_loss": 3.5943285098080433,
      "tokens_seen": 1175368704
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.00032519558676028084,
      "loss": 2.6465,
      "theoretical_loss": 3.594310003308099,
      "tokens_seen": 1175434240
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.00032518555667001,
      "loss": 2.893,
      "theoretical_loss": 3.594291498128843,
      "tokens_seen": 1175499776
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.0003251755265797392,
      "loss": 2.506,
      "theoretical_loss": 3.5942729942701077,
      "tokens_seen": 1175565312
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.00032516549648946844,
      "loss": 2.6732,
      "theoretical_loss": 3.594254491731725,
      "tokens_seen": 1175630848
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.00032515546639919757,
      "loss": 2.6036,
      "theoretical_loss": 3.5942359905135275,
      "tokens_seen": 1175696384
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.0003251454363089268,
      "loss": 2.6352,
      "theoretical_loss": 3.5942174906153466,
      "tokens_seen": 1175761920
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.00032513540621865593,
      "loss": 2.4589,
      "theoretical_loss": 3.5941989920370148,
      "tokens_seen": 1175827456
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.00032512537612838517,
      "loss": 2.6806,
      "theoretical_loss": 3.594180494778365,
      "tokens_seen": 1175892992
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.00032511534603811435,
      "loss": 2.6492,
      "theoretical_loss": 3.5941619988392293,
      "tokens_seen": 1175958528
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.00032510531594784353,
      "loss": 2.6898,
      "theoretical_loss": 3.5941435042194394,
      "tokens_seen": 1176024064
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.0003250952858575727,
      "loss": 2.6345,
      "theoretical_loss": 3.5941250109188285,
      "tokens_seen": 1176089600
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.0003250852557673019,
      "loss": 2.4554,
      "theoretical_loss": 3.5941065189372288,
      "tokens_seen": 1176155136
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.00032507522567703107,
      "loss": 2.2642,
      "theoretical_loss": 3.594088028274472,
      "tokens_seen": 1176220672
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.0003250651955867603,
      "loss": 2.7183,
      "theoretical_loss": 3.5940695389303916,
      "tokens_seen": 1176286208
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.0003250551654964895,
      "loss": 2.3044,
      "theoretical_loss": 3.5940510509048194,
      "tokens_seen": 1176351744
    },
    {
      "epoch": 3.09,
      "objective/train/docs_used": 1282424,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.1171867847442627,
      "objective/train/theoretical_loss": 3.5940464291044236,
      "objective/train/tokens_used": 1187029472,
      "theoretical_loss": 3.5940464291044236,
      "tokens_seen": 1176368128
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.00032504513540621867,
      "loss": 2.7721,
      "theoretical_loss": 3.5940325641975885,
      "tokens_seen": 1176417280
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.00032503510531594785,
      "loss": 2.6341,
      "theoretical_loss": 3.594014078808531,
      "tokens_seen": 1176482816
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.00032502507522567703,
      "loss": 2.5146,
      "theoretical_loss": 3.5939955947374793,
      "tokens_seen": 1176548352
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.00032501504513540627,
      "loss": 2.6868,
      "theoretical_loss": 3.5939771119842665,
      "tokens_seen": 1176613888
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.0003250050150451354,
      "loss": 2.3658,
      "theoretical_loss": 3.593958630548725,
      "tokens_seen": 1176679424
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.00032499498495486463,
      "loss": 2.6863,
      "theoretical_loss": 3.593940150430688,
      "tokens_seen": 1176744960
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.0003249849548645938,
      "loss": 2.6961,
      "theoretical_loss": 3.5939216716299875,
      "tokens_seen": 1176810496
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.000324974924774323,
      "loss": 2.5861,
      "theoretical_loss": 3.5939031941464563,
      "tokens_seen": 1176876032
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.0003249648946840522,
      "loss": 2.667,
      "theoretical_loss": 3.5938847179799276,
      "tokens_seen": 1176941568
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.00032495486459378135,
      "loss": 2.5671,
      "theoretical_loss": 3.593866243130234,
      "tokens_seen": 1177007104
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.00032494483450351054,
      "loss": 2.2381,
      "theoretical_loss": 3.5938477695972084,
      "tokens_seen": 1177072640
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.00032493480441323977,
      "loss": 2.5901,
      "theoretical_loss": 3.593829297380683,
      "tokens_seen": 1177138176
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.0003249247743229689,
      "loss": 2.635,
      "theoretical_loss": 3.593810826480492,
      "tokens_seen": 1177203712
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.00032491474423269813,
      "loss": 2.5594,
      "theoretical_loss": 3.5937923568964676,
      "tokens_seen": 1177269248
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.00032490471414242726,
      "loss": 2.6957,
      "theoretical_loss": 3.5937738886284425,
      "tokens_seen": 1177334784
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.0003248946840521565,
      "loss": 2.7526,
      "theoretical_loss": 3.5937554216762497,
      "tokens_seen": 1177400320
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.0003248846539618857,
      "loss": 2.8548,
      "theoretical_loss": 3.5937369560397228,
      "tokens_seen": 1177465856
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.00032487462387161486,
      "loss": 2.6545,
      "theoretical_loss": 3.593718491718694,
      "tokens_seen": 1177531392
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.00032486459378134404,
      "loss": 2.6746,
      "theoretical_loss": 3.5937000287129974,
      "tokens_seen": 1177596928
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.0003248545636910732,
      "loss": 2.6491,
      "theoretical_loss": 3.5936815670224656,
      "tokens_seen": 1177662464
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.0003248445336008024,
      "loss": 2.5586,
      "theoretical_loss": 3.5936631066469316,
      "tokens_seen": 1177728000
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.00032483450351053164,
      "loss": 2.8073,
      "theoretical_loss": 3.5936446475862285,
      "tokens_seen": 1177793536
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.00032482447342026076,
      "loss": 2.401,
      "theoretical_loss": 3.59362618984019,
      "tokens_seen": 1177859072
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.00032481444332999,
      "loss": 2.5078,
      "theoretical_loss": 3.593607733408649,
      "tokens_seen": 1177924608
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.0003248044132397192,
      "loss": 2.5333,
      "theoretical_loss": 3.5935892782914385,
      "tokens_seen": 1177990144
    },
    {
      "epoch": 3.09,
      "objective/train/docs_used": 1282424,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5997087955474854,
      "objective/train/theoretical_loss": 3.5935846647174805,
      "objective/train/tokens_used": 1187029472,
      "theoretical_loss": 3.5935846647174805,
      "tokens_seen": 1178006528
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.00032479438314944836,
      "loss": 2.6478,
      "theoretical_loss": 3.593570824488392,
      "tokens_seen": 1178055680
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.00032478435305917754,
      "loss": 2.6641,
      "theoretical_loss": 3.593552371999343,
      "tokens_seen": 1178121216
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.0003247743229689067,
      "loss": 2.6132,
      "theoretical_loss": 3.5935339208241253,
      "tokens_seen": 1178186752
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.0003247642928786359,
      "loss": 2.5248,
      "theoretical_loss": 3.593515470962571,
      "tokens_seen": 1178252288
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.00032475426278836514,
      "loss": 2.4717,
      "theoretical_loss": 3.5934970224145144,
      "tokens_seen": 1178317824
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.00032474423269809427,
      "loss": 2.4565,
      "theoretical_loss": 3.5934785751797893,
      "tokens_seen": 1178383360
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.0003247342026078235,
      "loss": 2.5356,
      "theoretical_loss": 3.593460129258228,
      "tokens_seen": 1178448896
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.00032472417251755263,
      "loss": 2.4481,
      "theoretical_loss": 3.5934416846496644,
      "tokens_seen": 1178514432
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.00032471414242728186,
      "loss": 2.679,
      "theoretical_loss": 3.593423241353933,
      "tokens_seen": 1178579968
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.00032470411233701104,
      "loss": 2.4818,
      "theoretical_loss": 3.593404799370867,
      "tokens_seen": 1178645504
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.0003246940822467402,
      "loss": 2.6474,
      "theoretical_loss": 3.5933863587002985,
      "tokens_seen": 1178711040
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.0003246840521564694,
      "loss": 2.5939,
      "theoretical_loss": 3.5933679193420627,
      "tokens_seen": 1178776576
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.00032467402206619864,
      "loss": 2.6959,
      "theoretical_loss": 3.593349481295993,
      "tokens_seen": 1178842112
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.00032466399197592777,
      "loss": 2.6948,
      "theoretical_loss": 3.5933310445619226,
      "tokens_seen": 1178907648
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.000324653961885657,
      "loss": 2.3513,
      "theoretical_loss": 3.593312609139686,
      "tokens_seen": 1178973184
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.00032464393179538613,
      "loss": 2.438,
      "theoretical_loss": 3.593294175029116,
      "tokens_seen": 1179038720
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.00032463390170511537,
      "loss": 2.4049,
      "theoretical_loss": 3.593275742230047,
      "tokens_seen": 1179104256
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.00032462387161484455,
      "loss": 2.7499,
      "theoretical_loss": 3.5932573107423127,
      "tokens_seen": 1179169792
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.00032461384152457373,
      "loss": 2.6122,
      "theoretical_loss": 3.593238880565747,
      "tokens_seen": 1179235328
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.0003246038114343029,
      "loss": 2.5427,
      "theoretical_loss": 3.5932204517001836,
      "tokens_seen": 1179300864
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.0003245937813440321,
      "loss": 2.5502,
      "theoretical_loss": 3.593202024145456,
      "tokens_seen": 1179366400
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.00032458375125376127,
      "loss": 2.5535,
      "theoretical_loss": 3.593183597901399,
      "tokens_seen": 1179431936
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.0003245737211634905,
      "loss": 2.6025,
      "theoretical_loss": 3.593165172967846,
      "tokens_seen": 1179497472
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.00032456369107321963,
      "loss": 2.736,
      "theoretical_loss": 3.5931467493446307,
      "tokens_seen": 1179563008
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.00032455366098294887,
      "loss": 2.3614,
      "theoretical_loss": 3.593128327031588,
      "tokens_seen": 1179628544
    },
    {
      "epoch": 3.09,
      "objective/train/docs_used": 1282424,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.648859977722168,
      "objective/train/theoretical_loss": 3.5931237216580225,
      "objective/train/tokens_used": 1187029472,
      "theoretical_loss": 3.5931237216580225,
      "tokens_seen": 1179644928
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.000324543630892678,
      "loss": 2.65,
      "theoretical_loss": 3.593109906028552,
      "tokens_seen": 1179694080
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.00032453360080240723,
      "loss": 2.6016,
      "theoretical_loss": 3.5930914863353554,
      "tokens_seen": 1179759616
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.0003245235707121364,
      "loss": 2.6044,
      "theoretical_loss": 3.5930730679518335,
      "tokens_seen": 1179825152
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.0003245135406218656,
      "loss": 2.4911,
      "theoretical_loss": 3.5930546508778196,
      "tokens_seen": 1179890688
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.0003245035105315948,
      "loss": 2.5624,
      "theoretical_loss": 3.593036235113149,
      "tokens_seen": 1179956224
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.000324493480441324,
      "loss": 2.6682,
      "theoretical_loss": 3.5930178206576553,
      "tokens_seen": 1180021760
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.00032448345035105314,
      "loss": 2.8837,
      "theoretical_loss": 3.592999407511172,
      "tokens_seen": 1180087296
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.0003244734202607824,
      "loss": 2.643,
      "theoretical_loss": 3.592980995673535,
      "tokens_seen": 1180152832
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.0003244633901705115,
      "loss": 2.408,
      "theoretical_loss": 3.5929625851445772,
      "tokens_seen": 1180218368
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.00032445336008024074,
      "loss": 2.7895,
      "theoretical_loss": 3.592944175924133,
      "tokens_seen": 1180283904
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.0003244433299899699,
      "loss": 2.629,
      "theoretical_loss": 3.592925768012038,
      "tokens_seen": 1180349440
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.0003244332998996991,
      "loss": 2.3755,
      "theoretical_loss": 3.592907361408125,
      "tokens_seen": 1180414976
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.0003244232698094283,
      "loss": 2.7623,
      "theoretical_loss": 3.5928889561122292,
      "tokens_seen": 1180480512
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.00032441323971915746,
      "loss": 2.7398,
      "theoretical_loss": 3.592870552124185,
      "tokens_seen": 1180546048
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.00032440320962888664,
      "loss": 2.6753,
      "theoretical_loss": 3.592852149443827,
      "tokens_seen": 1180611584
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.0003243931795386159,
      "loss": 2.5979,
      "theoretical_loss": 3.5928337480709898,
      "tokens_seen": 1180677120
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.000324383149448345,
      "loss": 2.5014,
      "theoretical_loss": 3.592815348005507,
      "tokens_seen": 1180742656
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.00032437311935807424,
      "loss": 2.2792,
      "theoretical_loss": 3.592796949247214,
      "tokens_seen": 1180808192
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.00032436308926780337,
      "loss": 2.755,
      "theoretical_loss": 3.592778551795945,
      "tokens_seen": 1180873728
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.0003243530591775326,
      "loss": 2.6092,
      "theoretical_loss": 3.5927601556515354,
      "tokens_seen": 1180939264
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.0003243430290872618,
      "loss": 2.592,
      "theoretical_loss": 3.5927417608138192,
      "tokens_seen": 1181004800
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.00032433299899699096,
      "loss": 2.7324,
      "theoretical_loss": 3.5927233672826304,
      "tokens_seen": 1181070336
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.0003243229689067202,
      "loss": 2.6515,
      "theoretical_loss": 3.5927049750578055,
      "tokens_seen": 1181135872
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.0003243129388164494,
      "loss": 2.7052,
      "theoretical_loss": 3.5926865841391775,
      "tokens_seen": 1181201408
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.00032430290872617856,
      "loss": 2.8725,
      "theoretical_loss": 3.592668194526582,
      "tokens_seen": 1181266944
    },
    {
      "epoch": 3.09,
      "objective/train/docs_used": 1282424,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.2769789695739746,
      "objective/train/theoretical_loss": 3.5926635973274816,
      "objective/train/tokens_used": 1187029472,
      "theoretical_loss": 3.5926635973274816,
      "tokens_seen": 1181283328
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.00032429287863590774,
      "loss": 2.4443,
      "theoretical_loss": 3.592649806219854,
      "tokens_seen": 1181332480
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.0003242828485456369,
      "loss": 2.4296,
      "theoretical_loss": 3.5926314192188276,
      "tokens_seen": 1181398016
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.0003242728184553661,
      "loss": 2.5088,
      "theoretical_loss": 3.592613033523339,
      "tokens_seen": 1181463552
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.00032426278836509534,
      "loss": 2.3114,
      "theoretical_loss": 3.592594649133221,
      "tokens_seen": 1181529088
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.00032425275827482447,
      "loss": 2.7753,
      "theoretical_loss": 3.59257626604831,
      "tokens_seen": 1181594624
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.0003242427281845537,
      "loss": 2.4749,
      "theoretical_loss": 3.592557884268441,
      "tokens_seen": 1181660160
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.00032423269809428283,
      "loss": 2.7379,
      "theoretical_loss": 3.592539503793448,
      "tokens_seen": 1181725696
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.00032422266800401206,
      "loss": 2.6699,
      "theoretical_loss": 3.5925211246231674,
      "tokens_seen": 1181791232
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.00032421263791374125,
      "loss": 2.4808,
      "theoretical_loss": 3.5925027467574333,
      "tokens_seen": 1181856768
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.0003242026078234704,
      "loss": 2.5653,
      "theoretical_loss": 3.5924843701960807,
      "tokens_seen": 1181922304
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.0003241925777331996,
      "loss": 2.6213,
      "theoretical_loss": 3.5924659949389453,
      "tokens_seen": 1181987840
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.00032418254764292884,
      "loss": 2.4843,
      "theoretical_loss": 3.592447620985862,
      "tokens_seen": 1182053376
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.00032417251755265797,
      "loss": 2.5115,
      "theoretical_loss": 3.592429248336665,
      "tokens_seen": 1182118912
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.0003241624874623872,
      "loss": 2.8637,
      "theoretical_loss": 3.592410876991191,
      "tokens_seen": 1182184448
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.00032415245737211633,
      "loss": 2.6917,
      "theoretical_loss": 3.5923925069492744,
      "tokens_seen": 1182249984
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.00032414242728184557,
      "loss": 2.5962,
      "theoretical_loss": 3.592374138210751,
      "tokens_seen": 1182315520
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.00032413239719157475,
      "loss": 2.2916,
      "theoretical_loss": 3.5923557707754554,
      "tokens_seen": 1182381056
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.00032412236710130393,
      "loss": 2.2378,
      "theoretical_loss": 3.5923374046432226,
      "tokens_seen": 1182446592
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.0003241123370110331,
      "loss": 2.6255,
      "theoretical_loss": 3.5923190398138893,
      "tokens_seen": 1182512128
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.0003241023069207623,
      "loss": 2.641,
      "theoretical_loss": 3.59230067628729,
      "tokens_seen": 1182577664
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.00032409227683049147,
      "loss": 2.6365,
      "theoretical_loss": 3.59228231406326,
      "tokens_seen": 1182643200
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.0003240822467402207,
      "loss": 2.683,
      "theoretical_loss": 3.5922639531416354,
      "tokens_seen": 1182708736
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.00032407221664994984,
      "loss": 2.7049,
      "theoretical_loss": 3.592245593522251,
      "tokens_seen": 1182774272
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.00032406218655967907,
      "loss": 2.6362,
      "theoretical_loss": 3.592227235204942,
      "tokens_seen": 1182839808
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.0003240521564694082,
      "loss": 2.5488,
      "theoretical_loss": 3.5922088781895454,
      "tokens_seen": 1182905344
    },
    {
      "epoch": 3.09,
      "objective/train/docs_used": 1282424,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5676214694976807,
      "objective/train/theoretical_loss": 3.5922042891391004,
      "objective/train/tokens_used": 1187029472,
      "theoretical_loss": 3.5922042891391004,
      "tokens_seen": 1182921728
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.00032404212637913743,
      "loss": 2.6056,
      "theoretical_loss": 3.592190522475895,
      "tokens_seen": 1182970880
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.0003240320962888666,
      "loss": 2.6205,
      "theoretical_loss": 3.5921721680638274,
      "tokens_seen": 1183036416
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.0003240220661985958,
      "loss": 2.6659,
      "theoretical_loss": 3.5921538149531784,
      "tokens_seen": 1183101952
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.000324012036108325,
      "loss": 2.4522,
      "theoretical_loss": 3.5921354631437827,
      "tokens_seen": 1183167488
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.0003240020060180542,
      "loss": 2.6169,
      "theoretical_loss": 3.5921171126354765,
      "tokens_seen": 1183233024
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.00032399197592778334,
      "loss": 2.527,
      "theoretical_loss": 3.592098763428096,
      "tokens_seen": 1183298560
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.0003239819458375126,
      "loss": 2.5272,
      "theoretical_loss": 3.5920804155214756,
      "tokens_seen": 1183364096
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.0003239719157472417,
      "loss": 2.5678,
      "theoretical_loss": 3.5920620689154523,
      "tokens_seen": 1183429632
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.00032396188565697094,
      "loss": 2.5987,
      "theoretical_loss": 3.5920437236098612,
      "tokens_seen": 1183495168
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.0003239518555667001,
      "loss": 2.8055,
      "theoretical_loss": 3.592025379604539,
      "tokens_seen": 1183560704
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.0003239418254764293,
      "loss": 2.6494,
      "theoretical_loss": 3.5920070368993207,
      "tokens_seen": 1183626240
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.0003239317953861585,
      "loss": 2.6893,
      "theoretical_loss": 3.5919886954940425,
      "tokens_seen": 1183691776
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.00032392176529588766,
      "loss": 2.1192,
      "theoretical_loss": 3.59197035538854,
      "tokens_seen": 1183757312
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.00032391173520561684,
      "loss": 2.5386,
      "theoretical_loss": 3.591952016582649,
      "tokens_seen": 1183822848
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.0003239017051153461,
      "loss": 2.7836,
      "theoretical_loss": 3.591933679076207,
      "tokens_seen": 1183888384
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.0003238916750250752,
      "loss": 2.5445,
      "theoretical_loss": 3.5919153428690476,
      "tokens_seen": 1183953920
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.00032388164493480444,
      "loss": 2.7289,
      "theoretical_loss": 3.591897007961008,
      "tokens_seen": 1184019456
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.00032387161484453357,
      "loss": 2.8839,
      "theoretical_loss": 3.5918786743519258,
      "tokens_seen": 1184084992
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.0003238615847542628,
      "loss": 2.4107,
      "theoretical_loss": 3.5918603420416346,
      "tokens_seen": 1184150528
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.000323851554663992,
      "loss": 2.5578,
      "theoretical_loss": 3.5918420110299714,
      "tokens_seen": 1184216064
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.00032384152457372116,
      "loss": 2.6188,
      "theoretical_loss": 3.5918236813167725,
      "tokens_seen": 1184281600
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.00032383149448345034,
      "loss": 2.7958,
      "theoretical_loss": 3.591805352901874,
      "tokens_seen": 1184347136
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.0003238214643931796,
      "loss": 2.8208,
      "theoretical_loss": 3.5917870257851128,
      "tokens_seen": 1184412672
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.0003238114343029087,
      "loss": 2.4828,
      "theoretical_loss": 3.591768699966324,
      "tokens_seen": 1184478208
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.00032380140421263794,
      "loss": 2.5075,
      "theoretical_loss": 3.5917503754453444,
      "tokens_seen": 1184543744
    },
    {
      "epoch": 3.09,
      "objective/train/docs_used": 1282424,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8725781440734863,
      "objective/train/theoretical_loss": 3.5917457945178626,
      "objective/train/tokens_used": 1187029472,
      "theoretical_loss": 3.5917457945178626,
      "tokens_seen": 1184560128
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.00032379137412236707,
      "loss": 2.8068,
      "theoretical_loss": 3.59173205222201,
      "tokens_seen": 1184609280
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.0003237813440320963,
      "loss": 2.456,
      "theoretical_loss": 3.591713730296158,
      "tokens_seen": 1184674816
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.0003237713139418255,
      "loss": 2.7284,
      "theoretical_loss": 3.591695409667624,
      "tokens_seen": 1184740352
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.00032376128385155467,
      "loss": 2.3995,
      "theoretical_loss": 3.591677090336244,
      "tokens_seen": 1184805888
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.00032375125376128385,
      "loss": 2.7092,
      "theoretical_loss": 3.591658772301855,
      "tokens_seen": 1184871424
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.00032374122367101303,
      "loss": 2.6974,
      "theoretical_loss": 3.591640455564294,
      "tokens_seen": 1184936960
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.0003237311935807422,
      "loss": 2.286,
      "theoretical_loss": 3.5916221401233965,
      "tokens_seen": 1185002496
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.00032372116349047145,
      "loss": 2.3438,
      "theoretical_loss": 3.591603825978999,
      "tokens_seen": 1185068032
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.00032371113340020057,
      "loss": 2.4572,
      "theoretical_loss": 3.591585513130939,
      "tokens_seen": 1185133568
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.0003237011033099298,
      "loss": 2.7635,
      "theoretical_loss": 3.5915672015790525,
      "tokens_seen": 1185199104
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.00032369107321965893,
      "loss": 2.781,
      "theoretical_loss": 3.5915488913231757,
      "tokens_seen": 1185264640
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.00032368104312938817,
      "loss": 2.649,
      "theoretical_loss": 3.5915305823631454,
      "tokens_seen": 1185330176
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.00032367101303911735,
      "loss": 2.5419,
      "theoretical_loss": 3.5915122746987986,
      "tokens_seen": 1185395712
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.00032366098294884653,
      "loss": 2.563,
      "theoretical_loss": 3.5914939683299716,
      "tokens_seen": 1185461248
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.0003236509528585757,
      "loss": 2.7375,
      "theoretical_loss": 3.591475663256502,
      "tokens_seen": 1185526784
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.00032364092276830495,
      "loss": 2.3776,
      "theoretical_loss": 3.591457359478225,
      "tokens_seen": 1185592320
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.0003236308926780341,
      "loss": 2.5519,
      "theoretical_loss": 3.591439056994979,
      "tokens_seen": 1185657856
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.0003236208625877633,
      "loss": 2.798,
      "theoretical_loss": 3.5914207558066,
      "tokens_seen": 1185723392
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.00032361083249749244,
      "loss": 2.7447,
      "theoretical_loss": 3.591402455912924,
      "tokens_seen": 1185788928
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.0003236008024072217,
      "loss": 2.5904,
      "theoretical_loss": 3.5913841573137892,
      "tokens_seen": 1185854464
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.00032359077231695085,
      "loss": 2.6773,
      "theoretical_loss": 3.5913658600090326,
      "tokens_seen": 1185920000
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.00032358074222668004,
      "loss": 2.6649,
      "theoretical_loss": 3.5913475639984895,
      "tokens_seen": 1185985536
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.00032357071213640927,
      "loss": 2.5266,
      "theoretical_loss": 3.5913292692819985,
      "tokens_seen": 1186051072
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.0003235606820461384,
      "loss": 2.7652,
      "theoretical_loss": 3.591310975859396,
      "tokens_seen": 1186116608
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.00032355065195586763,
      "loss": 2.641,
      "theoretical_loss": 3.5912926837305186,
      "tokens_seen": 1186182144
    },
    {
      "epoch": 3.09,
      "objective/train/docs_used": 1282424,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.350708246231079,
      "objective/train/theoretical_loss": 3.5912881109004244,
      "objective/train/tokens_used": 1187029472,
      "theoretical_loss": 3.5912881109004244,
      "tokens_seen": 1186198528
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.0003235406218655968,
      "loss": 2.4416,
      "theoretical_loss": 3.591274392895204,
      "tokens_seen": 1186247680
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.000323530591775326,
      "loss": 2.4279,
      "theoretical_loss": 3.5912561033532886,
      "tokens_seen": 1186313216
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.0003235205616850552,
      "loss": 2.6713,
      "theoretical_loss": 3.59123781510461,
      "tokens_seen": 1186378752
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.0003235105315947844,
      "loss": 2.8526,
      "theoretical_loss": 3.591219528149005,
      "tokens_seen": 1186444288
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.00032350050150451354,
      "loss": 2.4494,
      "theoretical_loss": 3.5912012424863113,
      "tokens_seen": 1186509824
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.0003234904714142428,
      "loss": 2.661,
      "theoretical_loss": 3.5911829581163657,
      "tokens_seen": 1186575360
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.0003234804413239719,
      "loss": 2.5418,
      "theoretical_loss": 3.591164675039005,
      "tokens_seen": 1186640896
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.00032347041123370114,
      "loss": 2.596,
      "theoretical_loss": 3.5911463932540677,
      "tokens_seen": 1186706432
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.0003234603811434303,
      "loss": 2.421,
      "theoretical_loss": 3.5911281127613894,
      "tokens_seen": 1186771968
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.0003234503510531595,
      "loss": 2.4505,
      "theoretical_loss": 3.591109833560809,
      "tokens_seen": 1186837504
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.0003234403209628887,
      "loss": 2.5962,
      "theoretical_loss": 3.591091555652162,
      "tokens_seen": 1186903040
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.00032343029087261786,
      "loss": 2.823,
      "theoretical_loss": 3.591073279035288,
      "tokens_seen": 1186968576
    },
    {
      "epoch": 3.09,
      "learning_rate": 0.00032342026078234704,
      "loss": 2.4926,
      "theoretical_loss": 3.591055003710023,
      "tokens_seen": 1187034112
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0003234102306920763,
      "loss": 3.3894,
      "theoretical_loss": 3.5910324468863224,
      "tokens_seen": 1187115008
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0003234002006018054,
      "loss": 2.6997,
      "theoretical_loss": 3.5910141744464092,
      "tokens_seen": 1187180544
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.00032339017051153464,
      "loss": 2.7738,
      "theoretical_loss": 3.5909959032975793,
      "tokens_seen": 1187246080
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.00032338014042126377,
      "loss": 2.723,
      "theoretical_loss": 3.590977633439671,
      "tokens_seen": 1187311616
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.000323370110330993,
      "loss": 2.6097,
      "theoretical_loss": 3.5909593648725204,
      "tokens_seen": 1187377152
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0003233600802407222,
      "loss": 2.744,
      "theoretical_loss": 3.5909410975959664,
      "tokens_seen": 1187442688
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.00032335005015045136,
      "loss": 2.8156,
      "theoretical_loss": 3.590922831609846,
      "tokens_seen": 1187508224
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.00032334002006018054,
      "loss": 2.8714,
      "theoretical_loss": 3.5909045669139967,
      "tokens_seen": 1187573760
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0003233299899699098,
      "loss": 2.5544,
      "theoretical_loss": 3.590886303508257,
      "tokens_seen": 1187639296
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0003233199598796389,
      "loss": 2.7742,
      "theoretical_loss": 3.5908680413924636,
      "tokens_seen": 1187704832
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.00032330992978936814,
      "loss": 2.8224,
      "theoretical_loss": 3.5908497805664545,
      "tokens_seen": 1187770368
    },
    {
      "epoch": 4.0,
      "objective/train/docs_used": 1347233,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8168835639953613,
      "objective/train/theoretical_loss": 3.5908315210300685,
      "objective/train/tokens_used": 1208295904,
      "theoretical_loss": 3.5908315210300685,
      "tokens_seen": 1187835904
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.00032329989969909727,
      "loss": 2.6189,
      "theoretical_loss": 3.5908315210300685,
      "tokens_seen": 1187835904
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0003232898696088265,
      "loss": 2.699,
      "theoretical_loss": 3.5908132627831417,
      "tokens_seen": 1187901440
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0003232798395185557,
      "loss": 2.7227,
      "theoretical_loss": 3.5907950058255125,
      "tokens_seen": 1187966976
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.00032326980942828487,
      "loss": 2.7502,
      "theoretical_loss": 3.590776750157019,
      "tokens_seen": 1188032512
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.00032325977933801405,
      "loss": 2.6845,
      "theoretical_loss": 3.5907584957774996,
      "tokens_seen": 1188098048
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.00032324974924774323,
      "loss": 2.8323,
      "theoretical_loss": 3.590740242686791,
      "tokens_seen": 1188163584
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0003232397191574724,
      "loss": 2.6911,
      "theoretical_loss": 3.590721990884732,
      "tokens_seen": 1188229120
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.00032322968906720165,
      "loss": 2.8486,
      "theoretical_loss": 3.5907037403711604,
      "tokens_seen": 1188294656
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.00032321965897693077,
      "loss": 2.7744,
      "theoretical_loss": 3.590685491145914,
      "tokens_seen": 1188360192
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.00032320962888666,
      "loss": 2.7593,
      "theoretical_loss": 3.5906672432088302,
      "tokens_seen": 1188425728
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.00032319959879638913,
      "loss": 2.7644,
      "theoretical_loss": 3.5906489965597483,
      "tokens_seen": 1188491264
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.00032318956870611837,
      "loss": 2.7176,
      "theoretical_loss": 3.5906307511985056,
      "tokens_seen": 1188556800
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.00032317953861584755,
      "loss": 2.8788,
      "theoretical_loss": 3.5906125071249404,
      "tokens_seen": 1188622336
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.00032316950852557673,
      "loss": 2.7639,
      "theoretical_loss": 3.5905942643388906,
      "tokens_seen": 1188687872
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0003231594784353059,
      "loss": 2.5787,
      "theoretical_loss": 3.590576022840195,
      "tokens_seen": 1188753408
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.00032314944834503515,
      "loss": 2.7243,
      "theoretical_loss": 3.590557782628691,
      "tokens_seen": 1188818944
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0003231394182547643,
      "loss": 2.7841,
      "theoretical_loss": 3.5905395437042174,
      "tokens_seen": 1188884480
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0003231293881644935,
      "loss": 2.5921,
      "theoretical_loss": 3.5905213060666124,
      "tokens_seen": 1188950016
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.00032311935807422264,
      "loss": 2.8621,
      "theoretical_loss": 3.590503069715714,
      "tokens_seen": 1189015552
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0003231093279839519,
      "loss": 2.7728,
      "theoretical_loss": 3.5904848346513605,
      "tokens_seen": 1189081088
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.00032309929789368105,
      "loss": 2.8418,
      "theoretical_loss": 3.5904666008733903,
      "tokens_seen": 1189146624
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.00032308926780341024,
      "loss": 2.7546,
      "theoretical_loss": 3.590448368381642,
      "tokens_seen": 1189212160
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0003230792377131394,
      "loss": 2.4472,
      "theoretical_loss": 3.5904301371759533,
      "tokens_seen": 1189277696
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0003230692076228686,
      "loss": 2.7317,
      "theoretical_loss": 3.590411907256164,
      "tokens_seen": 1189343232
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0003230591775325978,
      "loss": 2.7694,
      "theoretical_loss": 3.590393678622111,
      "tokens_seen": 1189408768
    },
    {
      "epoch": 4.0,
      "objective/train/docs_used": 1352239,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.0367116928100586,
      "objective/train/theoretical_loss": 3.5903754512736334,
      "objective/train/tokens_used": 1209934304,
      "theoretical_loss": 3.5903754512736334,
      "tokens_seen": 1189474304
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.000323049147442327,
      "loss": 2.7382,
      "theoretical_loss": 3.5903754512736334,
      "tokens_seen": 1189474304
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.00032303911735205614,
      "loss": 2.4576,
      "theoretical_loss": 3.59035722521057,
      "tokens_seen": 1189539840
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0003230290872617854,
      "loss": 2.648,
      "theoretical_loss": 3.5903390004327593,
      "tokens_seen": 1189605376
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0003230190571715145,
      "loss": 2.7522,
      "theoretical_loss": 3.5903207769400396,
      "tokens_seen": 1189670912
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.00032300902708124374,
      "loss": 2.87,
      "theoretical_loss": 3.5903025547322494,
      "tokens_seen": 1189736448
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0003229989969909729,
      "loss": 2.7883,
      "theoretical_loss": 3.590284333809228,
      "tokens_seen": 1189801984
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0003229889669007021,
      "loss": 2.8014,
      "theoretical_loss": 3.590266114170813,
      "tokens_seen": 1189867520
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0003229789368104313,
      "loss": 2.9751,
      "theoretical_loss": 3.5902478958168436,
      "tokens_seen": 1189933056
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0003229689067201605,
      "loss": 2.8034,
      "theoretical_loss": 3.590229678747159,
      "tokens_seen": 1189998592
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.00032295887662988964,
      "loss": 2.6914,
      "theoretical_loss": 3.590211462961597,
      "tokens_seen": 1190064128
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0003229488465396189,
      "loss": 2.7211,
      "theoretical_loss": 3.5901932484599977,
      "tokens_seen": 1190129664
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.000322938816449348,
      "loss": 2.7461,
      "theoretical_loss": 3.590175035242199,
      "tokens_seen": 1190195200
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.00032292878635907724,
      "loss": 2.6821,
      "theoretical_loss": 3.5901568233080394,
      "tokens_seen": 1190260736
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0003229187562688064,
      "loss": 2.7034,
      "theoretical_loss": 3.5901386126573582,
      "tokens_seen": 1190326272
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0003229087261785356,
      "loss": 2.6984,
      "theoretical_loss": 3.590120403289994,
      "tokens_seen": 1190391808
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0003228986960882648,
      "loss": 2.8506,
      "theoretical_loss": 3.5901021952057866,
      "tokens_seen": 1190457344
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.00032288866599799397,
      "loss": 2.8324,
      "theoretical_loss": 3.590083988404574,
      "tokens_seen": 1190522880
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.00032287863590772315,
      "loss": 2.7436,
      "theoretical_loss": 3.590065782886196,
      "tokens_seen": 1190588416
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0003228686058174524,
      "loss": 2.669,
      "theoretical_loss": 3.590047578650491,
      "tokens_seen": 1190653952
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0003228585757271815,
      "loss": 2.7477,
      "theoretical_loss": 3.5900293756972976,
      "tokens_seen": 1190719488
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.00032284854563691074,
      "loss": 2.7421,
      "theoretical_loss": 3.590011174026456,
      "tokens_seen": 1190785024
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0003228385155466399,
      "loss": 2.7889,
      "theoretical_loss": 3.5899929736378047,
      "tokens_seen": 1190850560
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0003228284854563691,
      "loss": 2.8669,
      "theoretical_loss": 3.589974774531182,
      "tokens_seen": 1190916096
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.00032281845536609834,
      "loss": 2.7807,
      "theoretical_loss": 3.5899565767064288,
      "tokens_seen": 1190981632
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.00032280842527582747,
      "loss": 2.707,
      "theoretical_loss": 3.5899383801633835,
      "tokens_seen": 1191047168
    },
    {
      "epoch": 4.0,
      "objective/train/docs_used": 1357258,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.0113937854766846,
      "objective/train/theoretical_loss": 3.5899201849018842,
      "objective/train/tokens_used": 1211572704,
      "theoretical_loss": 3.5899201849018842,
      "tokens_seen": 1191112704
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0003227983951855567,
      "loss": 2.8706,
      "theoretical_loss": 3.5899201849018842,
      "tokens_seen": 1191112704
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0003227883650952859,
      "loss": 2.815,
      "theoretical_loss": 3.5899019909217724,
      "tokens_seen": 1191178240
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.00032277833500501507,
      "loss": 2.82,
      "theoretical_loss": 3.589883798222885,
      "tokens_seen": 1191243776
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.00032276830491474425,
      "loss": 2.4651,
      "theoretical_loss": 3.5898656068050627,
      "tokens_seen": 1191309312
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.00032275827482447343,
      "loss": 2.6,
      "theoretical_loss": 3.589847416668145,
      "tokens_seen": 1191374848
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0003227482447342026,
      "loss": 2.7682,
      "theoretical_loss": 3.58982922781197,
      "tokens_seen": 1191440384
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.00032273821464393185,
      "loss": 2.7798,
      "theoretical_loss": 3.5898110402363788,
      "tokens_seen": 1191505920
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.00032272818455366097,
      "loss": 2.6923,
      "theoretical_loss": 3.5897928539412094,
      "tokens_seen": 1191571456
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0003227181544633902,
      "loss": 2.6961,
      "theoretical_loss": 3.5897746689263017,
      "tokens_seen": 1191636992
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.00032270812437311933,
      "loss": 2.7842,
      "theoretical_loss": 3.5897564851914954,
      "tokens_seen": 1191702528
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.00032269809428284857,
      "loss": 2.6957,
      "theoretical_loss": 3.589738302736629,
      "tokens_seen": 1191768064
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.00032268806419257775,
      "loss": 2.8474,
      "theoretical_loss": 3.5897201215615437,
      "tokens_seen": 1191833600
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.00032267803410230693,
      "loss": 2.754,
      "theoretical_loss": 3.589701941666078,
      "tokens_seen": 1191899136
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0003226680040120361,
      "loss": 2.6007,
      "theoretical_loss": 3.589683763050071,
      "tokens_seen": 1191964672
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.00032265797392176535,
      "loss": 2.9095,
      "theoretical_loss": 3.5896655857133637,
      "tokens_seen": 1192030208
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0003226479438314945,
      "loss": 2.7656,
      "theoretical_loss": 3.589647409655795,
      "tokens_seen": 1192095744
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0003226379137412237,
      "loss": 2.7206,
      "theoretical_loss": 3.5896292348772043,
      "tokens_seen": 1192161280
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.00032262788365095284,
      "loss": 2.9005,
      "theoretical_loss": 3.5896110613774317,
      "tokens_seen": 1192226816
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0003226178535606821,
      "loss": 2.8052,
      "theoretical_loss": 3.5895928891563167,
      "tokens_seen": 1192292352
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.00032260782347041125,
      "loss": 2.7781,
      "theoretical_loss": 3.589574718213699,
      "tokens_seen": 1192357888
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.00032259779338014044,
      "loss": 2.7395,
      "theoretical_loss": 3.589556548549419,
      "tokens_seen": 1192423424
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0003225877632898696,
      "loss": 2.931,
      "theoretical_loss": 3.589538380163316,
      "tokens_seen": 1192488960
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0003225777331995988,
      "loss": 2.7551,
      "theoretical_loss": 3.5895202130552297,
      "tokens_seen": 1192554496
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.000322567703109328,
      "loss": 2.7762,
      "theoretical_loss": 3.5895020472250003,
      "tokens_seen": 1192620032
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0003225576730190572,
      "loss": 2.6019,
      "theoretical_loss": 3.589483882672468,
      "tokens_seen": 1192685568
    },
    {
      "epoch": 4.0,
      "objective/train/docs_used": 1362192,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9670515060424805,
      "objective/train/theoretical_loss": 3.589465719397472,
      "objective/train/tokens_used": 1213211104,
      "theoretical_loss": 3.589465719397472,
      "tokens_seen": 1192751104
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.00032254764292878634,
      "loss": 2.8065,
      "theoretical_loss": 3.589465719397472,
      "tokens_seen": 1192751104
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0003225376128385156,
      "loss": 2.6473,
      "theoretical_loss": 3.5894475573998528,
      "tokens_seen": 1192816640
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0003225275827482447,
      "loss": 2.7868,
      "theoretical_loss": 3.58942939667945,
      "tokens_seen": 1192882176
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.00032251755265797394,
      "loss": 2.7406,
      "theoretical_loss": 3.589411237236104,
      "tokens_seen": 1192947712
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0003225075225677031,
      "loss": 2.8387,
      "theoretical_loss": 3.589393079069654,
      "tokens_seen": 1193013248
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0003224974924774323,
      "loss": 2.6082,
      "theoretical_loss": 3.5893749221799416,
      "tokens_seen": 1193078784
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0003224874623871615,
      "loss": 2.7082,
      "theoretical_loss": 3.589356766566806,
      "tokens_seen": 1193144320
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0003224774322968907,
      "loss": 2.7234,
      "theoretical_loss": 3.5893386122300868,
      "tokens_seen": 1193209856
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.00032246740220661984,
      "loss": 2.6437,
      "theoretical_loss": 3.5893204591696253,
      "tokens_seen": 1193275392
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0003224573721163491,
      "loss": 2.6607,
      "theoretical_loss": 3.5893023073852612,
      "tokens_seen": 1193340928
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0003224473420260782,
      "loss": 2.6923,
      "theoretical_loss": 3.5892841568768343,
      "tokens_seen": 1193406464
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.00032243731193580744,
      "loss": 2.758,
      "theoretical_loss": 3.589266007644185,
      "tokens_seen": 1193472000
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0003224272818455366,
      "loss": 2.7518,
      "theoretical_loss": 3.5892478596871547,
      "tokens_seen": 1193537536
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0003224172517552658,
      "loss": 2.7189,
      "theoretical_loss": 3.589229713005582,
      "tokens_seen": 1193603072
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.000322407221664995,
      "loss": 2.7066,
      "theoretical_loss": 3.5892115675993077,
      "tokens_seen": 1193668608
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.00032239719157472417,
      "loss": 2.6636,
      "theoretical_loss": 3.5891934234681733,
      "tokens_seen": 1193734144
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.00032238716148445335,
      "loss": 2.6847,
      "theoretical_loss": 3.589175280612018,
      "tokens_seen": 1193799680
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0003223771313941826,
      "loss": 2.7743,
      "theoretical_loss": 3.589157139030683,
      "tokens_seen": 1193865216
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0003223671013039117,
      "loss": 2.8292,
      "theoretical_loss": 3.589138998724008,
      "tokens_seen": 1193930752
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.00032235707121364095,
      "loss": 2.8121,
      "theoretical_loss": 3.5891208596918345,
      "tokens_seen": 1193996288
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0003223470411233701,
      "loss": 2.7214,
      "theoretical_loss": 3.589102721934002,
      "tokens_seen": 1194061824
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0003223370110330993,
      "loss": 2.7198,
      "theoretical_loss": 3.589084585450351,
      "tokens_seen": 1194127360
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0003223269809428285,
      "loss": 2.7321,
      "theoretical_loss": 3.589066450240723,
      "tokens_seen": 1194192896
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.00032231695085255767,
      "loss": 2.669,
      "theoretical_loss": 3.5890483163049574,
      "tokens_seen": 1194258432
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.00032230692076228685,
      "loss": 2.8583,
      "theoretical_loss": 3.5890301836428957,
      "tokens_seen": 1194323968
    },
    {
      "epoch": 4.0,
      "objective/train/docs_used": 1367200,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5147886276245117,
      "objective/train/theoretical_loss": 3.589012052254379,
      "objective/train/tokens_used": 1214849504,
      "theoretical_loss": 3.589012052254379,
      "tokens_seen": 1194389504
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0003222968906720161,
      "loss": 2.6539,
      "theoretical_loss": 3.589012052254379,
      "tokens_seen": 1194389504
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0003222868605817452,
      "loss": 2.8133,
      "theoretical_loss": 3.5889939221392466,
      "tokens_seen": 1194455040
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.00032227683049147445,
      "loss": 2.7431,
      "theoretical_loss": 3.5889757932973403,
      "tokens_seen": 1194520576
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0003222668004012036,
      "loss": 2.7153,
      "theoretical_loss": 3.5889576657285,
      "tokens_seen": 1194586112
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0003222567703109328,
      "loss": 2.691,
      "theoretical_loss": 3.5889395394325674,
      "tokens_seen": 1194651648
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.000322246740220662,
      "loss": 2.784,
      "theoretical_loss": 3.5889214144093824,
      "tokens_seen": 1194717184
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.00032223671013039117,
      "loss": 2.6861,
      "theoretical_loss": 3.588903290658787,
      "tokens_seen": 1194782720
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.00032222668004012035,
      "loss": 2.567,
      "theoretical_loss": 3.5888851681806204,
      "tokens_seen": 1194848256
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.00032221664994984953,
      "loss": 2.6833,
      "theoretical_loss": 3.588867046974725,
      "tokens_seen": 1194913792
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0003222066198595787,
      "loss": 2.7416,
      "theoretical_loss": 3.588848927040941,
      "tokens_seen": 1194979328
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.00032219658976930795,
      "loss": 2.6881,
      "theoretical_loss": 3.588830808379109,
      "tokens_seen": 1195044864
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0003221865596790371,
      "loss": 2.5894,
      "theoretical_loss": 3.588812690989071,
      "tokens_seen": 1195110400
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0003221765295887663,
      "loss": 2.8525,
      "theoretical_loss": 3.588794574870667,
      "tokens_seen": 1195175936
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0003221664994984955,
      "loss": 2.6549,
      "theoretical_loss": 3.588776460023739,
      "tokens_seen": 1195241472
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0003221564694082247,
      "loss": 2.8704,
      "theoretical_loss": 3.5887583464481265,
      "tokens_seen": 1195307008
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.00032214643931795386,
      "loss": 2.7525,
      "theoretical_loss": 3.5887402341436725,
      "tokens_seen": 1195372544
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.00032213640922768304,
      "loss": 2.745,
      "theoretical_loss": 3.588722123110217,
      "tokens_seen": 1195438080
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0003221263791374122,
      "loss": 2.8724,
      "theoretical_loss": 3.588704013347601,
      "tokens_seen": 1195503616
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.00032211634904714145,
      "loss": 2.6772,
      "theoretical_loss": 3.588685904855666,
      "tokens_seen": 1195569152
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0003221063189568706,
      "loss": 2.5685,
      "theoretical_loss": 3.588667797634253,
      "tokens_seen": 1195634688
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0003220962888665998,
      "loss": 2.6183,
      "theoretical_loss": 3.5886496916832034,
      "tokens_seen": 1195700224
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.00032208625877632894,
      "loss": 2.8578,
      "theoretical_loss": 3.588631587002359,
      "tokens_seen": 1195765760
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0003220762286860582,
      "loss": 2.4787,
      "theoretical_loss": 3.58861348359156,
      "tokens_seen": 1195831296
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0003220661985957874,
      "loss": 2.4697,
      "theoretical_loss": 3.5885953814506486,
      "tokens_seen": 1195896832
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.00032205616850551654,
      "loss": 2.6127,
      "theoretical_loss": 3.588577280579466,
      "tokens_seen": 1195962368
    },
    {
      "epoch": 4.0,
      "objective/train/docs_used": 1372299,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.829193592071533,
      "objective/train/theoretical_loss": 3.5885591809778523,
      "objective/train/tokens_used": 1216487904,
      "theoretical_loss": 3.5885591809778523,
      "tokens_seen": 1196027904
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0003220461384152458,
      "loss": 2.8615,
      "theoretical_loss": 3.5885591809778523,
      "tokens_seen": 1196027904
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0003220361083249749,
      "loss": 2.649,
      "theoretical_loss": 3.588541082645651,
      "tokens_seen": 1196093440
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.00032202607823470414,
      "loss": 2.816,
      "theoretical_loss": 3.588522985582702,
      "tokens_seen": 1196158976
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0003220160481444333,
      "loss": 2.7417,
      "theoretical_loss": 3.5885048897888474,
      "tokens_seen": 1196224512
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0003220060180541625,
      "loss": 2.7062,
      "theoretical_loss": 3.588486795263928,
      "tokens_seen": 1196290048
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0003219959879638917,
      "loss": 2.618,
      "theoretical_loss": 3.588468702007787,
      "tokens_seen": 1196355584
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0003219859578736209,
      "loss": 2.4866,
      "theoretical_loss": 3.5884506100202636,
      "tokens_seen": 1196421120
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.00032197592778335004,
      "loss": 2.5016,
      "theoretical_loss": 3.5884325193012008,
      "tokens_seen": 1196486656
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0003219658976930793,
      "loss": 2.5037,
      "theoretical_loss": 3.58841442985044,
      "tokens_seen": 1196552192
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0003219558676028084,
      "loss": 2.5967,
      "theoretical_loss": 3.588396341667823,
      "tokens_seen": 1196617728
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.00032194583751253764,
      "loss": 2.6184,
      "theoretical_loss": 3.5883782547531906,
      "tokens_seen": 1196683264
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0003219358074222668,
      "loss": 2.8586,
      "theoretical_loss": 3.5883601691063856,
      "tokens_seen": 1196748800
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.000321925777331996,
      "loss": 2.7405,
      "theoretical_loss": 3.5883420847272487,
      "tokens_seen": 1196814336
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0003219157472417252,
      "loss": 2.8422,
      "theoretical_loss": 3.5883240016156224,
      "tokens_seen": 1196879872
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.00032190571715145437,
      "loss": 2.7227,
      "theoretical_loss": 3.588305919771348,
      "tokens_seen": 1196945408
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.00032189568706118355,
      "loss": 2.8192,
      "theoretical_loss": 3.5882878391942676,
      "tokens_seen": 1197010944
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0003218856569709128,
      "loss": 2.6226,
      "theoretical_loss": 3.5882697598842226,
      "tokens_seen": 1197076480
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0003218756268806419,
      "loss": 2.6056,
      "theoretical_loss": 3.5882516818410557,
      "tokens_seen": 1197142016
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.00032186559679037115,
      "loss": 2.751,
      "theoretical_loss": 3.588233605064608,
      "tokens_seen": 1197207552
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0003218555667001003,
      "loss": 2.5782,
      "theoretical_loss": 3.588215529554722,
      "tokens_seen": 1197273088
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0003218455366098295,
      "loss": 2.5359,
      "theoretical_loss": 3.5881974553112386,
      "tokens_seen": 1197338624
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0003218355065195587,
      "loss": 2.6353,
      "theoretical_loss": 3.5881793823340007,
      "tokens_seen": 1197404160
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.00032182547642928787,
      "loss": 2.4617,
      "theoretical_loss": 3.58816131062285,
      "tokens_seen": 1197469696
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.00032181544633901705,
      "loss": 2.8249,
      "theoretical_loss": 3.5881432401776285,
      "tokens_seen": 1197535232
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0003218054162487463,
      "loss": 2.753,
      "theoretical_loss": 3.588125170998178,
      "tokens_seen": 1197600768
    },
    {
      "epoch": 4.0,
      "objective/train/docs_used": 1377331,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7911863327026367,
      "objective/train/theoretical_loss": 3.588107103084342,
      "objective/train/tokens_used": 1218126304,
      "theoretical_loss": 3.588107103084342,
      "tokens_seen": 1197666304
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0003217953861584754,
      "loss": 2.8607,
      "theoretical_loss": 3.588107103084342,
      "tokens_seen": 1197666304
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.00032178535606820465,
      "loss": 2.597,
      "theoretical_loss": 3.58808903643596,
      "tokens_seen": 1197731840
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0003217753259779338,
      "loss": 2.684,
      "theoretical_loss": 3.5880709710528764,
      "tokens_seen": 1197797376
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.000321765295887663,
      "loss": 2.819,
      "theoretical_loss": 3.5880529069349327,
      "tokens_seen": 1197862912
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0003217552657973922,
      "loss": 2.6415,
      "theoretical_loss": 3.588034844081971,
      "tokens_seen": 1197928448
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.00032174523570712137,
      "loss": 2.6505,
      "theoretical_loss": 3.588016782493833,
      "tokens_seen": 1197993984
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.00032173520561685055,
      "loss": 2.799,
      "theoretical_loss": 3.587998722170362,
      "tokens_seen": 1198059520
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.00032172517552657974,
      "loss": 2.7641,
      "theoretical_loss": 3.5879806631113995,
      "tokens_seen": 1198125056
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0003217151454363089,
      "loss": 2.7987,
      "theoretical_loss": 3.587962605316788,
      "tokens_seen": 1198190592
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.00032170511534603815,
      "loss": 2.6003,
      "theoretical_loss": 3.5879445487863695,
      "tokens_seen": 1198256128
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0003216950852557673,
      "loss": 2.8261,
      "theoretical_loss": 3.587926493519987,
      "tokens_seen": 1198321664
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0003216850551654965,
      "loss": 2.6754,
      "theoretical_loss": 3.5879084395174825,
      "tokens_seen": 1198387200
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0003216750250752257,
      "loss": 2.7249,
      "theoretical_loss": 3.5878903867786986,
      "tokens_seen": 1198452736
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0003216649949849549,
      "loss": 2.6604,
      "theoretical_loss": 3.5878723353034783,
      "tokens_seen": 1198518272
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.00032165496489468406,
      "loss": 2.925,
      "theoretical_loss": 3.5878542850916624,
      "tokens_seen": 1198583808
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.00032164493480441324,
      "loss": 2.7317,
      "theoretical_loss": 3.587836236143095,
      "tokens_seen": 1198649344
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0003216349047141424,
      "loss": 2.8008,
      "theoretical_loss": 3.587818188457618,
      "tokens_seen": 1198714880
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.00032162487462387165,
      "loss": 2.689,
      "theoretical_loss": 3.587800142035074,
      "tokens_seen": 1198780416
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0003216148445336008,
      "loss": 2.7237,
      "theoretical_loss": 3.5877820968753054,
      "tokens_seen": 1198845952
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.00032160481444333,
      "loss": 2.6511,
      "theoretical_loss": 3.587764052978155,
      "tokens_seen": 1198911488
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.00032159478435305914,
      "loss": 2.5991,
      "theoretical_loss": 3.587746010343466,
      "tokens_seen": 1198977024
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0003215847542627884,
      "loss": 2.7523,
      "theoretical_loss": 3.5877279689710804,
      "tokens_seen": 1199042560
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.00032157472417251756,
      "loss": 2.8926,
      "theoretical_loss": 3.587709928860841,
      "tokens_seen": 1199108096
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.00032156469408224674,
      "loss": 2.6584,
      "theoretical_loss": 3.5876918900125903,
      "tokens_seen": 1199173632
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0003215546639919759,
      "loss": 2.5422,
      "theoretical_loss": 3.5876738524261715,
      "tokens_seen": 1199239168
    },
    {
      "epoch": 4.0,
      "objective/train/docs_used": 1382178,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.804316282272339,
      "objective/train/theoretical_loss": 3.587655816101427,
      "objective/train/tokens_used": 1219764704,
      "theoretical_loss": 3.587655816101427,
      "tokens_seen": 1199304704
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0003215446339017051,
      "loss": 2.7885,
      "theoretical_loss": 3.587655816101427,
      "tokens_seen": 1199304704
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0003215346038114343,
      "loss": 2.8044,
      "theoretical_loss": 3.5876377810382003,
      "tokens_seen": 1199370240
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0003215245737211635,
      "loss": 2.7155,
      "theoretical_loss": 3.5876197472363334,
      "tokens_seen": 1199435776
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.00032151454363089265,
      "loss": 2.5803,
      "theoretical_loss": 3.58760171469567,
      "tokens_seen": 1199501312
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0003215045135406219,
      "loss": 2.6165,
      "theoretical_loss": 3.587583683416052,
      "tokens_seen": 1199566848
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.00032149448345035106,
      "loss": 2.7678,
      "theoretical_loss": 3.5875656533973235,
      "tokens_seen": 1199632384
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.00032148445336008024,
      "loss": 2.8196,
      "theoretical_loss": 3.5875476246393263,
      "tokens_seen": 1199697920
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0003214744232698094,
      "loss": 2.549,
      "theoretical_loss": 3.587529597141904,
      "tokens_seen": 1199763456
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0003214643931795386,
      "loss": 2.85,
      "theoretical_loss": 3.5875115709048995,
      "tokens_seen": 1199828992
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0003214543630892678,
      "loss": 2.6543,
      "theoretical_loss": 3.587493545928156,
      "tokens_seen": 1199894528
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.000321444332998997,
      "loss": 2.7943,
      "theoretical_loss": 3.587475522211516,
      "tokens_seen": 1199960064
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.00032143430290872615,
      "loss": 2.6293,
      "theoretical_loss": 3.587457499754823,
      "tokens_seen": 1200025600
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0003214242728184554,
      "loss": 2.7308,
      "theoretical_loss": 3.5874394785579207,
      "tokens_seen": 1200091136
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0003214142427281845,
      "loss": 2.6226,
      "theoretical_loss": 3.587421458620651,
      "tokens_seen": 1200156672
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.00032140421263791375,
      "loss": 2.615,
      "theoretical_loss": 3.5874034399428583,
      "tokens_seen": 1200222208
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.00032139418254764293,
      "loss": 2.7029,
      "theoretical_loss": 3.5873854225243855,
      "tokens_seen": 1200287744
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0003213841524573721,
      "loss": 2.6413,
      "theoretical_loss": 3.587367406365075,
      "tokens_seen": 1200353280
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0003213741223671013,
      "loss": 2.6256,
      "theoretical_loss": 3.5873493914647705,
      "tokens_seen": 1200418816
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0003213640922768305,
      "loss": 2.756,
      "theoretical_loss": 3.5873313778233165,
      "tokens_seen": 1200484352
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.00032135406218655965,
      "loss": 2.6721,
      "theoretical_loss": 3.5873133654405542,
      "tokens_seen": 1200549888
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0003213440320962889,
      "loss": 2.6894,
      "theoretical_loss": 3.5872953543163284,
      "tokens_seen": 1200615424
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.000321334002006018,
      "loss": 2.7949,
      "theoretical_loss": 3.587277344450482,
      "tokens_seen": 1200680960
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.00032132397191574725,
      "loss": 2.8017,
      "theoretical_loss": 3.587259335842859,
      "tokens_seen": 1200746496
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0003213139418254765,
      "loss": 2.7304,
      "theoretical_loss": 3.5872413284933016,
      "tokens_seen": 1200812032
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0003213039117352056,
      "loss": 2.8873,
      "theoretical_loss": 3.5872233224016545,
      "tokens_seen": 1200877568
    },
    {
      "epoch": 4.0,
      "objective/train/docs_used": 1387236,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.934062957763672,
      "objective/train/theoretical_loss": 3.5872053175677605,
      "objective/train/tokens_used": 1221403104,
      "theoretical_loss": 3.5872053175677605,
      "tokens_seen": 1200943104
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.00032129388164493485,
      "loss": 2.7567,
      "theoretical_loss": 3.5872053175677605,
      "tokens_seen": 1200943104
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.000321283851554664,
      "loss": 2.6779,
      "theoretical_loss": 3.587187313991463,
      "tokens_seen": 1201008640
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0003212738214643932,
      "loss": 2.7564,
      "theoretical_loss": 3.587169311672606,
      "tokens_seen": 1201074176
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0003212637913741224,
      "loss": 2.8282,
      "theoretical_loss": 3.5871513106110333,
      "tokens_seen": 1201139712
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0003212537612838516,
      "loss": 2.5829,
      "theoretical_loss": 3.587133310806588,
      "tokens_seen": 1201205248
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.00032124373119358075,
      "loss": 2.6074,
      "theoretical_loss": 3.5871153122591135,
      "tokens_seen": 1201270784
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.00032123370110330994,
      "loss": 2.5451,
      "theoretical_loss": 3.587097314968454,
      "tokens_seen": 1201336320
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0003212236710130391,
      "loss": 2.7113,
      "theoretical_loss": 3.5870793189344523,
      "tokens_seen": 1201401856
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.00032121364092276835,
      "loss": 2.4835,
      "theoretical_loss": 3.587061324156954,
      "tokens_seen": 1201467392
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0003212036108324975,
      "loss": 2.5967,
      "theoretical_loss": 3.587043330635801,
      "tokens_seen": 1201532928
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0003211935807422267,
      "loss": 2.6397,
      "theoretical_loss": 3.5870253383708377,
      "tokens_seen": 1201598464
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0003211835506519559,
      "loss": 2.7351,
      "theoretical_loss": 3.5870073473619075,
      "tokens_seen": 1201664000
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0003211735205616851,
      "loss": 2.6823,
      "theoretical_loss": 3.5869893576088545,
      "tokens_seen": 1201729536
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.00032116349047141426,
      "loss": 2.5033,
      "theoretical_loss": 3.5869713691115237,
      "tokens_seen": 1201795072
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.00032115346038114344,
      "loss": 2.7195,
      "theoretical_loss": 3.586953381869757,
      "tokens_seen": 1201860608
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0003211434302908726,
      "loss": 2.691,
      "theoretical_loss": 3.5869353958833994,
      "tokens_seen": 1201926144
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.00032113340020060185,
      "loss": 2.7591,
      "theoretical_loss": 3.586917411152295,
      "tokens_seen": 1201991680
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.000321123370110331,
      "loss": 2.7759,
      "theoretical_loss": 3.586899427676287,
      "tokens_seen": 1202057216
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0003211133400200602,
      "loss": 2.5341,
      "theoretical_loss": 3.5868814454552203,
      "tokens_seen": 1202122752
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.00032110330992978934,
      "loss": 2.7228,
      "theoretical_loss": 3.586863464488938,
      "tokens_seen": 1202188288
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0003210932798395186,
      "loss": 2.7055,
      "theoretical_loss": 3.5868454847772853,
      "tokens_seen": 1202253824
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.00032108324974924776,
      "loss": 2.6004,
      "theoretical_loss": 3.586827506320105,
      "tokens_seen": 1202319360
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.00032107321965897694,
      "loss": 2.7184,
      "theoretical_loss": 3.5868095291172413,
      "tokens_seen": 1202384896
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0003210631895687061,
      "loss": 2.7188,
      "theoretical_loss": 3.5867915531685397,
      "tokens_seen": 1202450432
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0003210531594784353,
      "loss": 2.7433,
      "theoretical_loss": 3.586773578473843,
      "tokens_seen": 1202515968
    },
    {
      "epoch": 4.0,
      "objective/train/docs_used": 1392256,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8471696376800537,
      "objective/train/theoretical_loss": 3.586755605032996,
      "objective/train/tokens_used": 1223041504,
      "theoretical_loss": 3.586755605032996,
      "tokens_seen": 1202581504
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0003210431293881645,
      "loss": 2.7683,
      "theoretical_loss": 3.586755605032996,
      "tokens_seen": 1202581504
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0003210330992978937,
      "loss": 2.6809,
      "theoretical_loss": 3.5867376328458427,
      "tokens_seen": 1202647040
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.00032102306920762285,
      "loss": 2.7461,
      "theoretical_loss": 3.5867196619122272,
      "tokens_seen": 1202712576
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0003210130391173521,
      "loss": 2.7872,
      "theoretical_loss": 3.586701692231994,
      "tokens_seen": 1202778112
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.00032100300902708126,
      "loss": 2.7223,
      "theoretical_loss": 3.5866837238049873,
      "tokens_seen": 1202843648
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.00032099297893681044,
      "loss": 2.6087,
      "theoretical_loss": 3.586665756631052,
      "tokens_seen": 1202909184
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0003209829488465396,
      "loss": 2.6453,
      "theoretical_loss": 3.5866477907100314,
      "tokens_seen": 1202974720
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0003209729187562688,
      "loss": 2.6654,
      "theoretical_loss": 3.5866298260417704,
      "tokens_seen": 1203040256
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.000320962888665998,
      "loss": 2.6581,
      "theoretical_loss": 3.586611862626114,
      "tokens_seen": 1203105792
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0003209528585757272,
      "loss": 2.6549,
      "theoretical_loss": 3.5865939004629057,
      "tokens_seen": 1203171328
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.00032094282848545635,
      "loss": 2.7228,
      "theoretical_loss": 3.58657593955199,
      "tokens_seen": 1203236864
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0003209327983951856,
      "loss": 2.8043,
      "theoretical_loss": 3.586557979893212,
      "tokens_seen": 1203302400
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0003209227683049147,
      "loss": 2.5721,
      "theoretical_loss": 3.5865400214864156,
      "tokens_seen": 1203367936
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.00032091273821464395,
      "loss": 2.6316,
      "theoretical_loss": 3.5865220643314464,
      "tokens_seen": 1203433472
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.00032090270812437313,
      "loss": 2.6906,
      "theoretical_loss": 3.586504108428148,
      "tokens_seen": 1203499008
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003208926780341023,
      "loss": 2.6847,
      "theoretical_loss": 3.586486153776365,
      "tokens_seen": 1203564544
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003208826479438315,
      "loss": 2.7254,
      "theoretical_loss": 3.586468200375942,
      "tokens_seen": 1203630080
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003208726178535607,
      "loss": 2.7631,
      "theoretical_loss": 3.5864502482267246,
      "tokens_seen": 1203695616
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00032086258776328985,
      "loss": 2.6306,
      "theoretical_loss": 3.5864322973285567,
      "tokens_seen": 1203761152
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003208525576730191,
      "loss": 2.806,
      "theoretical_loss": 3.5864143476812824,
      "tokens_seen": 1203826688
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003208425275827482,
      "loss": 2.6084,
      "theoretical_loss": 3.586396399284748,
      "tokens_seen": 1203892224
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00032083249749247745,
      "loss": 2.7025,
      "theoretical_loss": 3.5863784521387974,
      "tokens_seen": 1203957760
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00032082246740220663,
      "loss": 2.7901,
      "theoretical_loss": 3.5863605062432753,
      "tokens_seen": 1204023296
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003208124373119358,
      "loss": 2.7009,
      "theoretical_loss": 3.586342561598027,
      "tokens_seen": 1204088832
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.000320802407221665,
      "loss": 2.6494,
      "theoretical_loss": 3.5863246182028963,
      "tokens_seen": 1204154368
    },
    {
      "epoch": 4.01,
      "objective/train/docs_used": 1397361,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8406195640563965,
      "objective/train/theoretical_loss": 3.5863066760577293,
      "objective/train/tokens_used": 1224679904,
      "theoretical_loss": 3.5863066760577293,
      "tokens_seen": 1204219904
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003207923771313942,
      "loss": 2.6758,
      "theoretical_loss": 3.5863066760577293,
      "tokens_seen": 1204219904
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00032078234704112336,
      "loss": 2.831,
      "theoretical_loss": 3.5862887351623702,
      "tokens_seen": 1204285440
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003207723169508526,
      "loss": 2.8496,
      "theoretical_loss": 3.5862707955166644,
      "tokens_seen": 1204350976
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003207622868605817,
      "loss": 2.6733,
      "theoretical_loss": 3.5862528571204564,
      "tokens_seen": 1204416512
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00032075225677031095,
      "loss": 2.7109,
      "theoretical_loss": 3.5862349199735912,
      "tokens_seen": 1204482048
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003207422266800401,
      "loss": 2.7204,
      "theoretical_loss": 3.5862169840759144,
      "tokens_seen": 1204547584
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003207321965897693,
      "loss": 2.6926,
      "theoretical_loss": 3.586199049427271,
      "tokens_seen": 1204613120
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003207221664994985,
      "loss": 2.6064,
      "theoretical_loss": 3.586181116027505,
      "tokens_seen": 1204678656
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003207121364092277,
      "loss": 2.5865,
      "theoretical_loss": 3.5861631838764625,
      "tokens_seen": 1204744192
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00032070210631895686,
      "loss": 2.5053,
      "theoretical_loss": 3.5861452529739886,
      "tokens_seen": 1204809728
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003206920762286861,
      "loss": 2.751,
      "theoretical_loss": 3.5861273233199276,
      "tokens_seen": 1204875264
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003206820461384152,
      "loss": 2.8412,
      "theoretical_loss": 3.586109394914126,
      "tokens_seen": 1204940800
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00032067201604814446,
      "loss": 2.79,
      "theoretical_loss": 3.586091467756428,
      "tokens_seen": 1205006336
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003206619859578736,
      "loss": 2.7373,
      "theoretical_loss": 3.586073541846679,
      "tokens_seen": 1205071872
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003206519558676028,
      "loss": 2.7581,
      "theoretical_loss": 3.5860556171847247,
      "tokens_seen": 1205137408
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.000320641925777332,
      "loss": 2.9763,
      "theoretical_loss": 3.58603769377041,
      "tokens_seen": 1205202944
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003206318956870612,
      "loss": 2.7726,
      "theoretical_loss": 3.5860197716035804,
      "tokens_seen": 1205268480
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00032062186559679036,
      "loss": 2.6571,
      "theoretical_loss": 3.586001850684081,
      "tokens_seen": 1205334016
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00032061183550651954,
      "loss": 2.6722,
      "theoretical_loss": 3.5859839310117576,
      "tokens_seen": 1205399552
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003206018054162487,
      "loss": 2.6794,
      "theoretical_loss": 3.5859660125864554,
      "tokens_seen": 1205465088
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00032059177532597796,
      "loss": 2.7808,
      "theoretical_loss": 3.5859480954080194,
      "tokens_seen": 1205530624
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003205817452357071,
      "loss": 2.7927,
      "theoretical_loss": 3.585930179476296,
      "tokens_seen": 1205596160
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003205717151454363,
      "loss": 2.6525,
      "theoretical_loss": 3.58591226479113,
      "tokens_seen": 1205661696
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003205616850551655,
      "loss": 2.7671,
      "theoretical_loss": 3.5858943513523664,
      "tokens_seen": 1205727232
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003205516549648947,
      "loss": 2.771,
      "theoretical_loss": 3.5858764391598523,
      "tokens_seen": 1205792768
    },
    {
      "epoch": 4.01,
      "objective/train/docs_used": 1402482,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.825957775115967,
      "objective/train/theoretical_loss": 3.585858528213432,
      "objective/train/tokens_used": 1226318304,
      "theoretical_loss": 3.585858528213432,
      "tokens_seen": 1205858304
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003205416248746239,
      "loss": 2.6716,
      "theoretical_loss": 3.585858528213432,
      "tokens_seen": 1205858304
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00032053159478435305,
      "loss": 2.5004,
      "theoretical_loss": 3.5858406185129517,
      "tokens_seen": 1205923840
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003205215646940823,
      "loss": 2.6681,
      "theoretical_loss": 3.585822710058257,
      "tokens_seen": 1205989376
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00032051153460381146,
      "loss": 2.6805,
      "theoretical_loss": 3.585804802849193,
      "tokens_seen": 1206054912
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00032050150451354064,
      "loss": 2.557,
      "theoretical_loss": 3.5857868968856055,
      "tokens_seen": 1206120448
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003204914744232698,
      "loss": 2.7872,
      "theoretical_loss": 3.585768992167341,
      "tokens_seen": 1206185984
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.000320481444332999,
      "loss": 2.696,
      "theoretical_loss": 3.585751088694244,
      "tokens_seen": 1206251520
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003204714142427282,
      "loss": 2.8453,
      "theoretical_loss": 3.585733186466161,
      "tokens_seen": 1206317056
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003204613841524574,
      "loss": 2.7537,
      "theoretical_loss": 3.585715285482938,
      "tokens_seen": 1206382592
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00032045135406218655,
      "loss": 2.7335,
      "theoretical_loss": 3.585697385744421,
      "tokens_seen": 1206448128
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003204413239719158,
      "loss": 2.2961,
      "theoretical_loss": 3.585679487250455,
      "tokens_seen": 1206513664
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003204312938816449,
      "loss": 2.6176,
      "theoretical_loss": 3.5856615900008864,
      "tokens_seen": 1206579200
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00032042126379137415,
      "loss": 2.553,
      "theoretical_loss": 3.5856436939955607,
      "tokens_seen": 1206644736
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00032041123370110333,
      "loss": 2.5838,
      "theoretical_loss": 3.5856257992343243,
      "tokens_seen": 1206710272
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003204012036108325,
      "loss": 2.813,
      "theoretical_loss": 3.5856079057170227,
      "tokens_seen": 1206775808
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003203911735205617,
      "loss": 2.719,
      "theoretical_loss": 3.585590013443502,
      "tokens_seen": 1206841344
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003203811434302909,
      "loss": 2.7019,
      "theoretical_loss": 3.585572122413609,
      "tokens_seen": 1206906880
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00032037111334002005,
      "loss": 2.4981,
      "theoretical_loss": 3.5855542326271888,
      "tokens_seen": 1206972416
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003203610832497493,
      "loss": 2.6035,
      "theoretical_loss": 3.5855363440840873,
      "tokens_seen": 1207037952
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003203510531594784,
      "loss": 2.6704,
      "theoretical_loss": 3.5855184567841514,
      "tokens_seen": 1207103488
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00032034102306920765,
      "loss": 2.5376,
      "theoretical_loss": 3.585500570727227,
      "tokens_seen": 1207169024
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00032033099297893683,
      "loss": 2.5535,
      "theoretical_loss": 3.58548268591316,
      "tokens_seen": 1207234560
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.000320320962888666,
      "loss": 2.7184,
      "theoretical_loss": 3.5854648023417965,
      "tokens_seen": 1207300096
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003203109327983952,
      "loss": 2.746,
      "theoretical_loss": 3.585446920012983,
      "tokens_seen": 1207365632
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003203009027081244,
      "loss": 2.7058,
      "theoretical_loss": 3.5854290389265655,
      "tokens_seen": 1207431168
    },
    {
      "epoch": 4.01,
      "objective/train/docs_used": 1407494,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6062726974487305,
      "objective/train/theoretical_loss": 3.5854111590823905,
      "objective/train/tokens_used": 1227956704,
      "theoretical_loss": 3.5854111590823905,
      "tokens_seen": 1207496704
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00032029087261785356,
      "loss": 2.688,
      "theoretical_loss": 3.5854111590823905,
      "tokens_seen": 1207496704
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003202808425275828,
      "loss": 2.8863,
      "theoretical_loss": 3.5853932804803037,
      "tokens_seen": 1207562240
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003202708124373119,
      "loss": 2.6487,
      "theoretical_loss": 3.5853754031201523,
      "tokens_seen": 1207627776
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00032026078234704115,
      "loss": 2.7036,
      "theoretical_loss": 3.585357527001782,
      "tokens_seen": 1207693312
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003202507522567703,
      "loss": 2.7126,
      "theoretical_loss": 3.5853396521250396,
      "tokens_seen": 1207758848
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003202407221664995,
      "loss": 2.543,
      "theoretical_loss": 3.585321778489771,
      "tokens_seen": 1207824384
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003202306920762287,
      "loss": 2.7911,
      "theoretical_loss": 3.585303906095823,
      "tokens_seen": 1207889920
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003202206619859579,
      "loss": 2.6455,
      "theoretical_loss": 3.585286034943042,
      "tokens_seen": 1207955456
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00032021063189568706,
      "loss": 2.6103,
      "theoretical_loss": 3.585268165031274,
      "tokens_seen": 1208020992
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003202006018054163,
      "loss": 2.6886,
      "theoretical_loss": 3.5852502963603667,
      "tokens_seen": 1208086528
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003201905717151454,
      "loss": 2.6231,
      "theoretical_loss": 3.585232428930165,
      "tokens_seen": 1208152064
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00032018054162487466,
      "loss": 2.7009,
      "theoretical_loss": 3.585214562740517,
      "tokens_seen": 1208217600
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003201705115346038,
      "loss": 2.8047,
      "theoretical_loss": 3.585196697791268,
      "tokens_seen": 1208283136
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.000320160481444333,
      "loss": 2.7623,
      "theoretical_loss": 3.5851788340822655,
      "tokens_seen": 1208348672
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003201504513540622,
      "loss": 2.7484,
      "theoretical_loss": 3.5851609716133552,
      "tokens_seen": 1208414208
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003201404212637914,
      "loss": 2.6339,
      "theoretical_loss": 3.585143110384385,
      "tokens_seen": 1208479744
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00032013039117352056,
      "loss": 2.6358,
      "theoretical_loss": 3.585125250395201,
      "tokens_seen": 1208545280
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00032012036108324974,
      "loss": 2.6491,
      "theoretical_loss": 3.58510739164565,
      "tokens_seen": 1208610816
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003201103309929789,
      "loss": 2.7829,
      "theoretical_loss": 3.585089534135578,
      "tokens_seen": 1208676352
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00032010030090270816,
      "loss": 2.5954,
      "theoretical_loss": 3.585071677864833,
      "tokens_seen": 1208741888
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003200902708124373,
      "loss": 2.6858,
      "theoretical_loss": 3.585053822833261,
      "tokens_seen": 1208807424
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003200802407221665,
      "loss": 2.7955,
      "theoretical_loss": 3.585035969040709,
      "tokens_seen": 1208872960
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00032007021063189565,
      "loss": 2.7785,
      "theoretical_loss": 3.5850181164870243,
      "tokens_seen": 1208938496
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003200601805416249,
      "loss": 2.6245,
      "theoretical_loss": 3.5850002651720527,
      "tokens_seen": 1209004032
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00032005015045135407,
      "loss": 2.5587,
      "theoretical_loss": 3.5849824150956424,
      "tokens_seen": 1209069568
    },
    {
      "epoch": 4.01,
      "objective/train/docs_used": 1410378,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7569308280944824,
      "objective/train/theoretical_loss": 3.5849645662576393,
      "objective/train/tokens_used": 1229595104,
      "theoretical_loss": 3.5849645662576393,
      "tokens_seen": 1209135104
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00032004012036108325,
      "loss": 2.7375,
      "theoretical_loss": 3.5849645662576393,
      "tokens_seen": 1209135104
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00032003009027081243,
      "loss": 2.5791,
      "theoretical_loss": 3.584946718657891,
      "tokens_seen": 1209200640
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00032002006018054166,
      "loss": 2.6087,
      "theoretical_loss": 3.5849288722962447,
      "tokens_seen": 1209266176
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003200100300902708,
      "loss": 2.8597,
      "theoretical_loss": 3.584911027172547,
      "tokens_seen": 1209331712
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00032,
      "loss": 2.6192,
      "theoretical_loss": 3.584893183286644,
      "tokens_seen": 1209397248
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031998996990972915,
      "loss": 2.6508,
      "theoretical_loss": 3.5848753406383844,
      "tokens_seen": 1209462784
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003199799398194584,
      "loss": 2.564,
      "theoretical_loss": 3.5848574992276148,
      "tokens_seen": 1209528320
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031996990972918757,
      "loss": 2.6464,
      "theoretical_loss": 3.584839659054182,
      "tokens_seen": 1209593856
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031995987963891675,
      "loss": 2.6648,
      "theoretical_loss": 3.5848218201179334,
      "tokens_seen": 1209659392
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031994984954864593,
      "loss": 2.698,
      "theoretical_loss": 3.584803982418716,
      "tokens_seen": 1209724928
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003199398194583751,
      "loss": 2.6806,
      "theoretical_loss": 3.584786145956377,
      "tokens_seen": 1209790464
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003199297893681043,
      "loss": 2.7552,
      "theoretical_loss": 3.584768310730764,
      "tokens_seen": 1209856000
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031991975927783353,
      "loss": 2.7092,
      "theoretical_loss": 3.584750476741724,
      "tokens_seen": 1209921536
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031990972918756266,
      "loss": 2.6622,
      "theoretical_loss": 3.5847326439891045,
      "tokens_seen": 1209987072
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003198996990972919,
      "loss": 2.8497,
      "theoretical_loss": 3.584714812472752,
      "tokens_seen": 1210052608
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.000319889669007021,
      "loss": 2.592,
      "theoretical_loss": 3.584696982192515,
      "tokens_seen": 1210118144
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031987963891675025,
      "loss": 2.5142,
      "theoretical_loss": 3.5846791531482403,
      "tokens_seen": 1210183680
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031986960882647943,
      "loss": 2.7154,
      "theoretical_loss": 3.584661325339775,
      "tokens_seen": 1210249216
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003198595787362086,
      "loss": 2.7834,
      "theoretical_loss": 3.584643498766968,
      "tokens_seen": 1210314752
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003198495486459378,
      "loss": 2.5158,
      "theoretical_loss": 3.584625673429665,
      "tokens_seen": 1210380288
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031983951855566703,
      "loss": 2.6208,
      "theoretical_loss": 3.584607849327714,
      "tokens_seen": 1210445824
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031982948846539616,
      "loss": 2.6352,
      "theoretical_loss": 3.5845900264609623,
      "tokens_seen": 1210511360
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003198194583751254,
      "loss": 2.7371,
      "theoretical_loss": 3.584572204829258,
      "tokens_seen": 1210576896
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003198094282848546,
      "loss": 2.5,
      "theoretical_loss": 3.5845543844324483,
      "tokens_seen": 1210642432
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031979939819458376,
      "loss": 2.6279,
      "theoretical_loss": 3.5845365652703816,
      "tokens_seen": 1210707968
    },
    {
      "epoch": 4.01,
      "objective/train/docs_used": 1413162,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.503354787826538,
      "objective/train/theoretical_loss": 3.584518747342904,
      "objective/train/tokens_used": 1231233504,
      "theoretical_loss": 3.584518747342904,
      "tokens_seen": 1210773504
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.000319789368104313,
      "loss": 2.5154,
      "theoretical_loss": 3.584518747342904,
      "tokens_seen": 1210773504
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003197793380140421,
      "loss": 2.8354,
      "theoretical_loss": 3.5845009306498645,
      "tokens_seen": 1210839040
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031976930792377135,
      "loss": 2.9319,
      "theoretical_loss": 3.58448311519111,
      "tokens_seen": 1210904576
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003197592778335005,
      "loss": 2.6632,
      "theoretical_loss": 3.584465300966489,
      "tokens_seen": 1210970112
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003197492477432297,
      "loss": 2.6491,
      "theoretical_loss": 3.584447487975848,
      "tokens_seen": 1211035648
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003197392176529589,
      "loss": 2.7853,
      "theoretical_loss": 3.5844296762190355,
      "tokens_seen": 1211101184
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003197291875626881,
      "loss": 2.7017,
      "theoretical_loss": 3.5844118656958996,
      "tokens_seen": 1211166720
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031971915747241726,
      "loss": 2.7758,
      "theoretical_loss": 3.5843940564062873,
      "tokens_seen": 1211232256
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003197091273821465,
      "loss": 2.8573,
      "theoretical_loss": 3.584376248350047,
      "tokens_seen": 1211297792
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003196990972918756,
      "loss": 2.6947,
      "theoretical_loss": 3.5843584415270264,
      "tokens_seen": 1211363328
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031968906720160486,
      "loss": 2.6019,
      "theoretical_loss": 3.5843406359370737,
      "tokens_seen": 1211428864
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.000319679037111334,
      "loss": 2.7002,
      "theoretical_loss": 3.584322831580036,
      "tokens_seen": 1211494400
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003196690070210632,
      "loss": 2.4958,
      "theoretical_loss": 3.584305028455762,
      "tokens_seen": 1211559936
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003196589769307924,
      "loss": 2.7174,
      "theoretical_loss": 3.5842872265640993,
      "tokens_seen": 1211625472
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003196489468405216,
      "loss": 2.6841,
      "theoretical_loss": 3.584269425904896,
      "tokens_seen": 1211691008
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031963891675025076,
      "loss": 2.6407,
      "theoretical_loss": 3.584251626478,
      "tokens_seen": 1211756544
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031962888665997994,
      "loss": 2.8254,
      "theoretical_loss": 3.58423382828326,
      "tokens_seen": 1211822080
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003196188565697091,
      "loss": 2.7335,
      "theoretical_loss": 3.5842160313205236,
      "tokens_seen": 1211887616
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031960882647943836,
      "loss": 2.6786,
      "theoretical_loss": 3.5841982355896382,
      "tokens_seen": 1211953152
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003195987963891675,
      "loss": 2.8014,
      "theoretical_loss": 3.584180441090453,
      "tokens_seen": 1212018688
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003195887662988967,
      "loss": 2.6308,
      "theoretical_loss": 3.584162647822816,
      "tokens_seen": 1212084224
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031957873620862585,
      "loss": 2.566,
      "theoretical_loss": 3.5841448557865743,
      "tokens_seen": 1212149760
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003195687061183551,
      "loss": 2.5022,
      "theoretical_loss": 3.584127064981578,
      "tokens_seen": 1212215296
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031955867602808427,
      "loss": 2.831,
      "theoretical_loss": 3.5841092754076733,
      "tokens_seen": 1212280832
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031954864593781345,
      "loss": 2.3898,
      "theoretical_loss": 3.5840914870647103,
      "tokens_seen": 1212346368
    },
    {
      "epoch": 4.01,
      "objective/train/docs_used": 1414282,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5722250938415527,
      "objective/train/theoretical_loss": 3.5840736999525356,
      "objective/train/tokens_used": 1232871904,
      "theoretical_loss": 3.5840736999525356,
      "tokens_seen": 1212411904
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031953861584754263,
      "loss": 2.7378,
      "theoretical_loss": 3.5840736999525356,
      "tokens_seen": 1212411904
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031952858575727186,
      "loss": 2.5702,
      "theoretical_loss": 3.5840559140709987,
      "tokens_seen": 1212477440
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.000319518555667001,
      "loss": 2.7093,
      "theoretical_loss": 3.5840381294199477,
      "tokens_seen": 1212542976
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003195085255767302,
      "loss": 2.8855,
      "theoretical_loss": 3.584020345999231,
      "tokens_seen": 1212608512
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031949849548645935,
      "loss": 2.7143,
      "theoretical_loss": 3.5840025638086965,
      "tokens_seen": 1212674048
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003194884653961886,
      "loss": 2.7684,
      "theoretical_loss": 3.5839847828481934,
      "tokens_seen": 1212739584
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031947843530591777,
      "loss": 2.6283,
      "theoretical_loss": 3.5839670031175697,
      "tokens_seen": 1212805120
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031946840521564695,
      "loss": 2.4975,
      "theoretical_loss": 3.5839492246166733,
      "tokens_seen": 1212870656
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031945837512537613,
      "loss": 2.7088,
      "theoretical_loss": 3.583931447345354,
      "tokens_seen": 1212936192
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003194483450351053,
      "loss": 2.6488,
      "theoretical_loss": 3.5839136713034594,
      "tokens_seen": 1213001728
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003194383149448345,
      "loss": 2.6018,
      "theoretical_loss": 3.583895896490838,
      "tokens_seen": 1213067264
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031942828485456373,
      "loss": 2.7103,
      "theoretical_loss": 3.5838781229073398,
      "tokens_seen": 1213132800
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031941825476429286,
      "loss": 2.6472,
      "theoretical_loss": 3.5838603505528113,
      "tokens_seen": 1213198336
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003194082246740221,
      "loss": 2.5772,
      "theoretical_loss": 3.5838425794271025,
      "tokens_seen": 1213263872
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003193981945837512,
      "loss": 2.4455,
      "theoretical_loss": 3.5838248095300616,
      "tokens_seen": 1213329408
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031938816449348045,
      "loss": 2.7764,
      "theoretical_loss": 3.583807040861538,
      "tokens_seen": 1213394944
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031937813440320964,
      "loss": 2.7773,
      "theoretical_loss": 3.583789273421379,
      "tokens_seen": 1213460480
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003193681043129388,
      "loss": 2.7144,
      "theoretical_loss": 3.583771507209435,
      "tokens_seen": 1213526016
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.000319358074222668,
      "loss": 2.6453,
      "theoretical_loss": 3.5837537422255537,
      "tokens_seen": 1213591552
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031934804413239723,
      "loss": 2.6259,
      "theoretical_loss": 3.583735978469584,
      "tokens_seen": 1213657088
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031933801404212636,
      "loss": 2.6677,
      "theoretical_loss": 3.5837182159413747,
      "tokens_seen": 1213722624
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003193279839518556,
      "loss": 2.5048,
      "theoretical_loss": 3.583700454640775,
      "tokens_seen": 1213788160
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003193179538615847,
      "loss": 2.6941,
      "theoretical_loss": 3.583682694567634,
      "tokens_seen": 1213853696
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031930792377131396,
      "loss": 2.821,
      "theoretical_loss": 3.5836649357218002,
      "tokens_seen": 1213919232
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031929789368104314,
      "loss": 2.8494,
      "theoretical_loss": 3.5836471781031225,
      "tokens_seen": 1213984768
    },
    {
      "epoch": 4.01,
      "objective/train/docs_used": 1415070,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7974677085876465,
      "objective/train/theoretical_loss": 3.58362942171145,
      "objective/train/tokens_used": 1234510304,
      "theoretical_loss": 3.58362942171145,
      "tokens_seen": 1214050304
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003192878635907723,
      "loss": 2.8067,
      "theoretical_loss": 3.58362942171145,
      "tokens_seen": 1214050304
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003192778335005015,
      "loss": 2.7528,
      "theoretical_loss": 3.5836116665466315,
      "tokens_seen": 1214115840
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003192678034102307,
      "loss": 2.6348,
      "theoretical_loss": 3.583593912608516,
      "tokens_seen": 1214181376
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031925777331995986,
      "loss": 2.5474,
      "theoretical_loss": 3.5835761598969524,
      "tokens_seen": 1214246912
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003192477432296891,
      "loss": 2.6987,
      "theoretical_loss": 3.5835584084117906,
      "tokens_seen": 1214312448
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003192377131394182,
      "loss": 2.6957,
      "theoretical_loss": 3.583540658152879,
      "tokens_seen": 1214377984
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031922768304914746,
      "loss": 2.6414,
      "theoretical_loss": 3.583522909120067,
      "tokens_seen": 1214443520
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003192176529588766,
      "loss": 2.6054,
      "theoretical_loss": 3.583505161313204,
      "tokens_seen": 1214509056
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003192076228686058,
      "loss": 2.7971,
      "theoretical_loss": 3.5834874147321383,
      "tokens_seen": 1214574592
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.000319197592778335,
      "loss": 2.8113,
      "theoretical_loss": 3.5834696693767194,
      "tokens_seen": 1214640128
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003191875626880642,
      "loss": 2.7911,
      "theoretical_loss": 3.583451925246797,
      "tokens_seen": 1214705664
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031917753259779337,
      "loss": 2.7277,
      "theoretical_loss": 3.58343418234222,
      "tokens_seen": 1214771200
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003191675025075226,
      "loss": 2.6605,
      "theoretical_loss": 3.583416440662838,
      "tokens_seen": 1214836736
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031915747241725173,
      "loss": 2.6873,
      "theoretical_loss": 3.5833987002085,
      "tokens_seen": 1214902272
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031914744232698096,
      "loss": 2.6275,
      "theoretical_loss": 3.5833809609790555,
      "tokens_seen": 1214967808
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003191374122367101,
      "loss": 2.6332,
      "theoretical_loss": 3.5833632229743535,
      "tokens_seen": 1215033344
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003191273821464393,
      "loss": 2.6789,
      "theoretical_loss": 3.5833454861942435,
      "tokens_seen": 1215098880
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003191173520561685,
      "loss": 2.7704,
      "theoretical_loss": 3.583327750638576,
      "tokens_seen": 1215164416
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003191073219658977,
      "loss": 2.7698,
      "theoretical_loss": 3.5833100163071987,
      "tokens_seen": 1215229952
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031909729187562687,
      "loss": 2.8002,
      "theoretical_loss": 3.5832922831999623,
      "tokens_seen": 1215295488
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031908726178535605,
      "loss": 2.6626,
      "theoretical_loss": 3.5832745513167152,
      "tokens_seen": 1215361024
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031907723169508523,
      "loss": 2.6527,
      "theoretical_loss": 3.5832568206573088,
      "tokens_seen": 1215426560
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031906720160481447,
      "loss": 2.6216,
      "theoretical_loss": 3.58323909122159,
      "tokens_seen": 1215492096
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031905717151454365,
      "loss": 2.6442,
      "theoretical_loss": 3.583221363009411,
      "tokens_seen": 1215557632
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031904714142427283,
      "loss": 2.6907,
      "theoretical_loss": 3.5832036360206203,
      "tokens_seen": 1215623168
    },
    {
      "epoch": 4.01,
      "objective/train/docs_used": 1416232,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.866744041442871,
      "objective/train/theoretical_loss": 3.5831859102550667,
      "objective/train/tokens_used": 1236148704,
      "theoretical_loss": 3.5831859102550667,
      "tokens_seen": 1215688704
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031903711133400206,
      "loss": 2.7375,
      "theoretical_loss": 3.5831859102550667,
      "tokens_seen": 1215688704
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003190270812437312,
      "loss": 2.9198,
      "theoretical_loss": 3.583168185712601,
      "tokens_seen": 1215754240
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003190170511534604,
      "loss": 2.5112,
      "theoretical_loss": 3.5831504623930726,
      "tokens_seen": 1215819776
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031900702106318955,
      "loss": 2.8218,
      "theoretical_loss": 3.583132740296331,
      "tokens_seen": 1215885312
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003189969909729188,
      "loss": 2.6373,
      "theoretical_loss": 3.583115019422226,
      "tokens_seen": 1215950848
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031898696088264797,
      "loss": 2.6703,
      "theoretical_loss": 3.5830972997706074,
      "tokens_seen": 1216016384
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031897693079237715,
      "loss": 2.6533,
      "theoretical_loss": 3.583079581341325,
      "tokens_seen": 1216081920
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031896690070210633,
      "loss": 2.705,
      "theoretical_loss": 3.583061864134229,
      "tokens_seen": 1216147456
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003189568706118355,
      "loss": 2.8154,
      "theoretical_loss": 3.583044148149168,
      "tokens_seen": 1216212992
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003189468405215647,
      "loss": 2.7847,
      "theoretical_loss": 3.583026433385993,
      "tokens_seen": 1216278528
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031893681043129393,
      "loss": 2.7064,
      "theoretical_loss": 3.583008719844554,
      "tokens_seen": 1216344064
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031892678034102306,
      "loss": 2.677,
      "theoretical_loss": 3.5829910075247007,
      "tokens_seen": 1216409600
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003189167502507523,
      "loss": 2.5897,
      "theoretical_loss": 3.5829732964262826,
      "tokens_seen": 1216475136
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003189067201604814,
      "loss": 2.6573,
      "theoretical_loss": 3.58295558654915,
      "tokens_seen": 1216540672
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031889669007021065,
      "loss": 2.7557,
      "theoretical_loss": 3.582937877893153,
      "tokens_seen": 1216606208
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031888665997993984,
      "loss": 2.7416,
      "theoretical_loss": 3.582920170458141,
      "tokens_seen": 1216671744
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.000318876629889669,
      "loss": 2.6819,
      "theoretical_loss": 3.582902464243965,
      "tokens_seen": 1216737280
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003188665997993982,
      "loss": 2.6083,
      "theoretical_loss": 3.582884759250474,
      "tokens_seen": 1216802816
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031885656970912743,
      "loss": 2.7381,
      "theoretical_loss": 3.5828670554775197,
      "tokens_seen": 1216868352
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031884653961885656,
      "loss": 2.8159,
      "theoretical_loss": 3.5828493529249505,
      "tokens_seen": 1216933888
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003188365095285858,
      "loss": 2.5582,
      "theoretical_loss": 3.5828316515926177,
      "tokens_seen": 1216999424
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003188264794383149,
      "loss": 2.7293,
      "theoretical_loss": 3.5828139514803707,
      "tokens_seen": 1217064960
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031881644934804416,
      "loss": 2.5561,
      "theoretical_loss": 3.582796252588061,
      "tokens_seen": 1217130496
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031880641925777334,
      "loss": 2.8685,
      "theoretical_loss": 3.582778554915537,
      "tokens_seen": 1217196032
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003187963891675025,
      "loss": 2.5221,
      "theoretical_loss": 3.5827608584626502,
      "tokens_seen": 1217261568
    },
    {
      "epoch": 4.01,
      "objective/train/docs_used": 1416846,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6486992835998535,
      "objective/train/theoretical_loss": 3.5827431632292503,
      "objective/train/tokens_used": 1237787104,
      "theoretical_loss": 3.5827431632292503,
      "tokens_seen": 1217327104
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003187863590772317,
      "loss": 2.7341,
      "theoretical_loss": 3.5827431632292503,
      "tokens_seen": 1217327104
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003187763289869609,
      "loss": 2.8365,
      "theoretical_loss": 3.5827254692151884,
      "tokens_seen": 1217392640
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031876629889669006,
      "loss": 2.6498,
      "theoretical_loss": 3.582707776420315,
      "tokens_seen": 1217458176
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003187562688064193,
      "loss": 2.532,
      "theoretical_loss": 3.5826900848444785,
      "tokens_seen": 1217523712
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003187462387161484,
      "loss": 2.4932,
      "theoretical_loss": 3.582672394487531,
      "tokens_seen": 1217589248
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031873620862587766,
      "loss": 2.7149,
      "theoretical_loss": 3.582654705349323,
      "tokens_seen": 1217654784
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003187261785356068,
      "loss": 2.8224,
      "theoretical_loss": 3.5826370174297044,
      "tokens_seen": 1217720320
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.000318716148445336,
      "loss": 2.4685,
      "theoretical_loss": 3.5826193307285257,
      "tokens_seen": 1217785856
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003187061183550652,
      "loss": 2.5885,
      "theoretical_loss": 3.5826016452456377,
      "tokens_seen": 1217851392
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003186960882647944,
      "loss": 2.7836,
      "theoretical_loss": 3.5825839609808905,
      "tokens_seen": 1217916928
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031868605817452357,
      "loss": 2.8991,
      "theoretical_loss": 3.5825662779341347,
      "tokens_seen": 1217982464
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003186760280842528,
      "loss": 2.5322,
      "theoretical_loss": 3.5825485961052213,
      "tokens_seen": 1218048000
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031866599799398193,
      "loss": 2.5996,
      "theoretical_loss": 3.582530915494001,
      "tokens_seen": 1218113536
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031865596790371116,
      "loss": 2.7052,
      "theoretical_loss": 3.5825132361003234,
      "tokens_seen": 1218179072
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003186459378134403,
      "loss": 2.6446,
      "theoretical_loss": 3.58249555792404,
      "tokens_seen": 1218244608
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003186359077231695,
      "loss": 2.7679,
      "theoretical_loss": 3.5824778809650017,
      "tokens_seen": 1218310144
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003186258776328987,
      "loss": 2.6478,
      "theoretical_loss": 3.582460205223059,
      "tokens_seen": 1218375680
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003186158475426279,
      "loss": 2.5404,
      "theoretical_loss": 3.582442530698062,
      "tokens_seen": 1218441216
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031860581745235707,
      "loss": 2.7493,
      "theoretical_loss": 3.5824248573898623,
      "tokens_seen": 1218506752
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031859578736208625,
      "loss": 2.6604,
      "theoretical_loss": 3.58240718529831,
      "tokens_seen": 1218572288
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031858575727181543,
      "loss": 2.6678,
      "theoretical_loss": 3.582389514423257,
      "tokens_seen": 1218637824
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031857572718154467,
      "loss": 2.9431,
      "theoretical_loss": 3.5823718447645527,
      "tokens_seen": 1218703360
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003185656970912738,
      "loss": 2.8017,
      "theoretical_loss": 3.582354176322049,
      "tokens_seen": 1218768896
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031855566700100303,
      "loss": 2.5855,
      "theoretical_loss": 3.5823365090955965,
      "tokens_seen": 1218834432
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003185456369107322,
      "loss": 2.7062,
      "theoretical_loss": 3.582318843085046,
      "tokens_seen": 1218899968
    },
    {
      "epoch": 4.01,
      "objective/train/docs_used": 1418472,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8614342212677,
      "objective/train/theoretical_loss": 3.5823011782902485,
      "objective/train/tokens_used": 1239425504,
      "theoretical_loss": 3.5823011782902485,
      "tokens_seen": 1218965504
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003185356068204614,
      "loss": 2.7619,
      "theoretical_loss": 3.5823011782902485,
      "tokens_seen": 1218965504
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031852557673019057,
      "loss": 2.8181,
      "theoretical_loss": 3.5822835147110554,
      "tokens_seen": 1219031040
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031851554663991975,
      "loss": 2.5943,
      "theoretical_loss": 3.582265852347317,
      "tokens_seen": 1219096576
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031850551654964893,
      "loss": 2.6939,
      "theoretical_loss": 3.582248191198885,
      "tokens_seen": 1219162112
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031849548645937817,
      "loss": 2.5154,
      "theoretical_loss": 3.58223053126561,
      "tokens_seen": 1219227648
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003184854563691073,
      "loss": 2.5464,
      "theoretical_loss": 3.5822128725473434,
      "tokens_seen": 1219293184
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031847542627883653,
      "loss": 2.5821,
      "theoretical_loss": 3.582195215043936,
      "tokens_seen": 1219358720
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031846539618856566,
      "loss": 2.4363,
      "theoretical_loss": 3.582177558755239,
      "tokens_seen": 1219424256
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003184553660982949,
      "loss": 2.6333,
      "theoretical_loss": 3.5821599036811036,
      "tokens_seen": 1219489792
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003184453360080241,
      "loss": 2.7076,
      "theoretical_loss": 3.582142249821381,
      "tokens_seen": 1219555328
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031843530591775326,
      "loss": 2.626,
      "theoretical_loss": 3.5821245971759224,
      "tokens_seen": 1219620864
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031842527582748244,
      "loss": 2.4653,
      "theoretical_loss": 3.582106945744579,
      "tokens_seen": 1219686400
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003184152457372116,
      "loss": 2.5816,
      "theoretical_loss": 3.5820892955272026,
      "tokens_seen": 1219751936
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003184052156469408,
      "loss": 2.7463,
      "theoretical_loss": 3.5820716465236435,
      "tokens_seen": 1219817472
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031839518555667004,
      "loss": 2.65,
      "theoretical_loss": 3.5820539987337536,
      "tokens_seen": 1219883008
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031838515546639916,
      "loss": 2.8067,
      "theoretical_loss": 3.5820363521573846,
      "tokens_seen": 1219948544
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003183751253761284,
      "loss": 2.7794,
      "theoretical_loss": 3.582018706794387,
      "tokens_seen": 1220014080
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003183650952858576,
      "loss": 2.6482,
      "theoretical_loss": 3.582001062644613,
      "tokens_seen": 1220079616
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031835506519558676,
      "loss": 2.542,
      "theoretical_loss": 3.5819834197079135,
      "tokens_seen": 1220145152
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031834503510531594,
      "loss": 2.5878,
      "theoretical_loss": 3.58196577798414,
      "tokens_seen": 1220210688
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003183350050150451,
      "loss": 2.4764,
      "theoretical_loss": 3.581948137473144,
      "tokens_seen": 1220276224
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003183249749247743,
      "loss": 2.6947,
      "theoretical_loss": 3.5819304981747777,
      "tokens_seen": 1220341760
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031831494483450354,
      "loss": 2.6588,
      "theoretical_loss": 3.5819128600888916,
      "tokens_seen": 1220407296
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003183049147442327,
      "loss": 2.7199,
      "theoretical_loss": 3.581895223215337,
      "tokens_seen": 1220472832
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003182948846539619,
      "loss": 2.6414,
      "theoretical_loss": 3.5818775875539672,
      "tokens_seen": 1220538368
    },
    {
      "epoch": 4.01,
      "objective/train/docs_used": 1419245,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.898358106613159,
      "objective/train/theoretical_loss": 3.5818599531046322,
      "objective/train/tokens_used": 1241063904,
      "theoretical_loss": 3.5818599531046322,
      "tokens_seen": 1220603904
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003182848545636911,
      "loss": 2.807,
      "theoretical_loss": 3.5818599531046322,
      "tokens_seen": 1220603904
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031827482447342026,
      "loss": 2.5729,
      "theoretical_loss": 3.5818423198671847,
      "tokens_seen": 1220669440
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003182647943831495,
      "loss": 2.6056,
      "theoretical_loss": 3.5818246878414755,
      "tokens_seen": 1220734976
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003182547642928786,
      "loss": 2.5781,
      "theoretical_loss": 3.5818070570273566,
      "tokens_seen": 1220800512
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031824473420260786,
      "loss": 2.726,
      "theoretical_loss": 3.58178942742468,
      "tokens_seen": 1220866048
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.000318234704112337,
      "loss": 2.7488,
      "theoretical_loss": 3.581771799033297,
      "tokens_seen": 1220931584
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003182246740220662,
      "loss": 2.4787,
      "theoretical_loss": 3.58175417185306,
      "tokens_seen": 1220997120
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003182146439317954,
      "loss": 2.6335,
      "theoretical_loss": 3.5817365458838193,
      "tokens_seen": 1221062656
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003182046138415246,
      "loss": 2.6146,
      "theoretical_loss": 3.581718921125429,
      "tokens_seen": 1221128192
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031819458375125377,
      "loss": 2.8097,
      "theoretical_loss": 3.5817012975777387,
      "tokens_seen": 1221193728
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.000318184553660983,
      "loss": 2.6973,
      "theoretical_loss": 3.5816836752406016,
      "tokens_seen": 1221259264
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031817452357071213,
      "loss": 2.677,
      "theoretical_loss": 3.581666054113869,
      "tokens_seen": 1221324800
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031816449348044136,
      "loss": 2.6547,
      "theoretical_loss": 3.5816484341973935,
      "tokens_seen": 1221390336
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003181544633901705,
      "loss": 2.465,
      "theoretical_loss": 3.5816308154910264,
      "tokens_seen": 1221455872
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003181444332998997,
      "loss": 2.6044,
      "theoretical_loss": 3.58161319799462,
      "tokens_seen": 1221521408
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003181344032096289,
      "loss": 2.7337,
      "theoretical_loss": 3.5815955817080263,
      "tokens_seen": 1221586944
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003181243731193581,
      "loss": 2.8371,
      "theoretical_loss": 3.581577966631097,
      "tokens_seen": 1221652480
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031811434302908727,
      "loss": 2.5997,
      "theoretical_loss": 3.5815603527636846,
      "tokens_seen": 1221718016
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031810431293881645,
      "loss": 2.5927,
      "theoretical_loss": 3.5815427401056406,
      "tokens_seen": 1221783552
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031809428284854563,
      "loss": 2.7329,
      "theoretical_loss": 3.5815251286568177,
      "tokens_seen": 1221849088
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031808425275827487,
      "loss": 2.6258,
      "theoretical_loss": 3.5815075184170673,
      "tokens_seen": 1221914624
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.000318074222668004,
      "loss": 2.8222,
      "theoretical_loss": 3.5814899093862422,
      "tokens_seen": 1221980160
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031806419257773323,
      "loss": 3.0164,
      "theoretical_loss": 3.5814723015641947,
      "tokens_seen": 1222045696
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003180541624874624,
      "loss": 2.7932,
      "theoretical_loss": 3.5814546949507764,
      "tokens_seen": 1222111232
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003180441323971916,
      "loss": 2.6886,
      "theoretical_loss": 3.5814370895458403,
      "tokens_seen": 1222176768
    },
    {
      "epoch": 4.01,
      "objective/train/docs_used": 1420806,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.511540174484253,
      "objective/train/theoretical_loss": 3.5814194853492376,
      "objective/train/tokens_used": 1242702304,
      "theoretical_loss": 3.5814194853492376,
      "tokens_seen": 1222242304
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031803410230692077,
      "loss": 2.626,
      "theoretical_loss": 3.5814194853492376,
      "tokens_seen": 1222242304
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031802407221664995,
      "loss": 2.7816,
      "theoretical_loss": 3.5814018823608214,
      "tokens_seen": 1222307840
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031801404212637913,
      "loss": 2.6152,
      "theoretical_loss": 3.5813842805804437,
      "tokens_seen": 1222373376
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031800401203610837,
      "loss": 2.6969,
      "theoretical_loss": 3.581366680007957,
      "tokens_seen": 1222438912
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003179939819458375,
      "loss": 2.663,
      "theoretical_loss": 3.581349080643214,
      "tokens_seen": 1222504448
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031798395185556673,
      "loss": 2.8343,
      "theoretical_loss": 3.581331482486066,
      "tokens_seen": 1222569984
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031797392176529586,
      "loss": 2.7702,
      "theoretical_loss": 3.581313885536366,
      "tokens_seen": 1222635520
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003179638916750251,
      "loss": 2.7763,
      "theoretical_loss": 3.581296289793967,
      "tokens_seen": 1222701056
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003179538615847543,
      "loss": 2.8597,
      "theoretical_loss": 3.581278695258721,
      "tokens_seen": 1222766592
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031794383149448346,
      "loss": 2.7061,
      "theoretical_loss": 3.58126110193048,
      "tokens_seen": 1222832128
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031793380140421264,
      "loss": 2.643,
      "theoretical_loss": 3.5812435098090973,
      "tokens_seen": 1222897664
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003179237713139418,
      "loss": 2.85,
      "theoretical_loss": 3.5812259188944253,
      "tokens_seen": 1222963200
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.000317913741223671,
      "loss": 2.6904,
      "theoretical_loss": 3.5812083291863157,
      "tokens_seen": 1223028736
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031790371113340024,
      "loss": 2.6822,
      "theoretical_loss": 3.5811907406846224,
      "tokens_seen": 1223094272
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031789368104312936,
      "loss": 2.7356,
      "theoretical_loss": 3.5811731533891975,
      "tokens_seen": 1223159808
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003178836509528586,
      "loss": 2.7586,
      "theoretical_loss": 3.581155567299893,
      "tokens_seen": 1223225344
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003178736208625878,
      "loss": 2.8288,
      "theoretical_loss": 3.581137982416563,
      "tokens_seen": 1223290880
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031786359077231696,
      "loss": 2.5823,
      "theoretical_loss": 3.581120398739059,
      "tokens_seen": 1223356416
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031785356068204614,
      "loss": 2.5514,
      "theoretical_loss": 3.581102816267234,
      "tokens_seen": 1223421952
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003178435305917753,
      "loss": 2.6135,
      "theoretical_loss": 3.5810852350009403,
      "tokens_seen": 1223487488
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003178335005015045,
      "loss": 2.6177,
      "theoretical_loss": 3.5810676549400315,
      "tokens_seen": 1223553024
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031782347041123374,
      "loss": 2.7085,
      "theoretical_loss": 3.5810500760843604,
      "tokens_seen": 1223618560
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031781344032096287,
      "loss": 2.257,
      "theoretical_loss": 3.581032498433779,
      "tokens_seen": 1223684096
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003178034102306921,
      "loss": 2.7905,
      "theoretical_loss": 3.581014921988141,
      "tokens_seen": 1223749632
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031779338014042123,
      "loss": 2.6344,
      "theoretical_loss": 3.580997346747299,
      "tokens_seen": 1223815168
    },
    {
      "epoch": 4.01,
      "objective/train/docs_used": 1421310,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.000223159790039,
      "objective/train/theoretical_loss": 3.580979772711106,
      "objective/train/tokens_used": 1244340704,
      "theoretical_loss": 3.580979772711106,
      "tokens_seen": 1223880704
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031778335005015046,
      "loss": 2.6517,
      "theoretical_loss": 3.580979772711106,
      "tokens_seen": 1223880704
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031777331995987964,
      "loss": 2.7078,
      "theoretical_loss": 3.580962199879414,
      "tokens_seen": 1223946240
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003177632898696088,
      "loss": 2.8244,
      "theoretical_loss": 3.580944628252077,
      "tokens_seen": 1224011776
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.000317753259779338,
      "loss": 2.7608,
      "theoretical_loss": 3.5809270578289483,
      "tokens_seen": 1224077312
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003177432296890672,
      "loss": 2.671,
      "theoretical_loss": 3.58090948860988,
      "tokens_seen": 1224142848
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031773319959879637,
      "loss": 2.6531,
      "theoretical_loss": 3.580891920594725,
      "tokens_seen": 1224208384
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003177231695085256,
      "loss": 2.5167,
      "theoretical_loss": 3.5808743537833374,
      "tokens_seen": 1224273920
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031771313941825473,
      "loss": 2.7415,
      "theoretical_loss": 3.58085678817557,
      "tokens_seen": 1224339456
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031770310932798397,
      "loss": 2.5674,
      "theoretical_loss": 3.580839223771275,
      "tokens_seen": 1224404992
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031769307923771315,
      "loss": 2.7093,
      "theoretical_loss": 3.5808216605703063,
      "tokens_seen": 1224470528
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031768304914744233,
      "loss": 2.5677,
      "theoretical_loss": 3.580804098572517,
      "tokens_seen": 1224536064
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003176730190571715,
      "loss": 2.8175,
      "theoretical_loss": 3.58078653777776,
      "tokens_seen": 1224601600
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003176629889669007,
      "loss": 2.701,
      "theoretical_loss": 3.580768978185889,
      "tokens_seen": 1224667136
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031765295887662987,
      "loss": 2.7692,
      "theoretical_loss": 3.5807514197967567,
      "tokens_seen": 1224732672
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003176429287863591,
      "loss": 2.7475,
      "theoretical_loss": 3.580733862610217,
      "tokens_seen": 1224798208
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031763289869608823,
      "loss": 2.8054,
      "theoretical_loss": 3.5807163066261225,
      "tokens_seen": 1224863744
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031762286860581747,
      "loss": 2.6037,
      "theoretical_loss": 3.580698751844327,
      "tokens_seen": 1224929280
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003176128385155466,
      "loss": 2.5773,
      "theoretical_loss": 3.580681198264684,
      "tokens_seen": 1224994816
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031760280842527583,
      "loss": 2.7677,
      "theoretical_loss": 3.580663645887046,
      "tokens_seen": 1225060352
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.000317592778335005,
      "loss": 2.6867,
      "theoretical_loss": 3.580646094711267,
      "tokens_seen": 1225125888
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003175827482447342,
      "loss": 2.708,
      "theoretical_loss": 3.5806285447372006,
      "tokens_seen": 1225191424
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003175727181544634,
      "loss": 2.8949,
      "theoretical_loss": 3.5806109959647,
      "tokens_seen": 1225256960
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003175626880641926,
      "loss": 2.3915,
      "theoretical_loss": 3.5805934483936186,
      "tokens_seen": 1225322496
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003175526579739218,
      "loss": 2.7247,
      "theoretical_loss": 3.58057590202381,
      "tokens_seen": 1225388032
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031754262788365097,
      "loss": 2.745,
      "theoretical_loss": 3.5805583568551276,
      "tokens_seen": 1225453568
    },
    {
      "epoch": 4.01,
      "objective/train/docs_used": 1422763,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5950467586517334,
      "objective/train/theoretical_loss": 3.5805408128874254,
      "objective/train/tokens_used": 1245979104,
      "theoretical_loss": 3.5805408128874254,
      "tokens_seen": 1225519104
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031753259779338015,
      "loss": 2.619,
      "theoretical_loss": 3.5805408128874254,
      "tokens_seen": 1225519104
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031752256770310933,
      "loss": 2.767,
      "theoretical_loss": 3.5805232701205565,
      "tokens_seen": 1225584640
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031751253761283857,
      "loss": 2.8689,
      "theoretical_loss": 3.580505728554374,
      "tokens_seen": 1225650176
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003175025075225677,
      "loss": 2.6551,
      "theoretical_loss": 3.580488188188733,
      "tokens_seen": 1225715712
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031749247743229693,
      "loss": 2.7369,
      "theoretical_loss": 3.5804706490234857,
      "tokens_seen": 1225781248
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031748244734202606,
      "loss": 2.4896,
      "theoretical_loss": 3.580453111058487,
      "tokens_seen": 1225846784
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003174724172517553,
      "loss": 2.6095,
      "theoretical_loss": 3.58043557429359,
      "tokens_seen": 1225912320
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003174623871614845,
      "loss": 2.8006,
      "theoretical_loss": 3.5804180387286477,
      "tokens_seen": 1225977856
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031745235707121366,
      "loss": 2.6738,
      "theoretical_loss": 3.5804005043635154,
      "tokens_seen": 1226043392
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031744232698094284,
      "loss": 2.7993,
      "theoretical_loss": 3.5803829711980453,
      "tokens_seen": 1226108928
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.000317432296890672,
      "loss": 2.5852,
      "theoretical_loss": 3.580365439232092,
      "tokens_seen": 1226174464
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003174222668004012,
      "loss": 2.6334,
      "theoretical_loss": 3.5803479084655097,
      "tokens_seen": 1226240000
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031741223671013044,
      "loss": 2.8156,
      "theoretical_loss": 3.580330378898152,
      "tokens_seen": 1226305536
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031740220661985956,
      "loss": 2.5313,
      "theoretical_loss": 3.5803128505298725,
      "tokens_seen": 1226371072
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003173921765295888,
      "loss": 2.848,
      "theoretical_loss": 3.5802953233605246,
      "tokens_seen": 1226436608
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.000317382146439318,
      "loss": 2.7893,
      "theoretical_loss": 3.5802777973899635,
      "tokens_seen": 1226502144
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031737211634904716,
      "loss": 2.5006,
      "theoretical_loss": 3.580260272618043,
      "tokens_seen": 1226567680
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031736208625877634,
      "loss": 2.6533,
      "theoretical_loss": 3.580242749044616,
      "tokens_seen": 1226633216
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003173520561685055,
      "loss": 2.6162,
      "theoretical_loss": 3.580225226669537,
      "tokens_seen": 1226698752
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003173420260782347,
      "loss": 2.7069,
      "theoretical_loss": 3.5802077054926604,
      "tokens_seen": 1226764288
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031733199598796394,
      "loss": 2.8578,
      "theoretical_loss": 3.58019018551384,
      "tokens_seen": 1226829824
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031732196589769307,
      "loss": 2.5383,
      "theoretical_loss": 3.5801726667329294,
      "tokens_seen": 1226895360
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003173119358074223,
      "loss": 2.6596,
      "theoretical_loss": 3.5801551491497836,
      "tokens_seen": 1226960896
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031730190571715143,
      "loss": 2.6207,
      "theoretical_loss": 3.5801376327642567,
      "tokens_seen": 1227026432
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031729187562688066,
      "loss": 2.6281,
      "theoretical_loss": 3.5801201175762025,
      "tokens_seen": 1227091968
    },
    {
      "epoch": 4.01,
      "objective/train/docs_used": 1423509,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.779162645339966,
      "objective/train/theoretical_loss": 3.580102603585475,
      "objective/train/tokens_used": 1247617504,
      "theoretical_loss": 3.580102603585475,
      "tokens_seen": 1227157504
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031728184553660984,
      "loss": 2.7531,
      "theoretical_loss": 3.580102603585475,
      "tokens_seen": 1227157504
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.000317271815446339,
      "loss": 2.6181,
      "theoretical_loss": 3.5800850907919286,
      "tokens_seen": 1227223040
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003172617853560682,
      "loss": 2.6957,
      "theoretical_loss": 3.5800675791954175,
      "tokens_seen": 1227288576
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003172517552657974,
      "loss": 2.5925,
      "theoretical_loss": 3.5800500687957966,
      "tokens_seen": 1227354112
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031724172517552657,
      "loss": 2.5597,
      "theoretical_loss": 3.5800325595929188,
      "tokens_seen": 1227419648
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003172316950852558,
      "loss": 2.7134,
      "theoretical_loss": 3.58001505158664,
      "tokens_seen": 1227485184
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031722166499498493,
      "loss": 2.4009,
      "theoretical_loss": 3.5799975447768135,
      "tokens_seen": 1227550720
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031721163490471417,
      "loss": 2.7466,
      "theoretical_loss": 3.579980039163294,
      "tokens_seen": 1227616256
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031720160481444335,
      "loss": 2.7712,
      "theoretical_loss": 3.5799625347459356,
      "tokens_seen": 1227681792
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031719157472417253,
      "loss": 2.5761,
      "theoretical_loss": 3.579945031524593,
      "tokens_seen": 1227747328
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003171815446339017,
      "loss": 2.5698,
      "theoretical_loss": 3.5799275294991206,
      "tokens_seen": 1227812864
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003171715145436309,
      "loss": 2.7248,
      "theoretical_loss": 3.5799100286693735,
      "tokens_seen": 1227878400
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031716148445336007,
      "loss": 2.7414,
      "theoretical_loss": 3.579892529035205,
      "tokens_seen": 1227943936
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003171514543630893,
      "loss": 2.6472,
      "theoretical_loss": 3.57987503059647,
      "tokens_seen": 1228009472
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031714142427281843,
      "loss": 2.7689,
      "theoretical_loss": 3.579857533353024,
      "tokens_seen": 1228075008
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031713139418254767,
      "loss": 2.5172,
      "theoretical_loss": 3.57984003730472,
      "tokens_seen": 1228140544
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003171213640922768,
      "loss": 2.7424,
      "theoretical_loss": 3.5798225424514145,
      "tokens_seen": 1228206080
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031711133400200603,
      "loss": 2.6671,
      "theoretical_loss": 3.5798050487929602,
      "tokens_seen": 1228271616
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003171013039117352,
      "loss": 2.4961,
      "theoretical_loss": 3.5797875563292125,
      "tokens_seen": 1228337152
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003170912738214644,
      "loss": 2.8536,
      "theoretical_loss": 3.5797700650600266,
      "tokens_seen": 1228402688
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003170812437311936,
      "loss": 2.6944,
      "theoretical_loss": 3.579752574985257,
      "tokens_seen": 1228468224
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003170712136409228,
      "loss": 2.6289,
      "theoretical_loss": 3.5797350861047574,
      "tokens_seen": 1228533760
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031706118355065194,
      "loss": 2.8619,
      "theoretical_loss": 3.5797175984183838,
      "tokens_seen": 1228599296
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003170511534603812,
      "loss": 2.6812,
      "theoretical_loss": 3.5797001119259906,
      "tokens_seen": 1228664832
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003170411233701103,
      "loss": 2.5994,
      "theoretical_loss": 3.5796826266274326,
      "tokens_seen": 1228730368
    },
    {
      "epoch": 4.01,
      "objective/train/docs_used": 1425069,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5875046253204346,
      "objective/train/theoretical_loss": 3.5796651425225643,
      "objective/train/tokens_used": 1249255904,
      "theoretical_loss": 3.5796651425225643,
      "tokens_seen": 1228795904
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031703109327983954,
      "loss": 2.6018,
      "theoretical_loss": 3.5796651425225643,
      "tokens_seen": 1228795904
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003170210631895687,
      "loss": 2.8186,
      "theoretical_loss": 3.5796476596112408,
      "tokens_seen": 1228861440
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003170110330992979,
      "loss": 2.7534,
      "theoretical_loss": 3.5796301778933173,
      "tokens_seen": 1228926976
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003170010030090271,
      "loss": 2.7221,
      "theoretical_loss": 3.579612697368648,
      "tokens_seen": 1228992512
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031699097291875626,
      "loss": 2.6292,
      "theoretical_loss": 3.5795952180370882,
      "tokens_seen": 1229058048
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031698094282848544,
      "loss": 2.5586,
      "theoretical_loss": 3.579577739898493,
      "tokens_seen": 1229123584
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003169709127382147,
      "loss": 2.5425,
      "theoretical_loss": 3.5795602629527172,
      "tokens_seen": 1229189120
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003169608826479438,
      "loss": 2.4003,
      "theoretical_loss": 3.5795427871996157,
      "tokens_seen": 1229254656
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031695085255767304,
      "loss": 2.7679,
      "theoretical_loss": 3.579525312639044,
      "tokens_seen": 1229320192
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031694082246740217,
      "loss": 2.8011,
      "theoretical_loss": 3.579507839270857,
      "tokens_seen": 1229385728
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003169307923771314,
      "loss": 2.7587,
      "theoretical_loss": 3.579490367094909,
      "tokens_seen": 1229451264
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003169207622868606,
      "loss": 2.7321,
      "theoretical_loss": 3.579472896111056,
      "tokens_seen": 1229516800
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031691073219658976,
      "loss": 2.8625,
      "theoretical_loss": 3.5794554263191527,
      "tokens_seen": 1229582336
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031690070210631894,
      "loss": 2.5965,
      "theoretical_loss": 3.5794379577190547,
      "tokens_seen": 1229647872
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003168906720160482,
      "loss": 2.7067,
      "theoretical_loss": 3.579420490310617,
      "tokens_seen": 1229713408
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003168806419257773,
      "loss": 2.8471,
      "theoretical_loss": 3.5794030240936943,
      "tokens_seen": 1229778944
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031687061183550654,
      "loss": 2.7518,
      "theoretical_loss": 3.5793855590681423,
      "tokens_seen": 1229844480
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031686058174523567,
      "loss": 2.8173,
      "theoretical_loss": 3.579368095233816,
      "tokens_seen": 1229910016
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003168505516549649,
      "loss": 2.6816,
      "theoretical_loss": 3.579350632590571,
      "tokens_seen": 1229975552
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003168405215646941,
      "loss": 2.7107,
      "theoretical_loss": 3.5793331711382628,
      "tokens_seen": 1230041088
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031683049147442327,
      "loss": 2.7466,
      "theoretical_loss": 3.5793157108767457,
      "tokens_seen": 1230106624
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031682046138415245,
      "loss": 2.6286,
      "theoretical_loss": 3.5792982518058762,
      "tokens_seen": 1230172160
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031681043129388163,
      "loss": 2.683,
      "theoretical_loss": 3.579280793925509,
      "tokens_seen": 1230237696
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031680040120361086,
      "loss": 2.5633,
      "theoretical_loss": 3.5792633372355,
      "tokens_seen": 1230303232
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031679037111334004,
      "loss": 2.626,
      "theoretical_loss": 3.5792458817357042,
      "tokens_seen": 1230368768
    },
    {
      "epoch": 4.01,
      "objective/train/docs_used": 1425869,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.124709367752075,
      "objective/train/theoretical_loss": 3.579228427425978,
      "objective/train/tokens_used": 1250894304,
      "theoretical_loss": 3.579228427425978,
      "tokens_seen": 1230434304
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003167803410230692,
      "loss": 2.5447,
      "theoretical_loss": 3.579228427425978,
      "tokens_seen": 1230434304
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003167703109327984,
      "loss": 2.7377,
      "theoretical_loss": 3.579210974306175,
      "tokens_seen": 1230499840
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003167602808425276,
      "loss": 2.8242,
      "theoretical_loss": 3.5791935223761526,
      "tokens_seen": 1230565376
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031675025075225677,
      "loss": 2.8878,
      "theoretical_loss": 3.5791760716357652,
      "tokens_seen": 1230630912
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.000316740220661986,
      "loss": 2.6552,
      "theoretical_loss": 3.579158622084869,
      "tokens_seen": 1230696448
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031673019057171513,
      "loss": 2.7781,
      "theoretical_loss": 3.579141173723319,
      "tokens_seen": 1230761984
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031672016048144437,
      "loss": 2.572,
      "theoretical_loss": 3.5791237265509714,
      "tokens_seen": 1230827520
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031671013039117355,
      "loss": 2.5123,
      "theoretical_loss": 3.5791062805676814,
      "tokens_seen": 1230893056
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031670010030090273,
      "loss": 2.4936,
      "theoretical_loss": 3.5790888357733053,
      "tokens_seen": 1230958592
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003166900702106319,
      "loss": 2.703,
      "theoretical_loss": 3.5790713921676978,
      "tokens_seen": 1231024128
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003166800401203611,
      "loss": 2.8052,
      "theoretical_loss": 3.5790539497507154,
      "tokens_seen": 1231089664
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031667001003009027,
      "loss": 2.6364,
      "theoretical_loss": 3.579036508522213,
      "tokens_seen": 1231155200
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003166599799398195,
      "loss": 2.4755,
      "theoretical_loss": 3.5790190684820473,
      "tokens_seen": 1231220736
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031664994984954863,
      "loss": 2.8016,
      "theoretical_loss": 3.579001629630074,
      "tokens_seen": 1231286272
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031663991975927787,
      "loss": 2.4563,
      "theoretical_loss": 3.5789841919661485,
      "tokens_seen": 1231351808
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.000316629889669007,
      "loss": 2.9177,
      "theoretical_loss": 3.5789667554901268,
      "tokens_seen": 1231417344
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031661985957873623,
      "loss": 2.4134,
      "theoretical_loss": 3.5789493202018643,
      "tokens_seen": 1231482880
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003166098294884654,
      "loss": 2.7808,
      "theoretical_loss": 3.5789318861012176,
      "tokens_seen": 1231548416
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003165997993981946,
      "loss": 2.8649,
      "theoretical_loss": 3.5789144531880424,
      "tokens_seen": 1231613952
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003165897693079238,
      "loss": 2.8881,
      "theoretical_loss": 3.5788970214621942,
      "tokens_seen": 1231679488
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.000316579739217653,
      "loss": 2.8825,
      "theoretical_loss": 3.57887959092353,
      "tokens_seen": 1231745024
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031656970912738214,
      "loss": 2.644,
      "theoretical_loss": 3.5788621615719043,
      "tokens_seen": 1231810560
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003165596790371114,
      "loss": 2.6932,
      "theoretical_loss": 3.5788447334071747,
      "tokens_seen": 1231876096
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003165496489468405,
      "loss": 2.7675,
      "theoretical_loss": 3.578827306429196,
      "tokens_seen": 1231941632
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031653961885656974,
      "loss": 2.6962,
      "theoretical_loss": 3.5788098806378246,
      "tokens_seen": 1232007168
    },
    {
      "epoch": 4.01,
      "objective/train/docs_used": 1427266,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6803057193756104,
      "objective/train/theoretical_loss": 3.578792456032917,
      "objective/train/tokens_used": 1252532704,
      "theoretical_loss": 3.578792456032917,
      "tokens_seen": 1232072704
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003165295887662989,
      "loss": 2.6417,
      "theoretical_loss": 3.578792456032917,
      "tokens_seen": 1232072704
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003165195586760281,
      "loss": 2.4745,
      "theoretical_loss": 3.5787750326143293,
      "tokens_seen": 1232138240
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003165095285857573,
      "loss": 2.571,
      "theoretical_loss": 3.578757610381917,
      "tokens_seen": 1232203776
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031649949849548646,
      "loss": 2.7561,
      "theoretical_loss": 3.578740189335537,
      "tokens_seen": 1232269312
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031648946840521564,
      "loss": 2.5433,
      "theoretical_loss": 3.5787227694750445,
      "tokens_seen": 1232334848
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003164794383149449,
      "loss": 3.0216,
      "theoretical_loss": 3.578705350800297,
      "tokens_seen": 1232400384
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.000316469408224674,
      "loss": 2.4573,
      "theoretical_loss": 3.5786879333111496,
      "tokens_seen": 1232465920
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031645937813440324,
      "loss": 2.8524,
      "theoretical_loss": 3.5786705170074598,
      "tokens_seen": 1232531456
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031644934804413237,
      "loss": 2.7487,
      "theoretical_loss": 3.578653101889082,
      "tokens_seen": 1232596992
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003164393179538616,
      "loss": 2.6817,
      "theoretical_loss": 3.5786356879558747,
      "tokens_seen": 1232662528
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003164292878635908,
      "loss": 2.6149,
      "theoretical_loss": 3.5786182752076927,
      "tokens_seen": 1232728064
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031641925777331996,
      "loss": 2.6015,
      "theoretical_loss": 3.578600863644393,
      "tokens_seen": 1232793600
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031640922768304914,
      "loss": 2.5019,
      "theoretical_loss": 3.5785834532658316,
      "tokens_seen": 1232859136
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003163991975927784,
      "loss": 2.8141,
      "theoretical_loss": 3.5785660440718656,
      "tokens_seen": 1232924672
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003163891675025075,
      "loss": 2.5939,
      "theoretical_loss": 3.5785486360623504,
      "tokens_seen": 1232990208
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031637913741223674,
      "loss": 2.7359,
      "theoretical_loss": 3.5785312292371434,
      "tokens_seen": 1233055744
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031636910732196587,
      "loss": 2.5487,
      "theoretical_loss": 3.5785138235961007,
      "tokens_seen": 1233121280
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003163590772316951,
      "loss": 2.5529,
      "theoretical_loss": 3.578496419139079,
      "tokens_seen": 1233186816
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003163490471414243,
      "loss": 3.0039,
      "theoretical_loss": 3.5784790158659345,
      "tokens_seen": 1233252352
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031633901705115347,
      "loss": 2.5339,
      "theoretical_loss": 3.578461613776524,
      "tokens_seen": 1233317888
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031632898696088265,
      "loss": 2.7298,
      "theoretical_loss": 3.5784442128707044,
      "tokens_seen": 1233383424
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031631895687061183,
      "loss": 2.8187,
      "theoretical_loss": 3.578426813148332,
      "tokens_seen": 1233448960
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.000316308926780341,
      "loss": 2.7148,
      "theoretical_loss": 3.578409414609263,
      "tokens_seen": 1233514496
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031629889669007024,
      "loss": 2.6501,
      "theoretical_loss": 3.5783920172533543,
      "tokens_seen": 1233580032
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031628886659979937,
      "loss": 2.741,
      "theoretical_loss": 3.5783746210804637,
      "tokens_seen": 1233645568
    },
    {
      "epoch": 4.01,
      "objective/train/docs_used": 1427860,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9305169582366943,
      "objective/train/theoretical_loss": 3.5783572260904464,
      "objective/train/tokens_used": 1254171104,
      "theoretical_loss": 3.5783572260904464,
      "tokens_seen": 1233711104
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003162788365095286,
      "loss": 2.7442,
      "theoretical_loss": 3.5783572260904464,
      "tokens_seen": 1233711104
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031626880641925773,
      "loss": 2.633,
      "theoretical_loss": 3.57833983228316,
      "tokens_seen": 1233776640
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031625877632898697,
      "loss": 2.5109,
      "theoretical_loss": 3.5783224396584608,
      "tokens_seen": 1233842176
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031624874623871615,
      "loss": 2.4862,
      "theoretical_loss": 3.5783050482162055,
      "tokens_seen": 1233907712
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031623871614844533,
      "loss": 2.6074,
      "theoretical_loss": 3.5782876579562517,
      "tokens_seen": 1233973248
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003162286860581745,
      "loss": 2.5857,
      "theoretical_loss": 3.5782702688784553,
      "tokens_seen": 1234038784
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031621865596790375,
      "loss": 2.5521,
      "theoretical_loss": 3.578252880982674,
      "tokens_seen": 1234104320
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003162086258776329,
      "loss": 2.809,
      "theoretical_loss": 3.578235494268764,
      "tokens_seen": 1234169856
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003161985957873621,
      "loss": 2.7293,
      "theoretical_loss": 3.5782181087365825,
      "tokens_seen": 1234235392
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031618856569709124,
      "loss": 2.6321,
      "theoretical_loss": 3.5782007243859866,
      "tokens_seen": 1234300928
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031617853560682047,
      "loss": 2.3845,
      "theoretical_loss": 3.5781833412168336,
      "tokens_seen": 1234366464
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031616850551654965,
      "loss": 2.6426,
      "theoretical_loss": 3.578165959228979,
      "tokens_seen": 1234432000
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031615847542627883,
      "loss": 2.6782,
      "theoretical_loss": 3.5781485784222813,
      "tokens_seen": 1234497536
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.000316148445336008,
      "loss": 2.6913,
      "theoretical_loss": 3.5781311987965974,
      "tokens_seen": 1234563072
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003161384152457372,
      "loss": 2.6628,
      "theoretical_loss": 3.578113820351784,
      "tokens_seen": 1234628608
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003161283851554664,
      "loss": 2.8064,
      "theoretical_loss": 3.5780964430876976,
      "tokens_seen": 1234694144
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003161183550651956,
      "loss": 2.513,
      "theoretical_loss": 3.5780790670041966,
      "tokens_seen": 1234759680
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031610832497492474,
      "loss": 2.624,
      "theoretical_loss": 3.578061692101137,
      "tokens_seen": 1234825216
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.000316098294884654,
      "loss": 2.6117,
      "theoretical_loss": 3.5780443183783763,
      "tokens_seen": 1234890752
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003160882647943831,
      "loss": 2.6111,
      "theoretical_loss": 3.5780269458357723,
      "tokens_seen": 1234956288
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031607823470411234,
      "loss": 2.5321,
      "theoretical_loss": 3.5780095744731817,
      "tokens_seen": 1235021824
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003160682046138415,
      "loss": 2.5517,
      "theoretical_loss": 3.5779922042904615,
      "tokens_seen": 1235087360
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003160581745235707,
      "loss": 2.3564,
      "theoretical_loss": 3.5779748352874696,
      "tokens_seen": 1235152896
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031604814443329994,
      "loss": 2.561,
      "theoretical_loss": 3.577957467464062,
      "tokens_seen": 1235218432
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003160381143430291,
      "loss": 2.8883,
      "theoretical_loss": 3.577940100820098,
      "tokens_seen": 1235283968
    },
    {
      "epoch": 4.01,
      "objective/train/docs_used": 1429255,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.2701942920684814,
      "objective/train/theoretical_loss": 3.5779227353554335,
      "objective/train/tokens_used": 1255809504,
      "theoretical_loss": 3.5779227353554335,
      "tokens_seen": 1235349504
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003160280842527583,
      "loss": 2.5859,
      "theoretical_loss": 3.5779227353554335,
      "tokens_seen": 1235349504
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003160180541624875,
      "loss": 2.9854,
      "theoretical_loss": 3.5779053710699262,
      "tokens_seen": 1235415040
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031600802407221666,
      "loss": 2.5211,
      "theoretical_loss": 3.577888007963433,
      "tokens_seen": 1235480576
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031599799398194584,
      "loss": 2.6688,
      "theoretical_loss": 3.5778706460358127,
      "tokens_seen": 1235546112
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003159879638916751,
      "loss": 2.6484,
      "theoretical_loss": 3.5778532852869214,
      "tokens_seen": 1235611648
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003159779338014042,
      "loss": 2.6478,
      "theoretical_loss": 3.5778359257166175,
      "tokens_seen": 1235677184
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031596790371113344,
      "loss": 2.5795,
      "theoretical_loss": 3.5778185673247576,
      "tokens_seen": 1235742720
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031595787362086257,
      "loss": 2.5562,
      "theoretical_loss": 3.5778012101111996,
      "tokens_seen": 1235808256
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003159478435305918,
      "loss": 2.8139,
      "theoretical_loss": 3.577783854075801,
      "tokens_seen": 1235873792
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.000315937813440321,
      "loss": 2.7729,
      "theoretical_loss": 3.5777664992184195,
      "tokens_seen": 1235939328
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031592778335005016,
      "loss": 2.8157,
      "theoretical_loss": 3.5777491455389123,
      "tokens_seen": 1236004864
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031591775325977934,
      "loss": 2.5978,
      "theoretical_loss": 3.577731793037138,
      "tokens_seen": 1236070400
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003159077231695086,
      "loss": 2.8192,
      "theoretical_loss": 3.577714441712953,
      "tokens_seen": 1236135936
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003158976930792377,
      "loss": 2.5671,
      "theoretical_loss": 3.577697091566216,
      "tokens_seen": 1236201472
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031588766298896694,
      "loss": 2.6757,
      "theoretical_loss": 3.577679742596784,
      "tokens_seen": 1236267008
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031587763289869607,
      "loss": 2.6728,
      "theoretical_loss": 3.5776623948045145,
      "tokens_seen": 1236332544
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003158676028084253,
      "loss": 2.5845,
      "theoretical_loss": 3.5776450481892663,
      "tokens_seen": 1236398080
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.0003158575727181545,
      "loss": 2.5337,
      "theoretical_loss": 3.5776277027508963,
      "tokens_seen": 1236463616
    },
    {
      "epoch": 4.01,
      "learning_rate": 0.00031584754262788367,
      "loss": 2.8174,
      "theoretical_loss": 3.5776103584892622,
      "tokens_seen": 1236529152
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031583751253761285,
      "loss": 2.6758,
      "theoretical_loss": 3.577593015404222,
      "tokens_seen": 1236594688
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031582748244734203,
      "loss": 2.5004,
      "theoretical_loss": 3.577575673495634,
      "tokens_seen": 1236660224
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003158174523570712,
      "loss": 2.5351,
      "theoretical_loss": 3.5775583327633553,
      "tokens_seen": 1236725760
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031580742226680044,
      "loss": 2.7012,
      "theoretical_loss": 3.577540993207245,
      "tokens_seen": 1236791296
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031579739217652957,
      "loss": 2.4612,
      "theoretical_loss": 3.5775236548271594,
      "tokens_seen": 1236856832
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003157873620862588,
      "loss": 2.6514,
      "theoretical_loss": 3.5775063176229573,
      "tokens_seen": 1236922368
    },
    {
      "epoch": 4.02,
      "objective/train/docs_used": 1430103,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.250220537185669,
      "objective/train/theoretical_loss": 3.577488981594497,
      "objective/train/tokens_used": 1257447904,
      "theoretical_loss": 3.577488981594497,
      "tokens_seen": 1236987904
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031577733199598793,
      "loss": 2.5234,
      "theoretical_loss": 3.577488981594497,
      "tokens_seen": 1236987904
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031576730190571717,
      "loss": 2.7337,
      "theoretical_loss": 3.5774716467416354,
      "tokens_seen": 1237053440
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031575727181544635,
      "loss": 2.7496,
      "theoretical_loss": 3.5774543130642313,
      "tokens_seen": 1237118976
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031574724172517553,
      "loss": 2.5515,
      "theoretical_loss": 3.577436980562143,
      "tokens_seen": 1237184512
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003157372116349047,
      "loss": 2.8153,
      "theoretical_loss": 3.5774196492352277,
      "tokens_seen": 1237250048
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031572718154463395,
      "loss": 2.5813,
      "theoretical_loss": 3.577402319083344,
      "tokens_seen": 1237315584
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003157171514543631,
      "loss": 2.5498,
      "theoretical_loss": 3.57738499010635,
      "tokens_seen": 1237381120
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003157071213640923,
      "loss": 2.7301,
      "theoretical_loss": 3.5773676623041037,
      "tokens_seen": 1237446656
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031569709127382144,
      "loss": 2.7963,
      "theoretical_loss": 3.5773503356764635,
      "tokens_seen": 1237512192
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031568706118355067,
      "loss": 2.4556,
      "theoretical_loss": 3.577333010223287,
      "tokens_seen": 1237577728
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031567703109327985,
      "loss": 2.4602,
      "theoretical_loss": 3.5773156859444333,
      "tokens_seen": 1237643264
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031566700100300903,
      "loss": 2.5788,
      "theoretical_loss": 3.5772983628397594,
      "tokens_seen": 1237708800
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003156569709127382,
      "loss": 2.7927,
      "theoretical_loss": 3.5772810409091247,
      "tokens_seen": 1237774336
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003156469408224674,
      "loss": 2.7631,
      "theoretical_loss": 3.577263720152387,
      "tokens_seen": 1237839872
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003156369107321966,
      "loss": 2.6811,
      "theoretical_loss": 3.577246400569405,
      "tokens_seen": 1237905408
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003156268806419258,
      "loss": 2.5648,
      "theoretical_loss": 3.5772290821600357,
      "tokens_seen": 1237970944
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031561685055165494,
      "loss": 2.5189,
      "theoretical_loss": 3.5772117649241393,
      "tokens_seen": 1238036480
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003156068204613842,
      "loss": 2.9144,
      "theoretical_loss": 3.577194448861573,
      "tokens_seen": 1238102016
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003155967903711133,
      "loss": 2.7917,
      "theoretical_loss": 3.5771771339721954,
      "tokens_seen": 1238167552
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031558676028084254,
      "loss": 2.8326,
      "theoretical_loss": 3.577159820255865,
      "tokens_seen": 1238233088
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003155767301905717,
      "loss": 2.6731,
      "theoretical_loss": 3.5771425077124404,
      "tokens_seen": 1238298624
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003155667001003009,
      "loss": 2.6652,
      "theoretical_loss": 3.57712519634178,
      "tokens_seen": 1238364160
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003155566700100301,
      "loss": 2.5107,
      "theoretical_loss": 3.5771078861437418,
      "tokens_seen": 1238429696
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003155466399197593,
      "loss": 2.5704,
      "theoretical_loss": 3.577090577118185,
      "tokens_seen": 1238495232
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031553660982948844,
      "loss": 2.736,
      "theoretical_loss": 3.577073269264968,
      "tokens_seen": 1238560768
    },
    {
      "epoch": 4.02,
      "objective/train/docs_used": 1431596,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.694431781768799,
      "objective/train/theoretical_loss": 3.577055962583949,
      "objective/train/tokens_used": 1259086304,
      "theoretical_loss": 3.577055962583949,
      "tokens_seen": 1238626304
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003155265797392177,
      "loss": 2.7258,
      "theoretical_loss": 3.577055962583949,
      "tokens_seen": 1238626304
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003155165496489468,
      "loss": 2.8068,
      "theoretical_loss": 3.5770386570749872,
      "tokens_seen": 1238691840
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031550651955867604,
      "loss": 2.5443,
      "theoretical_loss": 3.5770213527379404,
      "tokens_seen": 1238757376
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003154964894684052,
      "loss": 2.6994,
      "theoretical_loss": 3.5770040495726687,
      "tokens_seen": 1238822912
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003154864593781344,
      "loss": 2.7348,
      "theoretical_loss": 3.576986747579029,
      "tokens_seen": 1238888448
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003154764292878636,
      "loss": 2.727,
      "theoretical_loss": 3.5769694467568813,
      "tokens_seen": 1238953984
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031546639919759277,
      "loss": 2.8486,
      "theoretical_loss": 3.5769521471060832,
      "tokens_seen": 1239019520
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031545636910732195,
      "loss": 2.7562,
      "theoretical_loss": 3.5769348486264945,
      "tokens_seen": 1239085056
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003154463390170512,
      "loss": 2.8638,
      "theoretical_loss": 3.5769175513179734,
      "tokens_seen": 1239150592
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003154363089267803,
      "loss": 2.577,
      "theoretical_loss": 3.5769002551803792,
      "tokens_seen": 1239216128
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031542627883650954,
      "loss": 2.6711,
      "theoretical_loss": 3.5768829602135694,
      "tokens_seen": 1239281664
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003154162487462387,
      "loss": 2.6202,
      "theoretical_loss": 3.5768656664174046,
      "tokens_seen": 1239347200
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003154062186559679,
      "loss": 2.7357,
      "theoretical_loss": 3.5768483737917425,
      "tokens_seen": 1239412736
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003153961885656971,
      "loss": 2.7097,
      "theoretical_loss": 3.576831082336443,
      "tokens_seen": 1239478272
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031538615847542627,
      "loss": 2.6539,
      "theoretical_loss": 3.5768137920513636,
      "tokens_seen": 1239543808
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031537612838515545,
      "loss": 2.5496,
      "theoretical_loss": 3.5767965029363644,
      "tokens_seen": 1239609344
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003153660982948847,
      "loss": 2.8014,
      "theoretical_loss": 3.5767792149913036,
      "tokens_seen": 1239674880
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003153560682046138,
      "loss": 2.5942,
      "theoretical_loss": 3.5767619282160403,
      "tokens_seen": 1239740416
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031534603811434305,
      "loss": 2.5592,
      "theoretical_loss": 3.576744642610435,
      "tokens_seen": 1239805952
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003153360080240722,
      "loss": 2.4833,
      "theoretical_loss": 3.5767273581743444,
      "tokens_seen": 1239871488
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003153259779338014,
      "loss": 2.8076,
      "theoretical_loss": 3.576710074907629,
      "tokens_seen": 1239937024
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003153159478435306,
      "loss": 2.751,
      "theoretical_loss": 3.576692792810147,
      "tokens_seen": 1240002560
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031530591775325977,
      "loss": 2.8131,
      "theoretical_loss": 3.576675511881759,
      "tokens_seen": 1240068096
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.000315295887662989,
      "loss": 2.7056,
      "theoretical_loss": 3.5766582321223224,
      "tokens_seen": 1240133632
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031528585757271813,
      "loss": 2.6935,
      "theoretical_loss": 3.5766409535316974,
      "tokens_seen": 1240199168
    },
    {
      "epoch": 4.02,
      "objective/train/docs_used": 1432295,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.533489227294922,
      "objective/train/theoretical_loss": 3.576623676109743,
      "objective/train/tokens_used": 1260724704,
      "theoretical_loss": 3.576623676109743,
      "tokens_seen": 1240264704
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031527582748244737,
      "loss": 2.8536,
      "theoretical_loss": 3.576623676109743,
      "tokens_seen": 1240264704
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031526579739217655,
      "loss": 2.5282,
      "theoretical_loss": 3.5766063998563187,
      "tokens_seen": 1240330240
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031525576730190573,
      "loss": 2.5231,
      "theoretical_loss": 3.5765891247712824,
      "tokens_seen": 1240395776
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003152457372116349,
      "loss": 2.6814,
      "theoretical_loss": 3.5765718508544952,
      "tokens_seen": 1240461312
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031523570712136415,
      "loss": 2.5898,
      "theoretical_loss": 3.5765545781058155,
      "tokens_seen": 1240526848
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003152256770310933,
      "loss": 2.8252,
      "theoretical_loss": 3.5765373065251023,
      "tokens_seen": 1240592384
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003152156469408225,
      "loss": 2.6618,
      "theoretical_loss": 3.576520036112215,
      "tokens_seen": 1240657920
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031520561685055164,
      "loss": 2.8285,
      "theoretical_loss": 3.5765027668670135,
      "tokens_seen": 1240723456
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031519558676028087,
      "loss": 2.5839,
      "theoretical_loss": 3.576485498789357,
      "tokens_seen": 1240788992
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031518555667001005,
      "loss": 2.6935,
      "theoretical_loss": 3.5764682318791046,
      "tokens_seen": 1240854528
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031517552657973924,
      "loss": 2.8678,
      "theoretical_loss": 3.576450966136116,
      "tokens_seen": 1240920064
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003151654964894684,
      "loss": 2.66,
      "theoretical_loss": 3.5764337015602505,
      "tokens_seen": 1240985600
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003151554663991976,
      "loss": 2.5715,
      "theoretical_loss": 3.5764164381513677,
      "tokens_seen": 1241051136
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003151454363089268,
      "loss": 2.6386,
      "theoretical_loss": 3.5763991759093265,
      "tokens_seen": 1241116672
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.000315135406218656,
      "loss": 2.6935,
      "theoretical_loss": 3.5763819148339877,
      "tokens_seen": 1241182208
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031512537612838514,
      "loss": 2.6195,
      "theoretical_loss": 3.57636465492521,
      "tokens_seen": 1241247744
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003151153460381144,
      "loss": 2.6906,
      "theoretical_loss": 3.576347396182853,
      "tokens_seen": 1241313280
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003151053159478435,
      "loss": 2.5668,
      "theoretical_loss": 3.5763301386067763,
      "tokens_seen": 1241378816
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031509528585757274,
      "loss": 2.9009,
      "theoretical_loss": 3.5763128821968393,
      "tokens_seen": 1241444352
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003150852557673019,
      "loss": 2.7263,
      "theoretical_loss": 3.576295626952902,
      "tokens_seen": 1241509888
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003150752256770311,
      "loss": 2.7097,
      "theoretical_loss": 3.5762783728748246,
      "tokens_seen": 1241575424
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003150651955867603,
      "loss": 2.636,
      "theoretical_loss": 3.576261119962466,
      "tokens_seen": 1241640960
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003150551654964895,
      "loss": 2.792,
      "theoretical_loss": 3.576243868215686,
      "tokens_seen": 1241706496
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031504513540621864,
      "loss": 2.7841,
      "theoretical_loss": 3.5762266176343447,
      "tokens_seen": 1241772032
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003150351053159479,
      "loss": 2.8194,
      "theoretical_loss": 3.576209368218301,
      "tokens_seen": 1241837568
    },
    {
      "epoch": 4.02,
      "objective/train/docs_used": 1433476,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.985485315322876,
      "objective/train/theoretical_loss": 3.576192119967416,
      "objective/train/tokens_used": 1262363104,
      "theoretical_loss": 3.576192119967416,
      "tokens_seen": 1241903104
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.000315025075225677,
      "loss": 2.8603,
      "theoretical_loss": 3.576192119967416,
      "tokens_seen": 1241903104
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031501504513540624,
      "loss": 2.7702,
      "theoretical_loss": 3.5761748728815483,
      "tokens_seen": 1241968640
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003150050150451354,
      "loss": 2.7794,
      "theoretical_loss": 3.576157626960559,
      "tokens_seen": 1242034176
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003149949849548646,
      "loss": 2.6736,
      "theoretical_loss": 3.576140382204307,
      "tokens_seen": 1242099712
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003149849548645938,
      "loss": 2.6884,
      "theoretical_loss": 3.576123138612652,
      "tokens_seen": 1242165248
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031497492477432297,
      "loss": 2.4335,
      "theoretical_loss": 3.576105896185455,
      "tokens_seen": 1242230784
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031496489468405215,
      "loss": 2.4949,
      "theoretical_loss": 3.5760886549225748,
      "tokens_seen": 1242296320
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003149548645937814,
      "loss": 2.5471,
      "theoretical_loss": 3.576071414823872,
      "tokens_seen": 1242361856
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003149448345035105,
      "loss": 2.7882,
      "theoretical_loss": 3.576054175889207,
      "tokens_seen": 1242427392
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031493480441323974,
      "loss": 2.6353,
      "theoretical_loss": 3.5760369381184383,
      "tokens_seen": 1242492928
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003149247743229689,
      "loss": 2.764,
      "theoretical_loss": 3.5760197015114277,
      "tokens_seen": 1242558464
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003149147442326981,
      "loss": 2.7535,
      "theoretical_loss": 3.576002466068034,
      "tokens_seen": 1242624000
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003149047141424273,
      "loss": 2.5901,
      "theoretical_loss": 3.575985231788118,
      "tokens_seen": 1242689536
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031489468405215647,
      "loss": 2.6292,
      "theoretical_loss": 3.575967998671539,
      "tokens_seen": 1242755072
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031488465396188565,
      "loss": 2.3893,
      "theoretical_loss": 3.575950766718158,
      "tokens_seen": 1242820608
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003148746238716149,
      "loss": 2.6584,
      "theoretical_loss": 3.5759335359278355,
      "tokens_seen": 1242886144
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.000314864593781344,
      "loss": 2.6964,
      "theoretical_loss": 3.5759163063004302,
      "tokens_seen": 1242951680
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031485456369107325,
      "loss": 2.7812,
      "theoretical_loss": 3.5758990778358033,
      "tokens_seen": 1243017216
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003148445336008024,
      "loss": 2.7795,
      "theoretical_loss": 3.575881850533815,
      "tokens_seen": 1243082752
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003148345035105316,
      "loss": 2.6807,
      "theoretical_loss": 3.5758646243943253,
      "tokens_seen": 1243148288
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003148244734202608,
      "loss": 2.715,
      "theoretical_loss": 3.5758473994171944,
      "tokens_seen": 1243213824
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031481444332998997,
      "loss": 2.5597,
      "theoretical_loss": 3.575830175602283,
      "tokens_seen": 1243279360
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031480441323971915,
      "loss": 2.8623,
      "theoretical_loss": 3.575812952949451,
      "tokens_seen": 1243344896
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031479438314944833,
      "loss": 2.3986,
      "theoretical_loss": 3.575795731458559,
      "tokens_seen": 1243410432
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003147843530591775,
      "loss": 2.5205,
      "theoretical_loss": 3.5757785111294673,
      "tokens_seen": 1243475968
    },
    {
      "epoch": 4.02,
      "objective/train/docs_used": 1434244,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7484705448150635,
      "objective/train/theoretical_loss": 3.5757612919620367,
      "objective/train/tokens_used": 1264001504,
      "theoretical_loss": 3.5757612919620367,
      "tokens_seen": 1243541504
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031477432296890675,
      "loss": 2.7885,
      "theoretical_loss": 3.5757612919620367,
      "tokens_seen": 1243541504
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003147642928786359,
      "loss": 2.8115,
      "theoretical_loss": 3.575744073956127,
      "tokens_seen": 1243607040
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003147542627883651,
      "loss": 2.6593,
      "theoretical_loss": 3.575726857111598,
      "tokens_seen": 1243672576
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003147442326980943,
      "loss": 2.73,
      "theoretical_loss": 3.575709641428312,
      "tokens_seen": 1243738112
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003147342026078235,
      "loss": 2.6826,
      "theoretical_loss": 3.5756924269061288,
      "tokens_seen": 1243803648
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031472417251755266,
      "loss": 2.8426,
      "theoretical_loss": 3.575675213544908,
      "tokens_seen": 1243869184
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031471414242728184,
      "loss": 2.8811,
      "theoretical_loss": 3.575658001344511,
      "tokens_seen": 1243934720
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.000314704112337011,
      "loss": 2.726,
      "theoretical_loss": 3.5756407903047984,
      "tokens_seen": 1244000256
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031469408224674025,
      "loss": 2.6208,
      "theoretical_loss": 3.5756235804256304,
      "tokens_seen": 1244065792
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003146840521564694,
      "loss": 2.5834,
      "theoretical_loss": 3.5756063717068676,
      "tokens_seen": 1244131328
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003146740220661986,
      "loss": 2.7418,
      "theoretical_loss": 3.5755891641483712,
      "tokens_seen": 1244196864
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031466399197592774,
      "loss": 2.7675,
      "theoretical_loss": 3.575571957750001,
      "tokens_seen": 1244262400
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.000314653961885657,
      "loss": 2.6723,
      "theoretical_loss": 3.5755547525116187,
      "tokens_seen": 1244327936
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031464393179538616,
      "loss": 2.5414,
      "theoretical_loss": 3.575537548433084,
      "tokens_seen": 1244393472
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031463390170511534,
      "loss": 2.5029,
      "theoretical_loss": 3.575520345514258,
      "tokens_seen": 1244459008
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003146238716148445,
      "loss": 2.8787,
      "theoretical_loss": 3.575503143755002,
      "tokens_seen": 1244524544
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003146138415245737,
      "loss": 2.6157,
      "theoretical_loss": 3.575485943155176,
      "tokens_seen": 1244590080
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003146038114343029,
      "loss": 2.671,
      "theoretical_loss": 3.5754687437146417,
      "tokens_seen": 1244655616
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003145937813440321,
      "loss": 2.527,
      "theoretical_loss": 3.5754515454332587,
      "tokens_seen": 1244721152
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031458375125376125,
      "loss": 2.3568,
      "theoretical_loss": 3.575434348310889,
      "tokens_seen": 1244786688
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003145737211634905,
      "loss": 2.4167,
      "theoretical_loss": 3.575417152347393,
      "tokens_seen": 1244852224
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003145636910732197,
      "loss": 2.5593,
      "theoretical_loss": 3.5753999575426314,
      "tokens_seen": 1244917760
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031455366098294884,
      "loss": 2.78,
      "theoretical_loss": 3.5753827638964655,
      "tokens_seen": 1244983296
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003145436308926781,
      "loss": 2.7571,
      "theoretical_loss": 3.575365571408756,
      "tokens_seen": 1245048832
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003145336008024072,
      "loss": 2.4951,
      "theoretical_loss": 3.5753483800793644,
      "tokens_seen": 1245114368
    },
    {
      "epoch": 4.02,
      "objective/train/docs_used": 1435669,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.428680419921875,
      "objective/train/theoretical_loss": 3.575331189908151,
      "objective/train/tokens_used": 1265639904,
      "theoretical_loss": 3.575331189908151,
      "tokens_seen": 1245179904
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031452357071213644,
      "loss": 2.7965,
      "theoretical_loss": 3.575331189908151,
      "tokens_seen": 1245179904
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003145135406218656,
      "loss": 2.6784,
      "theoretical_loss": 3.575314000894976,
      "tokens_seen": 1245245440
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003145035105315948,
      "loss": 2.4244,
      "theoretical_loss": 3.575296813039703,
      "tokens_seen": 1245310976
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.000314493480441324,
      "loss": 2.7376,
      "theoretical_loss": 3.575279626342191,
      "tokens_seen": 1245376512
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031448345035105317,
      "loss": 2.7399,
      "theoretical_loss": 3.5752624408023017,
      "tokens_seen": 1245442048
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031447342026078235,
      "loss": 2.7588,
      "theoretical_loss": 3.5752452564198967,
      "tokens_seen": 1245507584
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003144633901705116,
      "loss": 2.6889,
      "theoretical_loss": 3.575228073194836,
      "tokens_seen": 1245573120
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003144533600802407,
      "loss": 2.6662,
      "theoretical_loss": 3.575210891126982,
      "tokens_seen": 1245638656
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031444332998996994,
      "loss": 2.5576,
      "theoretical_loss": 3.575193710216195,
      "tokens_seen": 1245704192
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003144332998996991,
      "loss": 2.7878,
      "theoretical_loss": 3.575176530462337,
      "tokens_seen": 1245769728
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003144232698094283,
      "loss": 2.7706,
      "theoretical_loss": 3.5751593518652682,
      "tokens_seen": 1245835264
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003144132397191575,
      "loss": 2.7004,
      "theoretical_loss": 3.575142174424851,
      "tokens_seen": 1245900800
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031440320962888667,
      "loss": 2.8504,
      "theoretical_loss": 3.5751249981409456,
      "tokens_seen": 1245966336
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031439317953861585,
      "loss": 2.6992,
      "theoretical_loss": 3.575107823013414,
      "tokens_seen": 1246031872
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003143831494483451,
      "loss": 2.6023,
      "theoretical_loss": 3.5750906490421173,
      "tokens_seen": 1246097408
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003143731193580742,
      "loss": 2.5395,
      "theoretical_loss": 3.575073476226917,
      "tokens_seen": 1246162944
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031436308926780345,
      "loss": 2.5867,
      "theoretical_loss": 3.575056304567674,
      "tokens_seen": 1246228480
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003143530591775326,
      "loss": 2.572,
      "theoretical_loss": 3.5750391340642507,
      "tokens_seen": 1246294016
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003143430290872618,
      "loss": 2.6351,
      "theoretical_loss": 3.575021964716508,
      "tokens_seen": 1246359552
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.000314332998996991,
      "loss": 2.4288,
      "theoretical_loss": 3.5750047965243072,
      "tokens_seen": 1246425088
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031432296890672017,
      "loss": 2.5788,
      "theoretical_loss": 3.57498762948751,
      "tokens_seen": 1246490624
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031431293881644935,
      "loss": 2.7888,
      "theoretical_loss": 3.574970463605977,
      "tokens_seen": 1246556160
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031430290872617853,
      "loss": 2.682,
      "theoretical_loss": 3.5749532988795716,
      "tokens_seen": 1246621696
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003142928786359077,
      "loss": 2.5778,
      "theoretical_loss": 3.5749361353081537,
      "tokens_seen": 1246687232
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031428284854563695,
      "loss": 2.5332,
      "theoretical_loss": 3.5749189728915853,
      "tokens_seen": 1246752768
    },
    {
      "epoch": 4.02,
      "objective/train/docs_used": 1436465,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.797041893005371,
      "objective/train/theoretical_loss": 3.5749018116297284,
      "objective/train/tokens_used": 1267278304,
      "theoretical_loss": 3.5749018116297284,
      "tokens_seen": 1246818304
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003142728184553661,
      "loss": 2.7708,
      "theoretical_loss": 3.5749018116297284,
      "tokens_seen": 1246818304
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003142627883650953,
      "loss": 2.3856,
      "theoretical_loss": 3.574884651522444,
      "tokens_seen": 1246883840
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003142527582748245,
      "loss": 2.7342,
      "theoretical_loss": 3.5748674925695942,
      "tokens_seen": 1246949376
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003142427281845537,
      "loss": 2.5475,
      "theoretical_loss": 3.5748503347710407,
      "tokens_seen": 1247014912
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031423269809428286,
      "loss": 2.4742,
      "theoretical_loss": 3.574833178126645,
      "tokens_seen": 1247080448
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031422266800401204,
      "loss": 2.6937,
      "theoretical_loss": 3.574816022636269,
      "tokens_seen": 1247145984
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003142126379137412,
      "loss": 2.6389,
      "theoretical_loss": 3.5747988682997747,
      "tokens_seen": 1247211520
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031420260782347045,
      "loss": 2.6282,
      "theoretical_loss": 3.5747817151170227,
      "tokens_seen": 1247277056
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003141925777331996,
      "loss": 2.6403,
      "theoretical_loss": 3.574764563087876,
      "tokens_seen": 1247342592
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003141825476429288,
      "loss": 2.8371,
      "theoretical_loss": 3.5747474122121954,
      "tokens_seen": 1247408128
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031417251755265794,
      "loss": 2.5975,
      "theoretical_loss": 3.574730262489844,
      "tokens_seen": 1247473664
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003141624874623872,
      "loss": 2.5355,
      "theoretical_loss": 3.5747131139206827,
      "tokens_seen": 1247539200
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031415245737211636,
      "loss": 2.6023,
      "theoretical_loss": 3.5746959665045734,
      "tokens_seen": 1247604736
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031414242728184554,
      "loss": 2.7003,
      "theoretical_loss": 3.5746788202413784,
      "tokens_seen": 1247670272
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003141323971915747,
      "loss": 2.3663,
      "theoretical_loss": 3.57466167513096,
      "tokens_seen": 1247735808
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003141223671013039,
      "loss": 2.7268,
      "theoretical_loss": 3.5746445311731794,
      "tokens_seen": 1247801344
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003141123370110331,
      "loss": 2.6805,
      "theoretical_loss": 3.5746273883678983,
      "tokens_seen": 1247866880
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003141023069207623,
      "loss": 2.7947,
      "theoretical_loss": 3.57461024671498,
      "tokens_seen": 1247932416
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031409227683049145,
      "loss": 2.8079,
      "theoretical_loss": 3.574593106214285,
      "tokens_seen": 1247997952
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003140822467402207,
      "loss": 2.6496,
      "theoretical_loss": 3.5745759668656767,
      "tokens_seen": 1248063488
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031407221664994986,
      "loss": 2.6264,
      "theoretical_loss": 3.574558828669016,
      "tokens_seen": 1248129024
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031406218655967904,
      "loss": 2.2906,
      "theoretical_loss": 3.574541691624166,
      "tokens_seen": 1248194560
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003140521564694082,
      "loss": 2.902,
      "theoretical_loss": 3.574524555730988,
      "tokens_seen": 1248260096
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003140421263791374,
      "loss": 2.7177,
      "theoretical_loss": 3.5745074209893453,
      "tokens_seen": 1248325632
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003140320962888666,
      "loss": 2.7631,
      "theoretical_loss": 3.5744902873990982,
      "tokens_seen": 1248391168
    },
    {
      "epoch": 4.02,
      "objective/train/docs_used": 1437026,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7030203342437744,
      "objective/train/theoretical_loss": 3.5744731549601108,
      "objective/train/tokens_used": 1268916704,
      "theoretical_loss": 3.5744731549601108,
      "tokens_seen": 1248456704
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003140220661985958,
      "loss": 2.6616,
      "theoretical_loss": 3.5744731549601108,
      "tokens_seen": 1248456704
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031401203610832495,
      "loss": 2.6589,
      "theoretical_loss": 3.5744560236722442,
      "tokens_seen": 1248522240
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003140020060180542,
      "loss": 2.6114,
      "theoretical_loss": 3.574438893535361,
      "tokens_seen": 1248587776
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003139919759277833,
      "loss": 2.7657,
      "theoretical_loss": 3.5744217645493235,
      "tokens_seen": 1248653312
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031398194583751255,
      "loss": 2.6823,
      "theoretical_loss": 3.574404636713994,
      "tokens_seen": 1248718848
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031397191574724173,
      "loss": 2.5717,
      "theoretical_loss": 3.5743875100292346,
      "tokens_seen": 1248784384
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003139618856569709,
      "loss": 2.6765,
      "theoretical_loss": 3.574370384494908,
      "tokens_seen": 1248849920
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003139518555667001,
      "loss": 2.6439,
      "theoretical_loss": 3.574353260110876,
      "tokens_seen": 1248915456
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003139418254764293,
      "loss": 2.5065,
      "theoretical_loss": 3.574336136877001,
      "tokens_seen": 1248980992
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031393179538615845,
      "loss": 2.713,
      "theoretical_loss": 3.5743190147931463,
      "tokens_seen": 1249046528
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003139217652958877,
      "loss": 2.5401,
      "theoretical_loss": 3.5743018938591735,
      "tokens_seen": 1249112064
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003139117352056168,
      "loss": 2.6175,
      "theoretical_loss": 3.574284774074945,
      "tokens_seen": 1249177600
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031390170511534605,
      "loss": 2.6046,
      "theoretical_loss": 3.574267655440323,
      "tokens_seen": 1249243136
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031389167502507523,
      "loss": 2.6326,
      "theoretical_loss": 3.574250537955172,
      "tokens_seen": 1249308672
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003138816449348044,
      "loss": 2.4136,
      "theoretical_loss": 3.5742334216193523,
      "tokens_seen": 1249374208
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003138716148445336,
      "loss": 2.4814,
      "theoretical_loss": 3.574216306432727,
      "tokens_seen": 1249439744
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003138615847542628,
      "loss": 2.6682,
      "theoretical_loss": 3.5741991923951595,
      "tokens_seen": 1249505280
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031385155466399196,
      "loss": 2.7518,
      "theoretical_loss": 3.574182079506511,
      "tokens_seen": 1249570816
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003138415245737212,
      "loss": 2.5976,
      "theoretical_loss": 3.5741649677666456,
      "tokens_seen": 1249636352
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003138314944834503,
      "loss": 2.7658,
      "theoretical_loss": 3.574147857175425,
      "tokens_seen": 1249701888
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031382146439317955,
      "loss": 2.8964,
      "theoretical_loss": 3.574130747732712,
      "tokens_seen": 1249767424
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031381143430290873,
      "loss": 2.6613,
      "theoretical_loss": 3.5741136394383695,
      "tokens_seen": 1249832960
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003138014042126379,
      "loss": 2.5456,
      "theoretical_loss": 3.57409653229226,
      "tokens_seen": 1249898496
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031379137412236715,
      "loss": 2.7456,
      "theoretical_loss": 3.5740794262942464,
      "tokens_seen": 1249964032
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003137813440320963,
      "loss": 2.5863,
      "theoretical_loss": 3.574062321444192,
      "tokens_seen": 1250029568
    },
    {
      "epoch": 4.02,
      "objective/train/docs_used": 1438333,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5628535747528076,
      "objective/train/theoretical_loss": 3.5740452177419577,
      "objective/train/tokens_used": 1270555104,
      "theoretical_loss": 3.5740452177419577,
      "tokens_seen": 1250095104
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003137713139418255,
      "loss": 2.7452,
      "theoretical_loss": 3.5740452177419577,
      "tokens_seen": 1250095104
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003137612838515547,
      "loss": 2.5718,
      "theoretical_loss": 3.5740281151874087,
      "tokens_seen": 1250160640
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003137512537612839,
      "loss": 2.5954,
      "theoretical_loss": 3.5740110137804066,
      "tokens_seen": 1250226176
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031374122367101306,
      "loss": 2.577,
      "theoretical_loss": 3.5739939135208143,
      "tokens_seen": 1250291712
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031373119358074224,
      "loss": 2.5118,
      "theoretical_loss": 3.5739768144084945,
      "tokens_seen": 1250357248
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003137211634904714,
      "loss": 2.7736,
      "theoretical_loss": 3.5739597164433103,
      "tokens_seen": 1250422784
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031371113340020065,
      "loss": 2.8973,
      "theoretical_loss": 3.5739426196251247,
      "tokens_seen": 1250488320
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003137011033099298,
      "loss": 2.9025,
      "theoretical_loss": 3.5739255239538013,
      "tokens_seen": 1250553856
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.000313691073219659,
      "loss": 2.8072,
      "theoretical_loss": 3.573908429429202,
      "tokens_seen": 1250619392
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031368104312938814,
      "loss": 2.5561,
      "theoretical_loss": 3.5738913360511897,
      "tokens_seen": 1250684928
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003136710130391174,
      "loss": 2.9685,
      "theoretical_loss": 3.5738742438196285,
      "tokens_seen": 1250750464
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031366098294884656,
      "loss": 2.7795,
      "theoretical_loss": 3.5738571527343805,
      "tokens_seen": 1250816000
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031365095285857574,
      "loss": 2.7314,
      "theoretical_loss": 3.573840062795309,
      "tokens_seen": 1250881536
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003136409227683049,
      "loss": 2.8282,
      "theoretical_loss": 3.5738229740022778,
      "tokens_seen": 1250947072
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003136308926780341,
      "loss": 2.6157,
      "theoretical_loss": 3.5738058863551494,
      "tokens_seen": 1251012608
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003136208625877633,
      "loss": 2.8974,
      "theoretical_loss": 3.5737887998537867,
      "tokens_seen": 1251078144
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003136108324974925,
      "loss": 2.7745,
      "theoretical_loss": 3.5737717144980534,
      "tokens_seen": 1251143680
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031360080240722165,
      "loss": 2.7687,
      "theoretical_loss": 3.573754630287812,
      "tokens_seen": 1251209216
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003135907723169509,
      "loss": 2.6226,
      "theoretical_loss": 3.573737547222926,
      "tokens_seen": 1251274752
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031358074222668006,
      "loss": 2.568,
      "theoretical_loss": 3.5737204653032597,
      "tokens_seen": 1251340288
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031357071213640924,
      "loss": 2.5222,
      "theoretical_loss": 3.5737033845286748,
      "tokens_seen": 1251405824
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003135606820461384,
      "loss": 2.7934,
      "theoretical_loss": 3.573686304899035,
      "tokens_seen": 1251471360
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003135506519558676,
      "loss": 2.5712,
      "theoretical_loss": 3.5736692264142036,
      "tokens_seen": 1251536896
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003135406218655968,
      "loss": 2.4537,
      "theoretical_loss": 3.5736521490740447,
      "tokens_seen": 1251602432
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.000313530591775326,
      "loss": 2.4182,
      "theoretical_loss": 3.573635072878421,
      "tokens_seen": 1251667968
    },
    {
      "epoch": 4.02,
      "objective/train/docs_used": 1439071,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.880246639251709,
      "objective/train/theoretical_loss": 3.5736179978271956,
      "objective/train/tokens_used": 1272193504,
      "theoretical_loss": 3.5736179978271956,
      "tokens_seen": 1251733504
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031352056168505515,
      "loss": 2.4664,
      "theoretical_loss": 3.5736179978271956,
      "tokens_seen": 1251733504
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003135105315947844,
      "loss": 2.862,
      "theoretical_loss": 3.5736009239202327,
      "tokens_seen": 1251799040
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003135005015045135,
      "loss": 2.8991,
      "theoretical_loss": 3.5735838511573945,
      "tokens_seen": 1251864576
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031349047141424275,
      "loss": 2.9548,
      "theoretical_loss": 3.573566779538546,
      "tokens_seen": 1251930112
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031348044132397193,
      "loss": 2.5068,
      "theoretical_loss": 3.5735497090635495,
      "tokens_seen": 1251995648
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003134704112337011,
      "loss": 2.6644,
      "theoretical_loss": 3.573532639732269,
      "tokens_seen": 1252061184
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003134603811434303,
      "loss": 2.6921,
      "theoretical_loss": 3.5735155715445677,
      "tokens_seen": 1252126720
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003134503510531595,
      "loss": 2.7023,
      "theoretical_loss": 3.573498504500309,
      "tokens_seen": 1252192256
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031344032096288865,
      "loss": 2.725,
      "theoretical_loss": 3.573481438599357,
      "tokens_seen": 1252257792
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003134302908726179,
      "loss": 2.7769,
      "theoretical_loss": 3.5734643738415754,
      "tokens_seen": 1252323328
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.000313420260782347,
      "loss": 2.5722,
      "theoretical_loss": 3.573447310226827,
      "tokens_seen": 1252388864
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031341023069207625,
      "loss": 2.7007,
      "theoretical_loss": 3.5734302477549758,
      "tokens_seen": 1252454400
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031340020060180543,
      "loss": 2.6761,
      "theoretical_loss": 3.573413186425886,
      "tokens_seen": 1252519936
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003133901705115346,
      "loss": 2.5791,
      "theoretical_loss": 3.5733961262394205,
      "tokens_seen": 1252585472
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003133801404212638,
      "loss": 2.5603,
      "theoretical_loss": 3.5733790671954435,
      "tokens_seen": 1252651008
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.000313370110330993,
      "loss": 2.617,
      "theoretical_loss": 3.5733620092938185,
      "tokens_seen": 1252716544
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031336008024072216,
      "loss": 2.7296,
      "theoretical_loss": 3.573344952534409,
      "tokens_seen": 1252782080
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003133500501504514,
      "loss": 2.5422,
      "theoretical_loss": 3.573327896917079,
      "tokens_seen": 1252847616
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003133400200601805,
      "loss": 2.6829,
      "theoretical_loss": 3.5733108424416926,
      "tokens_seen": 1252913152
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031332998996990975,
      "loss": 2.48,
      "theoretical_loss": 3.5732937891081136,
      "tokens_seen": 1252978688
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003133199598796389,
      "loss": 2.9448,
      "theoretical_loss": 3.573276736916205,
      "tokens_seen": 1253044224
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003133099297893681,
      "loss": 2.6787,
      "theoretical_loss": 3.5732596858658314,
      "tokens_seen": 1253109760
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003132998996990973,
      "loss": 2.6809,
      "theoretical_loss": 3.5732426359568565,
      "tokens_seen": 1253175296
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003132898696088265,
      "loss": 2.4739,
      "theoretical_loss": 3.5732255871891443,
      "tokens_seen": 1253240832
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031327983951855566,
      "loss": 2.7654,
      "theoretical_loss": 3.5732085395625592,
      "tokens_seen": 1253306368
    },
    {
      "epoch": 4.02,
      "objective/train/docs_used": 1440666,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5984394550323486,
      "objective/train/theoretical_loss": 3.573191493076964,
      "objective/train/tokens_used": 1273831904,
      "theoretical_loss": 3.573191493076964,
      "tokens_seen": 1253371904
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003132698094282849,
      "loss": 2.6108,
      "theoretical_loss": 3.573191493076964,
      "tokens_seen": 1253371904
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.000313259779338014,
      "loss": 2.734,
      "theoretical_loss": 3.5731744477322236,
      "tokens_seen": 1253437440
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031324974924774326,
      "loss": 2.7576,
      "theoretical_loss": 3.5731574035282017,
      "tokens_seen": 1253502976
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003132397191574724,
      "loss": 2.8248,
      "theoretical_loss": 3.5731403604647625,
      "tokens_seen": 1253568512
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003132296890672016,
      "loss": 2.7149,
      "theoretical_loss": 3.5731233185417697,
      "tokens_seen": 1253634048
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003132196589769308,
      "loss": 2.5068,
      "theoretical_loss": 3.5731062777590874,
      "tokens_seen": 1253699584
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031320962888666,
      "loss": 2.4724,
      "theoretical_loss": 3.57308923811658,
      "tokens_seen": 1253765120
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031319959879638916,
      "loss": 2.5921,
      "theoretical_loss": 3.5730721996141117,
      "tokens_seen": 1253830656
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031318956870611834,
      "loss": 2.7723,
      "theoretical_loss": 3.5730551622515465,
      "tokens_seen": 1253896192
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003131795386158475,
      "loss": 2.5817,
      "theoretical_loss": 3.5730381260287483,
      "tokens_seen": 1253961728
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031316950852557676,
      "loss": 2.4434,
      "theoretical_loss": 3.5730210909455815,
      "tokens_seen": 1254027264
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003131594784353059,
      "loss": 2.8277,
      "theoretical_loss": 3.57300405700191,
      "tokens_seen": 1254092800
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003131494483450351,
      "loss": 2.4262,
      "theoretical_loss": 3.572987024197599,
      "tokens_seen": 1254158336
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031313941825476425,
      "loss": 2.32,
      "theoretical_loss": 3.572969992532512,
      "tokens_seen": 1254223872
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003131293881644935,
      "loss": 2.4445,
      "theoretical_loss": 3.5729529620065126,
      "tokens_seen": 1254289408
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031311935807422267,
      "loss": 2.7825,
      "theoretical_loss": 3.5729359326194663,
      "tokens_seen": 1254354944
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031310932798395185,
      "loss": 2.6967,
      "theoretical_loss": 3.572918904371237,
      "tokens_seen": 1254420480
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031309929789368103,
      "loss": 2.5315,
      "theoretical_loss": 3.57290187726169,
      "tokens_seen": 1254486016
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031308926780341026,
      "loss": 2.3605,
      "theoretical_loss": 3.5728848512906874,
      "tokens_seen": 1254551552
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003130792377131394,
      "loss": 2.5921,
      "theoretical_loss": 3.5728678264580958,
      "tokens_seen": 1254617088
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003130692076228686,
      "loss": 2.8032,
      "theoretical_loss": 3.572850802763778,
      "tokens_seen": 1254682624
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003130591775325978,
      "loss": 2.7727,
      "theoretical_loss": 3.5728337802076,
      "tokens_seen": 1254748160
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.000313049147442327,
      "loss": 2.5223,
      "theoretical_loss": 3.5728167587894255,
      "tokens_seen": 1254813696
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003130391173520562,
      "loss": 2.7716,
      "theoretical_loss": 3.5727997385091186,
      "tokens_seen": 1254879232
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031302908726178535,
      "loss": 2.7501,
      "theoretical_loss": 3.5727827193665442,
      "tokens_seen": 1254944768
    },
    {
      "epoch": 4.02,
      "objective/train/docs_used": 1441890,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5261950492858887,
      "objective/train/theoretical_loss": 3.5727657013615666,
      "objective/train/tokens_used": 1275470304,
      "theoretical_loss": 3.5727657013615666,
      "tokens_seen": 1255010304
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003130190571715146,
      "loss": 2.4896,
      "theoretical_loss": 3.5727657013615666,
      "tokens_seen": 1255010304
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003130090270812437,
      "loss": 2.6838,
      "theoretical_loss": 3.572748684494051,
      "tokens_seen": 1255075840
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031299899699097295,
      "loss": 2.6958,
      "theoretical_loss": 3.5727316687638613,
      "tokens_seen": 1255141376
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031298896690070213,
      "loss": 2.7371,
      "theoretical_loss": 3.5727146541708628,
      "tokens_seen": 1255206912
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003129789368104313,
      "loss": 2.6208,
      "theoretical_loss": 3.572697640714919,
      "tokens_seen": 1255272448
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003129689067201605,
      "loss": 2.7077,
      "theoretical_loss": 3.5726806283958963,
      "tokens_seen": 1255337984
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003129588766298897,
      "loss": 2.81,
      "theoretical_loss": 3.5726636172136574,
      "tokens_seen": 1255403520
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031294884653961885,
      "loss": 2.9799,
      "theoretical_loss": 3.5726466071680685,
      "tokens_seen": 1255469056
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003129388164493481,
      "loss": 2.8607,
      "theoretical_loss": 3.5726295982589935,
      "tokens_seen": 1255534592
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003129287863590772,
      "loss": 2.9026,
      "theoretical_loss": 3.5726125904862975,
      "tokens_seen": 1255600128
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031291875626880645,
      "loss": 2.6078,
      "theoretical_loss": 3.5725955838498455,
      "tokens_seen": 1255665664
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031290872617853563,
      "loss": 2.4904,
      "theoretical_loss": 3.5725785783495017,
      "tokens_seen": 1255731200
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003128986960882648,
      "loss": 2.7038,
      "theoretical_loss": 3.572561573985131,
      "tokens_seen": 1255796736
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.000312888665997994,
      "loss": 2.6649,
      "theoretical_loss": 3.5725445707565986,
      "tokens_seen": 1255862272
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003128786359077232,
      "loss": 2.5088,
      "theoretical_loss": 3.5725275686637694,
      "tokens_seen": 1255927808
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031286860581745236,
      "loss": 2.8002,
      "theoretical_loss": 3.572510567706508,
      "tokens_seen": 1255993344
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003128585757271816,
      "loss": 2.9148,
      "theoretical_loss": 3.572493567884679,
      "tokens_seen": 1256058880
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003128485456369107,
      "loss": 2.6956,
      "theoretical_loss": 3.572476569198148,
      "tokens_seen": 1256124416
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031283851554663995,
      "loss": 2.6343,
      "theoretical_loss": 3.57245957164678,
      "tokens_seen": 1256189952
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003128284854563691,
      "loss": 2.7083,
      "theoretical_loss": 3.5724425752304394,
      "tokens_seen": 1256255488
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003128184553660983,
      "loss": 2.7523,
      "theoretical_loss": 3.5724255799489915,
      "tokens_seen": 1256321024
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003128084252758275,
      "loss": 2.9487,
      "theoretical_loss": 3.572408585802301,
      "tokens_seen": 1256386560
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003127983951855567,
      "loss": 2.5786,
      "theoretical_loss": 3.5723915927902334,
      "tokens_seen": 1256452096
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031278836509528586,
      "loss": 2.5373,
      "theoretical_loss": 3.572374600912654,
      "tokens_seen": 1256517632
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003127783350050151,
      "loss": 2.825,
      "theoretical_loss": 3.572357610169427,
      "tokens_seen": 1256583168
    },
    {
      "epoch": 4.02,
      "objective/train/docs_used": 1442594,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8732402324676514,
      "objective/train/theoretical_loss": 3.5723406205604182,
      "objective/train/tokens_used": 1277108704,
      "theoretical_loss": 3.5723406205604182,
      "tokens_seen": 1256648704
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003127683049147442,
      "loss": 2.6564,
      "theoretical_loss": 3.5723406205604182,
      "tokens_seen": 1256648704
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031275827482447346,
      "loss": 2.6124,
      "theoretical_loss": 3.572323632085493,
      "tokens_seen": 1256714240
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003127482447342026,
      "loss": 2.7265,
      "theoretical_loss": 3.5723066447445158,
      "tokens_seen": 1256779776
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003127382146439318,
      "loss": 2.6958,
      "theoretical_loss": 3.572289658537352,
      "tokens_seen": 1256845312
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.000312728184553661,
      "loss": 2.7697,
      "theoretical_loss": 3.572272673463867,
      "tokens_seen": 1256910848
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003127181544633902,
      "loss": 2.5176,
      "theoretical_loss": 3.5722556895239257,
      "tokens_seen": 1256976384
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031270812437311936,
      "loss": 2.8485,
      "theoretical_loss": 3.5722387067173935,
      "tokens_seen": 1257041920
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031269809428284854,
      "loss": 2.9032,
      "theoretical_loss": 3.5722217250441366,
      "tokens_seen": 1257107456
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003126880641925777,
      "loss": 2.7659,
      "theoretical_loss": 3.5722047445040186,
      "tokens_seen": 1257172992
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031267803410230696,
      "loss": 2.6285,
      "theoretical_loss": 3.5721877650969063,
      "tokens_seen": 1257238528
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003126680040120361,
      "loss": 2.7843,
      "theoretical_loss": 3.572170786822664,
      "tokens_seen": 1257304064
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003126579739217653,
      "loss": 2.5797,
      "theoretical_loss": 3.572153809681158,
      "tokens_seen": 1257369600
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031264794383149445,
      "loss": 2.5359,
      "theoretical_loss": 3.572136833672253,
      "tokens_seen": 1257435136
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003126379137412237,
      "loss": 2.5155,
      "theoretical_loss": 3.5721198587958147,
      "tokens_seen": 1257500672
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031262788365095287,
      "loss": 2.6478,
      "theoretical_loss": 3.5721028850517085,
      "tokens_seen": 1257566208
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031261785356068205,
      "loss": 2.7653,
      "theoretical_loss": 3.5720859124397997,
      "tokens_seen": 1257631744
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031260782347041123,
      "loss": 2.6135,
      "theoretical_loss": 3.572068940959954,
      "tokens_seen": 1257697280
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031259779338014046,
      "loss": 2.6955,
      "theoretical_loss": 3.572051970612037,
      "tokens_seen": 1257762816
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003125877632898696,
      "loss": 2.7319,
      "theoretical_loss": 3.5720350013959137,
      "tokens_seen": 1257828352
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003125777331995988,
      "loss": 2.7793,
      "theoretical_loss": 3.5720180333114504,
      "tokens_seen": 1257893888
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031256770310932795,
      "loss": 2.6664,
      "theoretical_loss": 3.572001066358512,
      "tokens_seen": 1257959424
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003125576730190572,
      "loss": 2.7865,
      "theoretical_loss": 3.571984100536964,
      "tokens_seen": 1258024960
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031254764292878637,
      "loss": 2.5398,
      "theoretical_loss": 3.5719671358466734,
      "tokens_seen": 1258090496
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031253761283851555,
      "loss": 2.8289,
      "theoretical_loss": 3.5719501722875044,
      "tokens_seen": 1258156032
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031252758274824473,
      "loss": 2.6467,
      "theoretical_loss": 3.5719332098593233,
      "tokens_seen": 1258221568
    },
    {
      "epoch": 4.02,
      "objective/train/docs_used": 1443926,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4860124588012695,
      "objective/train/theoretical_loss": 3.5719162485619953,
      "objective/train/tokens_used": 1278747104,
      "theoretical_loss": 3.5719162485619953,
      "tokens_seen": 1258287104
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003125175526579739,
      "loss": 2.571,
      "theoretical_loss": 3.5719162485619953,
      "tokens_seen": 1258287104
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003125075225677031,
      "loss": 2.7123,
      "theoretical_loss": 3.5718992883953864,
      "tokens_seen": 1258352640
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031249749247743233,
      "loss": 2.6302,
      "theoretical_loss": 3.571882329359363,
      "tokens_seen": 1258418176
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031248746238716146,
      "loss": 2.7074,
      "theoretical_loss": 3.57186537145379,
      "tokens_seen": 1258483712
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003124774322968907,
      "loss": 2.6631,
      "theoretical_loss": 3.571848414678533,
      "tokens_seen": 1258549248
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003124674022066198,
      "loss": 2.6379,
      "theoretical_loss": 3.5718314590334583,
      "tokens_seen": 1258614784
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031245737211634905,
      "loss": 2.8167,
      "theoretical_loss": 3.571814504518432,
      "tokens_seen": 1258680320
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031244734202607823,
      "loss": 2.6393,
      "theoretical_loss": 3.5717975511333195,
      "tokens_seen": 1258745856
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003124373119358074,
      "loss": 2.586,
      "theoretical_loss": 3.5717805988779867,
      "tokens_seen": 1258811392
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003124272818455366,
      "loss": 2.6975,
      "theoretical_loss": 3.5717636477522996,
      "tokens_seen": 1258876928
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031241725175526583,
      "loss": 2.8738,
      "theoretical_loss": 3.5717466977561245,
      "tokens_seen": 1258942464
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031240722166499496,
      "loss": 2.5827,
      "theoretical_loss": 3.5717297488893265,
      "tokens_seen": 1259008000
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003123971915747242,
      "loss": 2.835,
      "theoretical_loss": 3.5717128011517723,
      "tokens_seen": 1259073536
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003123871614844533,
      "loss": 2.6795,
      "theoretical_loss": 3.5716958545433277,
      "tokens_seen": 1259139072
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031237713139418256,
      "loss": 2.5723,
      "theoretical_loss": 3.5716789090638583,
      "tokens_seen": 1259204608
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031236710130391174,
      "loss": 2.5694,
      "theoretical_loss": 3.5716619647132304,
      "tokens_seen": 1259270144
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003123570712136409,
      "loss": 2.7332,
      "theoretical_loss": 3.5716450214913102,
      "tokens_seen": 1259335680
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003123470411233701,
      "loss": 2.7726,
      "theoretical_loss": 3.5716280793979642,
      "tokens_seen": 1259401216
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003123370110330993,
      "loss": 2.6838,
      "theoretical_loss": 3.5716111384330578,
      "tokens_seen": 1259466752
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031232698094282846,
      "loss": 2.7907,
      "theoretical_loss": 3.5715941985964568,
      "tokens_seen": 1259532288
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003123169508525577,
      "loss": 2.6441,
      "theoretical_loss": 3.571577259888028,
      "tokens_seen": 1259597824
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003123069207622869,
      "loss": 2.5449,
      "theoretical_loss": 3.5715603223076378,
      "tokens_seen": 1259663360
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031229689067201606,
      "loss": 2.6923,
      "theoretical_loss": 3.5715433858551524,
      "tokens_seen": 1259728896
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003122868605817453,
      "loss": 2.7129,
      "theoretical_loss": 3.571526450530437,
      "tokens_seen": 1259794432
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003122768304914744,
      "loss": 2.7213,
      "theoretical_loss": 3.571509516333359,
      "tokens_seen": 1259859968
    },
    {
      "epoch": 4.02,
      "objective/train/docs_used": 1444504,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6051039695739746,
      "objective/train/theoretical_loss": 3.5714925832637836,
      "objective/train/tokens_used": 1280385504,
      "theoretical_loss": 3.5714925832637836,
      "tokens_seen": 1259925504
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031226680040120366,
      "loss": 2.6518,
      "theoretical_loss": 3.5714925832637836,
      "tokens_seen": 1259925504
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003122567703109328,
      "loss": 2.6353,
      "theoretical_loss": 3.571475651321578,
      "tokens_seen": 1259991040
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.000312246740220662,
      "loss": 2.8753,
      "theoretical_loss": 3.571458720506608,
      "tokens_seen": 1260056576
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003122367101303912,
      "loss": 2.6656,
      "theoretical_loss": 3.5714417908187404,
      "tokens_seen": 1260122112
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003122266800401204,
      "loss": 2.6615,
      "theoretical_loss": 3.5714248622578415,
      "tokens_seen": 1260187648
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031221664994984956,
      "loss": 2.6877,
      "theoretical_loss": 3.571407934823777,
      "tokens_seen": 1260253184
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031220661985957874,
      "loss": 2.7432,
      "theoretical_loss": 3.5713910085164136,
      "tokens_seen": 1260318720
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003121965897693079,
      "loss": 2.6483,
      "theoretical_loss": 3.571374083335618,
      "tokens_seen": 1260384256
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031218655967903716,
      "loss": 2.4913,
      "theoretical_loss": 3.5713571592812565,
      "tokens_seen": 1260449792
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003121765295887663,
      "loss": 2.5252,
      "theoretical_loss": 3.571340236353196,
      "tokens_seen": 1260515328
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003121664994984955,
      "loss": 2.9449,
      "theoretical_loss": 3.5713233145513024,
      "tokens_seen": 1260580864
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031215646940822465,
      "loss": 2.8003,
      "theoretical_loss": 3.5713063938754424,
      "tokens_seen": 1260646400
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003121464393179539,
      "loss": 2.5984,
      "theoretical_loss": 3.571289474325482,
      "tokens_seen": 1260711936
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031213640922768307,
      "loss": 2.7581,
      "theoretical_loss": 3.571272555901289,
      "tokens_seen": 1260777472
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031212637913741225,
      "loss": 2.8671,
      "theoretical_loss": 3.5712556386027288,
      "tokens_seen": 1260843008
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031211634904714143,
      "loss": 2.5729,
      "theoretical_loss": 3.571238722429669,
      "tokens_seen": 1260908544
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031210631895687066,
      "loss": 2.7379,
      "theoretical_loss": 3.571221807381975,
      "tokens_seen": 1260974080
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003120962888665998,
      "loss": 2.4806,
      "theoretical_loss": 3.571204893459515,
      "tokens_seen": 1261039616
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.000312086258776329,
      "loss": 2.725,
      "theoretical_loss": 3.5711879806621543,
      "tokens_seen": 1261105152
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031207622868605815,
      "loss": 2.7716,
      "theoretical_loss": 3.5711710689897602,
      "tokens_seen": 1261170688
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003120661985957874,
      "loss": 2.5052,
      "theoretical_loss": 3.5711541584421993,
      "tokens_seen": 1261236224
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031205616850551657,
      "loss": 2.6935,
      "theoretical_loss": 3.5711372490193383,
      "tokens_seen": 1261301760
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031204613841524575,
      "loss": 2.5666,
      "theoretical_loss": 3.571120340721044,
      "tokens_seen": 1261367296
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031203610832497493,
      "loss": 2.6823,
      "theoretical_loss": 3.571103433547184,
      "tokens_seen": 1261432832
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003120260782347041,
      "loss": 2.7387,
      "theoretical_loss": 3.5710865274976236,
      "tokens_seen": 1261498368
    },
    {
      "epoch": 4.02,
      "objective/train/docs_used": 1445499,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4663262367248535,
      "objective/train/theoretical_loss": 3.5710696225722307,
      "objective/train/tokens_used": 1282023904,
      "theoretical_loss": 3.5710696225722307,
      "tokens_seen": 1261563904
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003120160481444333,
      "loss": 2.7363,
      "theoretical_loss": 3.5710696225722307,
      "tokens_seen": 1261563904
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031200601805416253,
      "loss": 2.6551,
      "theoretical_loss": 3.571052718770871,
      "tokens_seen": 1261629440
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031199598796389166,
      "loss": 2.9558,
      "theoretical_loss": 3.5710358160934135,
      "tokens_seen": 1261694976
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003119859578736209,
      "loss": 2.7521,
      "theoretical_loss": 3.571018914539723,
      "tokens_seen": 1261760512
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031197592778335,
      "loss": 2.6555,
      "theoretical_loss": 3.5710020141096677,
      "tokens_seen": 1261826048
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031196589769307925,
      "loss": 2.6724,
      "theoretical_loss": 3.570985114803114,
      "tokens_seen": 1261891584
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031195586760280843,
      "loss": 2.766,
      "theoretical_loss": 3.570968216619929,
      "tokens_seen": 1261957120
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003119458375125376,
      "loss": 2.7561,
      "theoretical_loss": 3.5709513195599794,
      "tokens_seen": 1262022656
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003119358074222668,
      "loss": 2.7715,
      "theoretical_loss": 3.5709344236231324,
      "tokens_seen": 1262088192
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031192577733199603,
      "loss": 2.7651,
      "theoretical_loss": 3.5709175288092556,
      "tokens_seen": 1262153728
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031191574724172516,
      "loss": 2.5331,
      "theoretical_loss": 3.5709006351182153,
      "tokens_seen": 1262219264
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003119057171514544,
      "loss": 2.7933,
      "theoretical_loss": 3.5708837425498787,
      "tokens_seen": 1262284800
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003118956870611835,
      "loss": 2.5899,
      "theoretical_loss": 3.570866851104113,
      "tokens_seen": 1262350336
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031188565697091276,
      "loss": 2.7355,
      "theoretical_loss": 3.570849960780786,
      "tokens_seen": 1262415872
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031187562688064194,
      "loss": 2.7772,
      "theoretical_loss": 3.570833071579764,
      "tokens_seen": 1262481408
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003118655967903711,
      "loss": 2.3021,
      "theoretical_loss": 3.570816183500914,
      "tokens_seen": 1262546944
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003118555667001003,
      "loss": 2.7408,
      "theoretical_loss": 3.570799296544104,
      "tokens_seen": 1262612480
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003118455366098295,
      "loss": 2.6958,
      "theoretical_loss": 3.5707824107092003,
      "tokens_seen": 1262678016
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031183550651955866,
      "loss": 3.0141,
      "theoretical_loss": 3.5707655259960713,
      "tokens_seen": 1262743552
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003118254764292879,
      "loss": 2.7682,
      "theoretical_loss": 3.5707486424045833,
      "tokens_seen": 1262809088
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.000311815446339017,
      "loss": 2.6249,
      "theoretical_loss": 3.5707317599346036,
      "tokens_seen": 1262874624
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031180541624874626,
      "loss": 2.6679,
      "theoretical_loss": 3.5707148785860006,
      "tokens_seen": 1262940160
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003117953861584754,
      "loss": 2.719,
      "theoretical_loss": 3.57069799835864,
      "tokens_seen": 1263005696
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003117853560682046,
      "loss": 2.9154,
      "theoretical_loss": 3.5706811192523906,
      "tokens_seen": 1263071232
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003117753259779338,
      "loss": 2.6642,
      "theoretical_loss": 3.570664241267119,
      "tokens_seen": 1263136768
    },
    {
      "epoch": 4.02,
      "objective/train/docs_used": 1446182,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.874471664428711,
      "objective/train/theoretical_loss": 3.570647364402693,
      "objective/train/tokens_used": 1283662304,
      "theoretical_loss": 3.570647364402693,
      "tokens_seen": 1263202304
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.000311765295887663,
      "loss": 2.6386,
      "theoretical_loss": 3.570647364402693,
      "tokens_seen": 1263202304
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031175526579739217,
      "loss": 2.6591,
      "theoretical_loss": 3.5706304886589795,
      "tokens_seen": 1263267840
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003117452357071214,
      "loss": 2.8144,
      "theoretical_loss": 3.5706136140358464,
      "tokens_seen": 1263333376
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031173520561685053,
      "loss": 2.7972,
      "theoretical_loss": 3.570596740533161,
      "tokens_seen": 1263398912
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031172517552657976,
      "loss": 2.7657,
      "theoretical_loss": 3.5705798681507908,
      "tokens_seen": 1263464448
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003117151454363089,
      "loss": 2.8032,
      "theoretical_loss": 3.5705629968886035,
      "tokens_seen": 1263529984
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003117051153460381,
      "loss": 2.6886,
      "theoretical_loss": 3.570546126746466,
      "tokens_seen": 1263595520
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003116950852557673,
      "loss": 2.5839,
      "theoretical_loss": 3.570529257724247,
      "tokens_seen": 1263661056
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003116850551654965,
      "loss": 2.6547,
      "theoretical_loss": 3.5705123898218125,
      "tokens_seen": 1263726592
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031167502507522567,
      "loss": 2.67,
      "theoretical_loss": 3.5704955230390314,
      "tokens_seen": 1263792128
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031166499498495485,
      "loss": 2.6661,
      "theoretical_loss": 3.570478657375771,
      "tokens_seen": 1263857664
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031165496489468403,
      "loss": 2.6153,
      "theoretical_loss": 3.5704617928318987,
      "tokens_seen": 1263923200
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031164493480441327,
      "loss": 2.671,
      "theoretical_loss": 3.5704449294072824,
      "tokens_seen": 1263988736
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003116349047141424,
      "loss": 2.8392,
      "theoretical_loss": 3.5704280671017896,
      "tokens_seen": 1264054272
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031162487462387163,
      "loss": 2.9227,
      "theoretical_loss": 3.5704112059152884,
      "tokens_seen": 1264119808
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003116148445336008,
      "loss": 2.6377,
      "theoretical_loss": 3.570394345847646,
      "tokens_seen": 1264185344
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031160481444333,
      "loss": 2.5467,
      "theoretical_loss": 3.570377486898731,
      "tokens_seen": 1264250880
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031159478435305917,
      "loss": 2.7076,
      "theoretical_loss": 3.57036062906841,
      "tokens_seen": 1264316416
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031158475426278835,
      "loss": 2.6186,
      "theoretical_loss": 3.5703437723565514,
      "tokens_seen": 1264381952
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031157472417251753,
      "loss": 2.7294,
      "theoretical_loss": 3.5703269167630234,
      "tokens_seen": 1264447488
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031156469408224677,
      "loss": 2.7033,
      "theoretical_loss": 3.5703100622876933,
      "tokens_seen": 1264513024
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031155466399197595,
      "loss": 2.4419,
      "theoretical_loss": 3.570293208930429,
      "tokens_seen": 1264578560
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031154463390170513,
      "loss": 2.6084,
      "theoretical_loss": 3.5702763566910987,
      "tokens_seen": 1264644096
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003115346038114343,
      "loss": 2.5783,
      "theoretical_loss": 3.5702595055695703,
      "tokens_seen": 1264709632
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003115245737211635,
      "loss": 2.7979,
      "theoretical_loss": 3.5702426555657114,
      "tokens_seen": 1264775168
    },
    {
      "epoch": 4.02,
      "objective/train/docs_used": 1446866,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.502274751663208,
      "objective/train/theoretical_loss": 3.57022580667939,
      "objective/train/tokens_used": 1285300704,
      "theoretical_loss": 3.57022580667939,
      "tokens_seen": 1264840704
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031151454363089273,
      "loss": 2.7241,
      "theoretical_loss": 3.57022580667939,
      "tokens_seen": 1264840704
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031150451354062186,
      "loss": 2.6458,
      "theoretical_loss": 3.5702089589104746,
      "tokens_seen": 1264906240
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003114944834503511,
      "loss": 2.719,
      "theoretical_loss": 3.570192112258833,
      "tokens_seen": 1264971776
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003114844533600802,
      "loss": 2.7096,
      "theoretical_loss": 3.5701752667243327,
      "tokens_seen": 1265037312
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031147442326980945,
      "loss": 2.929,
      "theoretical_loss": 3.570158422306842,
      "tokens_seen": 1265102848
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031146439317953863,
      "loss": 3.0498,
      "theoretical_loss": 3.57014157900623,
      "tokens_seen": 1265168384
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003114543630892678,
      "loss": 2.6962,
      "theoretical_loss": 3.570124736822363,
      "tokens_seen": 1265233920
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.000311444332998997,
      "loss": 2.5755,
      "theoretical_loss": 3.5701078957551107,
      "tokens_seen": 1265299456
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031143430290872623,
      "loss": 2.7734,
      "theoretical_loss": 3.5700910558043404,
      "tokens_seen": 1265364992
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031142427281845536,
      "loss": 2.714,
      "theoretical_loss": 3.5700742169699202,
      "tokens_seen": 1265430528
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003114142427281846,
      "loss": 2.6307,
      "theoretical_loss": 3.5700573792517187,
      "tokens_seen": 1265496064
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003114042126379137,
      "loss": 2.5619,
      "theoretical_loss": 3.570040542649604,
      "tokens_seen": 1265561600
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031139418254764296,
      "loss": 2.8593,
      "theoretical_loss": 3.570023707163444,
      "tokens_seen": 1265627136
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031138415245737214,
      "loss": 2.5305,
      "theoretical_loss": 3.5700068727931074,
      "tokens_seen": 1265692672
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003113741223671013,
      "loss": 2.7596,
      "theoretical_loss": 3.569990039538462,
      "tokens_seen": 1265758208
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003113640922768305,
      "loss": 2.7541,
      "theoretical_loss": 3.569973207399377,
      "tokens_seen": 1265823744
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003113540621865597,
      "loss": 2.6803,
      "theoretical_loss": 3.5699563763757194,
      "tokens_seen": 1265889280
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031134403209628886,
      "loss": 2.9374,
      "theoretical_loss": 3.5699395464673587,
      "tokens_seen": 1265954816
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003113340020060181,
      "loss": 2.5742,
      "theoretical_loss": 3.569922717674163,
      "tokens_seen": 1266020352
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003113239719157472,
      "loss": 2.4029,
      "theoretical_loss": 3.569905889996,
      "tokens_seen": 1266085888
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031131394182547646,
      "loss": 2.7946,
      "theoretical_loss": 3.5698890634327385,
      "tokens_seen": 1266151424
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003113039117352056,
      "loss": 2.6271,
      "theoretical_loss": 3.5698722379842476,
      "tokens_seen": 1266216960
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003112938816449348,
      "loss": 2.8781,
      "theoretical_loss": 3.569855413650395,
      "tokens_seen": 1266282496
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.000311283851554664,
      "loss": 2.6544,
      "theoretical_loss": 3.569838590431049,
      "tokens_seen": 1266348032
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003112738214643932,
      "loss": 2.8127,
      "theoretical_loss": 3.5698217683260784,
      "tokens_seen": 1266413568
    },
    {
      "epoch": 4.02,
      "objective/train/docs_used": 1448419,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8012731075286865,
      "objective/train/theoretical_loss": 3.569804947335353,
      "objective/train/tokens_used": 1286939104,
      "theoretical_loss": 3.569804947335353,
      "tokens_seen": 1266479104
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031126379137412237,
      "loss": 2.6329,
      "theoretical_loss": 3.569804947335353,
      "tokens_seen": 1266479104
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003112537612838516,
      "loss": 2.7451,
      "theoretical_loss": 3.5697881274587386,
      "tokens_seen": 1266544640
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031124373119358073,
      "loss": 2.5991,
      "theoretical_loss": 3.5697713086961063,
      "tokens_seen": 1266610176
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031123370110330996,
      "loss": 2.6133,
      "theoretical_loss": 3.5697544910473233,
      "tokens_seen": 1266675712
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003112236710130391,
      "loss": 2.5417,
      "theoretical_loss": 3.569737674512259,
      "tokens_seen": 1266741248
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003112136409227683,
      "loss": 2.6459,
      "theoretical_loss": 3.569720859090781,
      "tokens_seen": 1266806784
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003112036108324975,
      "loss": 2.6508,
      "theoretical_loss": 3.569704044782759,
      "tokens_seen": 1266872320
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003111935807422267,
      "loss": 2.8711,
      "theoretical_loss": 3.5696872315880612,
      "tokens_seen": 1266937856
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031118355065195587,
      "loss": 2.6125,
      "theoretical_loss": 3.5696704195065565,
      "tokens_seen": 1267003392
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031117352056168505,
      "loss": 2.6055,
      "theoretical_loss": 3.569653608538113,
      "tokens_seen": 1267068928
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031116349047141423,
      "loss": 2.6969,
      "theoretical_loss": 3.5696367986826005,
      "tokens_seen": 1267134464
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031115346038114347,
      "loss": 2.6464,
      "theoretical_loss": 3.569619989939887,
      "tokens_seen": 1267200000
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003111434302908726,
      "loss": 2.5475,
      "theoretical_loss": 3.5696031823098413,
      "tokens_seen": 1267265536
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031113340020060183,
      "loss": 2.7453,
      "theoretical_loss": 3.5695863757923325,
      "tokens_seen": 1267331072
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.000311123370110331,
      "loss": 2.5926,
      "theoretical_loss": 3.5695695703872294,
      "tokens_seen": 1267396608
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003111133400200602,
      "loss": 2.687,
      "theoretical_loss": 3.5695527660944006,
      "tokens_seen": 1267462144
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031110330992978937,
      "loss": 2.6837,
      "theoretical_loss": 3.569535962913715,
      "tokens_seen": 1267527680
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031109327983951855,
      "loss": 2.942,
      "theoretical_loss": 3.5695191608450423,
      "tokens_seen": 1267593216
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031108324974924773,
      "loss": 2.4684,
      "theoretical_loss": 3.5695023598882507,
      "tokens_seen": 1267658752
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031107321965897697,
      "loss": 2.6351,
      "theoretical_loss": 3.5694855600432085,
      "tokens_seen": 1267724288
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003110631895687061,
      "loss": 2.4623,
      "theoretical_loss": 3.569468761309786,
      "tokens_seen": 1267789824
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031105315947843533,
      "loss": 2.663,
      "theoretical_loss": 3.5694519636878512,
      "tokens_seen": 1267855360
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031104312938816446,
      "loss": 2.5563,
      "theoretical_loss": 3.569435167177274,
      "tokens_seen": 1267920896
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003110330992978937,
      "loss": 2.6268,
      "theoretical_loss": 3.569418371777923,
      "tokens_seen": 1267986432
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003110230692076229,
      "loss": 2.6797,
      "theoretical_loss": 3.5694015774896664,
      "tokens_seen": 1268051968
    },
    {
      "epoch": 4.02,
      "objective/train/docs_used": 1449162,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7662107944488525,
      "objective/train/theoretical_loss": 3.5693847843123745,
      "objective/train/tokens_used": 1288577504,
      "theoretical_loss": 3.5693847843123745,
      "tokens_seen": 1268117504
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031101303911735206,
      "loss": 2.8176,
      "theoretical_loss": 3.5693847843123745,
      "tokens_seen": 1268117504
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031100300902708124,
      "loss": 2.7783,
      "theoretical_loss": 3.569367992245916,
      "tokens_seen": 1268183040
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003109929789368104,
      "loss": 2.613,
      "theoretical_loss": 3.56935120129016,
      "tokens_seen": 1268248576
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003109829488465396,
      "loss": 2.6012,
      "theoretical_loss": 3.5693344114449754,
      "tokens_seen": 1268314112
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031097291875626883,
      "loss": 2.6034,
      "theoretical_loss": 3.5693176227102317,
      "tokens_seen": 1268379648
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031096288866599796,
      "loss": 2.8255,
      "theoretical_loss": 3.569300835085798,
      "tokens_seen": 1268445184
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003109528585757272,
      "loss": 2.5776,
      "theoretical_loss": 3.5692840485715434,
      "tokens_seen": 1268510720
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003109428284854564,
      "loss": 2.5631,
      "theoretical_loss": 3.5692672631673372,
      "tokens_seen": 1268576256
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031093279839518556,
      "loss": 2.9934,
      "theoretical_loss": 3.569250478873049,
      "tokens_seen": 1268641792
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031092276830491474,
      "loss": 2.7714,
      "theoretical_loss": 3.5692336956885473,
      "tokens_seen": 1268707328
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003109127382146439,
      "loss": 2.7042,
      "theoretical_loss": 3.569216913613702,
      "tokens_seen": 1268772864
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003109027081243731,
      "loss": 2.6858,
      "theoretical_loss": 3.5692001326483824,
      "tokens_seen": 1268838400
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031089267803410234,
      "loss": 2.4343,
      "theoretical_loss": 3.569183352792458,
      "tokens_seen": 1268903936
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031088264794383147,
      "loss": 2.6434,
      "theoretical_loss": 3.569166574045797,
      "tokens_seen": 1268969472
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003108726178535607,
      "loss": 2.7497,
      "theoretical_loss": 3.5691497964082703,
      "tokens_seen": 1269035008
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031086258776328983,
      "loss": 2.7075,
      "theoretical_loss": 3.5691330198797466,
      "tokens_seen": 1269100544
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031085255767301906,
      "loss": 2.7228,
      "theoretical_loss": 3.569116244460096,
      "tokens_seen": 1269166080
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.00031084252758274824,
      "loss": 2.5925,
      "theoretical_loss": 3.5690994701491863,
      "tokens_seen": 1269231616
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003108324974924774,
      "loss": 2.6519,
      "theoretical_loss": 3.5690826969468885,
      "tokens_seen": 1269297152
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003108224674022066,
      "loss": 2.727,
      "theoretical_loss": 3.5690659248530716,
      "tokens_seen": 1269362688
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.0003108124373119358,
      "loss": 2.7352,
      "theoretical_loss": 3.5690491538676055,
      "tokens_seen": 1269428224
    },
    {
      "epoch": 4.02,
      "learning_rate": 0.000310802407221665,
      "loss": 2.61,
      "theoretical_loss": 3.5690323839903586,
      "tokens_seen": 1269493760
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003107923771313942,
      "loss": 2.6601,
      "theoretical_loss": 3.5690156152212023,
      "tokens_seen": 1269559296
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003107823470411234,
      "loss": 2.7277,
      "theoretical_loss": 3.568998847560004,
      "tokens_seen": 1269624832
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00031077231695085257,
      "loss": 2.6063,
      "theoretical_loss": 3.5689820810066353,
      "tokens_seen": 1269690368
    },
    {
      "epoch": 4.03,
      "objective/train/docs_used": 1449667,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5712544918060303,
      "objective/train/theoretical_loss": 3.568965315560965,
      "objective/train/tokens_used": 1290215904,
      "theoretical_loss": 3.568965315560965,
      "tokens_seen": 1269755904
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003107622868605818,
      "loss": 2.55,
      "theoretical_loss": 3.568965315560965,
      "tokens_seen": 1269755904
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00031075225677031093,
      "loss": 2.6712,
      "theoretical_loss": 3.5689485512228623,
      "tokens_seen": 1269821440
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00031074222668004016,
      "loss": 2.435,
      "theoretical_loss": 3.568931787992198,
      "tokens_seen": 1269886976
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003107321965897693,
      "loss": 2.6075,
      "theoretical_loss": 3.56891502586884,
      "tokens_seen": 1269952512
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003107221664994985,
      "loss": 2.5393,
      "theoretical_loss": 3.5688982648526597,
      "tokens_seen": 1270018048
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003107121364092277,
      "loss": 2.5827,
      "theoretical_loss": 3.5688815049435263,
      "tokens_seen": 1270083584
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003107021063189569,
      "loss": 2.6393,
      "theoretical_loss": 3.5688647461413097,
      "tokens_seen": 1270149120
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00031069207622868607,
      "loss": 2.7534,
      "theoretical_loss": 3.5688479884458797,
      "tokens_seen": 1270214656
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00031068204613841525,
      "loss": 2.713,
      "theoretical_loss": 3.5688312318571054,
      "tokens_seen": 1270280192
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00031067201604814443,
      "loss": 2.6002,
      "theoretical_loss": 3.5688144763748575,
      "tokens_seen": 1270345728
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00031066198595787367,
      "loss": 2.7751,
      "theoretical_loss": 3.5687977219990055,
      "tokens_seen": 1270411264
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003106519558676028,
      "loss": 2.5598,
      "theoretical_loss": 3.5687809687294187,
      "tokens_seen": 1270476800
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00031064192577733203,
      "loss": 2.7079,
      "theoretical_loss": 3.5687642165659685,
      "tokens_seen": 1270542336
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003106318956870612,
      "loss": 2.467,
      "theoretical_loss": 3.5687474655085234,
      "tokens_seen": 1270607872
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003106218655967904,
      "loss": 2.8467,
      "theoretical_loss": 3.5687307155569536,
      "tokens_seen": 1270673408
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00031061183550651957,
      "loss": 2.9576,
      "theoretical_loss": 3.56871396671113,
      "tokens_seen": 1270738944
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00031060180541624875,
      "loss": 2.5267,
      "theoretical_loss": 3.5686972189709216,
      "tokens_seen": 1270804480
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00031059177532597793,
      "loss": 2.7555,
      "theoretical_loss": 3.5686804723361982,
      "tokens_seen": 1270870016
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00031058174523570717,
      "loss": 2.6888,
      "theoretical_loss": 3.5686637268068306,
      "tokens_seen": 1270935552
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003105717151454363,
      "loss": 2.7137,
      "theoretical_loss": 3.568646982382689,
      "tokens_seen": 1271001088
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00031056168505516553,
      "loss": 2.8593,
      "theoretical_loss": 3.5686302390636424,
      "tokens_seen": 1271066624
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00031055165496489466,
      "loss": 2.7043,
      "theoretical_loss": 3.568613496849562,
      "tokens_seen": 1271132160
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003105416248746239,
      "loss": 2.8863,
      "theoretical_loss": 3.568596755740317,
      "tokens_seen": 1271197696
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003105315947843531,
      "loss": 2.6993,
      "theoretical_loss": 3.5685800157357788,
      "tokens_seen": 1271263232
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00031052156469408226,
      "loss": 2.6819,
      "theoretical_loss": 3.5685632768358158,
      "tokens_seen": 1271328768
    },
    {
      "epoch": 4.03,
      "objective/train/docs_used": 1450856,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6388139724731445,
      "objective/train/theoretical_loss": 3.5685465390402995,
      "objective/train/tokens_used": 1291854304,
      "theoretical_loss": 3.5685465390402995,
      "tokens_seen": 1271394304
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00031051153460381144,
      "loss": 2.5599,
      "theoretical_loss": 3.5685465390402995,
      "tokens_seen": 1271394304
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003105015045135406,
      "loss": 2.6993,
      "theoretical_loss": 3.5685298023491,
      "tokens_seen": 1271459840
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003104914744232698,
      "loss": 2.7917,
      "theoretical_loss": 3.568513066762087,
      "tokens_seen": 1271525376
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00031048144433299904,
      "loss": 2.7555,
      "theoretical_loss": 3.5684963322791305,
      "tokens_seen": 1271590912
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00031047141424272816,
      "loss": 2.6375,
      "theoretical_loss": 3.568479598900102,
      "tokens_seen": 1271656448
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003104613841524574,
      "loss": 2.581,
      "theoretical_loss": 3.5684628666248708,
      "tokens_seen": 1271721984
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003104513540621866,
      "loss": 2.6204,
      "theoretical_loss": 3.568446135453307,
      "tokens_seen": 1271787520
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00031044132397191576,
      "loss": 2.7401,
      "theoretical_loss": 3.568429405385282,
      "tokens_seen": 1271853056
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00031043129388164494,
      "loss": 2.6455,
      "theoretical_loss": 3.5684126764206656,
      "tokens_seen": 1271918592
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003104212637913741,
      "loss": 2.6757,
      "theoretical_loss": 3.5683959485593277,
      "tokens_seen": 1271984128
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003104112337011033,
      "loss": 2.7583,
      "theoretical_loss": 3.5683792218011394,
      "tokens_seen": 1272049664
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00031040120361083254,
      "loss": 2.589,
      "theoretical_loss": 3.568362496145971,
      "tokens_seen": 1272115200
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00031039117352056167,
      "loss": 2.8462,
      "theoretical_loss": 3.5683457715936924,
      "tokens_seen": 1272180736
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003103811434302909,
      "loss": 2.777,
      "theoretical_loss": 3.5683290481441747,
      "tokens_seen": 1272246272
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00031037111334002003,
      "loss": 2.5992,
      "theoretical_loss": 3.5683123257972884,
      "tokens_seen": 1272311808
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00031036108324974926,
      "loss": 2.7123,
      "theoretical_loss": 3.568295604552903,
      "tokens_seen": 1272377344
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00031035105315947844,
      "loss": 2.9384,
      "theoretical_loss": 3.5682788844108906,
      "tokens_seen": 1272442880
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003103410230692076,
      "loss": 2.812,
      "theoretical_loss": 3.5682621653711206,
      "tokens_seen": 1272508416
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003103309929789368,
      "loss": 2.7367,
      "theoretical_loss": 3.5682454474334637,
      "tokens_seen": 1272573952
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.000310320962888666,
      "loss": 2.761,
      "theoretical_loss": 3.568228730597791,
      "tokens_seen": 1272639488
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00031031093279839517,
      "loss": 2.8438,
      "theoretical_loss": 3.568212014863973,
      "tokens_seen": 1272705024
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003103009027081244,
      "loss": 3.026,
      "theoretical_loss": 3.56819530023188,
      "tokens_seen": 1272770560
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00031029087261785353,
      "loss": 2.6869,
      "theoretical_loss": 3.568178586701383,
      "tokens_seen": 1272836096
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00031028084252758277,
      "loss": 2.7515,
      "theoretical_loss": 3.568161874272352,
      "tokens_seen": 1272901632
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00031027081243731195,
      "loss": 2.6262,
      "theoretical_loss": 3.568145162944659,
      "tokens_seen": 1272967168
    },
    {
      "epoch": 4.03,
      "objective/train/docs_used": 1451655,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.884730577468872,
      "objective/train/theoretical_loss": 3.568128452718174,
      "objective/train/tokens_used": 1293492704,
      "theoretical_loss": 3.568128452718174,
      "tokens_seen": 1273032704
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00031026078234704113,
      "loss": 2.7017,
      "theoretical_loss": 3.568128452718174,
      "tokens_seen": 1273032704
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003102507522567703,
      "loss": 2.817,
      "theoretical_loss": 3.568111743592767,
      "tokens_seen": 1273098240
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003102407221664995,
      "loss": 2.6407,
      "theoretical_loss": 3.5680950355683096,
      "tokens_seen": 1273163776
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00031023069207622867,
      "loss": 2.6872,
      "theoretical_loss": 3.5680783286446727,
      "tokens_seen": 1273229312
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003102206619859579,
      "loss": 2.9628,
      "theoretical_loss": 3.568061622821727,
      "tokens_seen": 1273294848
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00031021063189568703,
      "loss": 2.7832,
      "theoretical_loss": 3.568044918099343,
      "tokens_seen": 1273360384
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00031020060180541627,
      "loss": 2.6355,
      "theoretical_loss": 3.5680282144773923,
      "tokens_seen": 1273425920
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003101905717151454,
      "loss": 2.9272,
      "theoretical_loss": 3.5680115119557447,
      "tokens_seen": 1273491456
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00031018054162487463,
      "loss": 2.7187,
      "theoretical_loss": 3.5679948105342714,
      "tokens_seen": 1273556992
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003101705115346038,
      "loss": 2.4965,
      "theoretical_loss": 3.5679781102128443,
      "tokens_seen": 1273622528
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.000310160481444333,
      "loss": 2.6559,
      "theoretical_loss": 3.5679614109913333,
      "tokens_seen": 1273688064
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003101504513540622,
      "loss": 2.5532,
      "theoretical_loss": 3.5679447128696102,
      "tokens_seen": 1273753600
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003101404212637914,
      "loss": 2.6703,
      "theoretical_loss": 3.567928015847545,
      "tokens_seen": 1273819136
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00031013039117352054,
      "loss": 2.7403,
      "theoretical_loss": 3.567911319925009,
      "tokens_seen": 1273884672
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00031012036108324977,
      "loss": 2.8405,
      "theoretical_loss": 3.567894625101874,
      "tokens_seen": 1273950208
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003101103309929789,
      "loss": 2.8008,
      "theoretical_loss": 3.5678779313780105,
      "tokens_seen": 1274015744
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00031010030090270813,
      "loss": 2.6953,
      "theoretical_loss": 3.567861238753289,
      "tokens_seen": 1274081280
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003100902708124373,
      "loss": 2.7145,
      "theoretical_loss": 3.5678445472275815,
      "tokens_seen": 1274146816
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003100802407221665,
      "loss": 2.7383,
      "theoretical_loss": 3.567827856800759,
      "tokens_seen": 1274212352
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003100702106318957,
      "loss": 2.6356,
      "theoretical_loss": 3.5678111674726924,
      "tokens_seen": 1274277888
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00031006018054162486,
      "loss": 2.8844,
      "theoretical_loss": 3.5677944792432528,
      "tokens_seen": 1274343424
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003100501504513541,
      "loss": 2.789,
      "theoretical_loss": 3.5677777921123113,
      "tokens_seen": 1274408960
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003100401203610833,
      "loss": 2.8134,
      "theoretical_loss": 3.5677611060797396,
      "tokens_seen": 1274474496
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00031003009027081246,
      "loss": 2.5342,
      "theoretical_loss": 3.5677444211454086,
      "tokens_seen": 1274540032
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00031002006018054164,
      "loss": 2.6298,
      "theoretical_loss": 3.5677277373091894,
      "tokens_seen": 1274605568
    },
    {
      "epoch": 4.03,
      "objective/train/docs_used": 1452721,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9962716102600098,
      "objective/train/theoretical_loss": 3.5677110545709536,
      "objective/train/tokens_used": 1295131104,
      "theoretical_loss": 3.5677110545709536,
      "tokens_seen": 1274671104
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003100100300902708,
      "loss": 2.7639,
      "theoretical_loss": 3.5677110545709536,
      "tokens_seen": 1274671104
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00031,
      "loss": 2.8508,
      "theoretical_loss": 3.5676943729305726,
      "tokens_seen": 1274736640
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030998996990972924,
      "loss": 2.6672,
      "theoretical_loss": 3.567677692387917,
      "tokens_seen": 1274802176
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030997993981945836,
      "loss": 2.6414,
      "theoretical_loss": 3.5676610129428585,
      "tokens_seen": 1274867712
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003099699097291876,
      "loss": 2.8492,
      "theoretical_loss": 3.5676443345952693,
      "tokens_seen": 1274933248
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003099598796389168,
      "loss": 2.8881,
      "theoretical_loss": 3.567627657345019,
      "tokens_seen": 1274998784
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030994984954864596,
      "loss": 2.8469,
      "theoretical_loss": 3.567610981191981,
      "tokens_seen": 1275064320
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030993981945837514,
      "loss": 2.5649,
      "theoretical_loss": 3.5675943061360256,
      "tokens_seen": 1275129856
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003099297893681043,
      "loss": 2.7739,
      "theoretical_loss": 3.5675776321770236,
      "tokens_seen": 1275195392
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003099197592778335,
      "loss": 2.5517,
      "theoretical_loss": 3.5675609593148483,
      "tokens_seen": 1275260928
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030990972918756274,
      "loss": 2.7727,
      "theoretical_loss": 3.56754428754937,
      "tokens_seen": 1275326464
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030989969909729187,
      "loss": 2.8845,
      "theoretical_loss": 3.5675276168804597,
      "tokens_seen": 1275392000
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003098896690070211,
      "loss": 2.5749,
      "theoretical_loss": 3.5675109473079902,
      "tokens_seen": 1275457536
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030987963891675023,
      "loss": 2.7791,
      "theoretical_loss": 3.5674942788318322,
      "tokens_seen": 1275523072
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030986960882647946,
      "loss": 2.5838,
      "theoretical_loss": 3.567477611451858,
      "tokens_seen": 1275588608
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030985957873620864,
      "loss": 2.7737,
      "theoretical_loss": 3.567460945167938,
      "tokens_seen": 1275654144
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003098495486459378,
      "loss": 2.6846,
      "theoretical_loss": 3.567444279979945,
      "tokens_seen": 1275719680
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.000309839518555667,
      "loss": 2.7138,
      "theoretical_loss": 3.56742761588775,
      "tokens_seen": 1275785216
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003098294884653962,
      "loss": 2.6993,
      "theoretical_loss": 3.5674109528912252,
      "tokens_seen": 1275850752
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030981945837512537,
      "loss": 2.8416,
      "theoretical_loss": 3.567394290990242,
      "tokens_seen": 1275916288
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003098094282848546,
      "loss": 2.7672,
      "theoretical_loss": 3.567377630184672,
      "tokens_seen": 1275981824
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030979939819458373,
      "loss": 2.9553,
      "theoretical_loss": 3.567360970474387,
      "tokens_seen": 1276047360
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030978936810431297,
      "loss": 2.5491,
      "theoretical_loss": 3.567344311859258,
      "tokens_seen": 1276112896
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030977933801404215,
      "loss": 2.9679,
      "theoretical_loss": 3.567327654339158,
      "tokens_seen": 1276178432
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030976930792377133,
      "loss": 2.7076,
      "theoretical_loss": 3.5673109979139586,
      "tokens_seen": 1276243968
    },
    {
      "epoch": 4.03,
      "objective/train/docs_used": 1453859,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.2308290004730225,
      "objective/train/theoretical_loss": 3.567294342583531,
      "objective/train/tokens_used": 1296769504,
      "theoretical_loss": 3.567294342583531,
      "tokens_seen": 1276309504
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003097592778335005,
      "loss": 2.9256,
      "theoretical_loss": 3.567294342583531,
      "tokens_seen": 1276309504
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003097492477432297,
      "loss": 2.9109,
      "theoretical_loss": 3.5672776883477475,
      "tokens_seen": 1276375040
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030973921765295887,
      "loss": 2.7835,
      "theoretical_loss": 3.5672610352064797,
      "tokens_seen": 1276440576
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003097291875626881,
      "loss": 2.6977,
      "theoretical_loss": 3.5672443831595997,
      "tokens_seen": 1276506112
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030971915747241723,
      "loss": 2.9232,
      "theoretical_loss": 3.567227732206979,
      "tokens_seen": 1276571648
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030970912738214647,
      "loss": 2.8935,
      "theoretical_loss": 3.56721108234849,
      "tokens_seen": 1276637184
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003096990972918756,
      "loss": 3.0577,
      "theoretical_loss": 3.5671944335840045,
      "tokens_seen": 1276702720
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030968906720160483,
      "loss": 2.892,
      "theoretical_loss": 3.567177785913394,
      "tokens_seen": 1276768256
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.000309679037111334,
      "loss": 2.8298,
      "theoretical_loss": 3.5671611393365312,
      "tokens_seen": 1276833792
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003096690070210632,
      "loss": 2.7139,
      "theoretical_loss": 3.567144493853288,
      "tokens_seen": 1276899328
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003096589769307924,
      "loss": 2.8356,
      "theoretical_loss": 3.567127849463536,
      "tokens_seen": 1276964864
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003096489468405216,
      "loss": 2.8401,
      "theoretical_loss": 3.5671112061671475,
      "tokens_seen": 1277030400
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030963891675025074,
      "loss": 2.8479,
      "theoretical_loss": 3.5670945639639946,
      "tokens_seen": 1277095936
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030962888665997997,
      "loss": 2.8155,
      "theoretical_loss": 3.567077922853949,
      "tokens_seen": 1277161472
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003096188565697091,
      "loss": 2.6403,
      "theoretical_loss": 3.567061282836884,
      "tokens_seen": 1277227008
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030960882647943833,
      "loss": 2.7317,
      "theoretical_loss": 3.56704464391267,
      "tokens_seen": 1277292544
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003095987963891675,
      "loss": 2.6416,
      "theoretical_loss": 3.5670280060811805,
      "tokens_seen": 1277358080
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003095887662988967,
      "loss": 2.8647,
      "theoretical_loss": 3.5670113693422874,
      "tokens_seen": 1277423616
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003095787362086259,
      "loss": 2.7733,
      "theoretical_loss": 3.5669947336958625,
      "tokens_seen": 1277489152
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030956870611835506,
      "loss": 2.9185,
      "theoretical_loss": 3.566978099141778,
      "tokens_seen": 1277554688
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030955867602808424,
      "loss": 2.8959,
      "theoretical_loss": 3.566961465679907,
      "tokens_seen": 1277620224
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003095486459378135,
      "loss": 2.9382,
      "theoretical_loss": 3.566944833310121,
      "tokens_seen": 1277685760
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003095386158475426,
      "loss": 2.7745,
      "theoretical_loss": 3.566928202032292,
      "tokens_seen": 1277751296
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030952858575727184,
      "loss": 2.6371,
      "theoretical_loss": 3.566911571846293,
      "tokens_seen": 1277816832
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030951855566700096,
      "loss": 2.8015,
      "theoretical_loss": 3.566894942751997,
      "tokens_seen": 1277882368
    },
    {
      "epoch": 4.03,
      "objective/train/docs_used": 1454164,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9620399475097656,
      "objective/train/theoretical_loss": 3.5668783147492746,
      "objective/train/tokens_used": 1298407904,
      "theoretical_loss": 3.5668783147492746,
      "tokens_seen": 1277947904
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003095085255767302,
      "loss": 2.9085,
      "theoretical_loss": 3.5668783147492746,
      "tokens_seen": 1277947904
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003094984954864594,
      "loss": 2.9253,
      "theoretical_loss": 3.5668616878379993,
      "tokens_seen": 1278013440
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030948846539618856,
      "loss": 2.6875,
      "theoretical_loss": 3.5668450620180425,
      "tokens_seen": 1278078976
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030947843530591774,
      "loss": 2.8784,
      "theoretical_loss": 3.566828437289278,
      "tokens_seen": 1278144512
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.000309468405215647,
      "loss": 2.7315,
      "theoretical_loss": 3.5668118136515776,
      "tokens_seen": 1278210048
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003094583751253761,
      "loss": 2.7856,
      "theoretical_loss": 3.5667951911048137,
      "tokens_seen": 1278275584
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030944834503510534,
      "loss": 2.8026,
      "theoretical_loss": 3.5667785696488585,
      "tokens_seen": 1278341120
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030943831494483447,
      "loss": 2.7517,
      "theoretical_loss": 3.566761949283585,
      "tokens_seen": 1278406656
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003094282848545637,
      "loss": 2.9017,
      "theoretical_loss": 3.5667453300088656,
      "tokens_seen": 1278472192
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003094182547642929,
      "loss": 2.6789,
      "theoretical_loss": 3.566728711824573,
      "tokens_seen": 1278537728
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030940822467402207,
      "loss": 2.7979,
      "theoretical_loss": 3.5667120947305793,
      "tokens_seen": 1278603264
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030939819458375125,
      "loss": 2.7914,
      "theoretical_loss": 3.5666954787267575,
      "tokens_seen": 1278668800
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030938816449348043,
      "loss": 2.665,
      "theoretical_loss": 3.56667886381298,
      "tokens_seen": 1278734336
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003093781344032096,
      "loss": 2.9661,
      "theoretical_loss": 3.566662249989119,
      "tokens_seen": 1278799872
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030936810431293884,
      "loss": 2.9095,
      "theoretical_loss": 3.566645637255048,
      "tokens_seen": 1278865408
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030935807422266797,
      "loss": 2.8724,
      "theoretical_loss": 3.5666290256106397,
      "tokens_seen": 1278930944
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003093480441323972,
      "loss": 2.8794,
      "theoretical_loss": 3.566612415055766,
      "tokens_seen": 1278996480
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030933801404212633,
      "loss": 2.7036,
      "theoretical_loss": 3.5665958055903,
      "tokens_seen": 1279062016
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030932798395185557,
      "loss": 2.6173,
      "theoretical_loss": 3.5665791972141148,
      "tokens_seen": 1279127552
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030931795386158475,
      "loss": 2.891,
      "theoretical_loss": 3.5665625899270825,
      "tokens_seen": 1279193088
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030930792377131393,
      "loss": 3.0057,
      "theoretical_loss": 3.5665459837290765,
      "tokens_seen": 1279258624
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030929789368104317,
      "loss": 2.7406,
      "theoretical_loss": 3.566529378619969,
      "tokens_seen": 1279324160
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030928786359077235,
      "loss": 2.5874,
      "theoretical_loss": 3.5665127745996332,
      "tokens_seen": 1279389696
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030927783350050153,
      "loss": 2.8983,
      "theoretical_loss": 3.566496171667942,
      "tokens_seen": 1279455232
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003092678034102307,
      "loss": 2.8011,
      "theoretical_loss": 3.566479569824768,
      "tokens_seen": 1279520768
    },
    {
      "epoch": 4.03,
      "objective/train/docs_used": 1454164,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.791790723800659,
      "objective/train/theoretical_loss": 3.566462969069984,
      "objective/train/tokens_used": 1300046304,
      "theoretical_loss": 3.566462969069984,
      "tokens_seen": 1279586304
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003092577733199599,
      "loss": 3.0594,
      "theoretical_loss": 3.566462969069984,
      "tokens_seen": 1279586304
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030924774322968907,
      "loss": 2.856,
      "theoretical_loss": 3.5664463694034634,
      "tokens_seen": 1279651840
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003092377131394183,
      "loss": 2.4826,
      "theoretical_loss": 3.5664297708250787,
      "tokens_seen": 1279717376
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030922768304914743,
      "loss": 2.989,
      "theoretical_loss": 3.5664131733347033,
      "tokens_seen": 1279782912
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030921765295887667,
      "loss": 2.8264,
      "theoretical_loss": 3.5663965769322097,
      "tokens_seen": 1279848448
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003092076228686058,
      "loss": 2.9237,
      "theoretical_loss": 3.5663799816174713,
      "tokens_seen": 1279913984
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030919759277833503,
      "loss": 2.7125,
      "theoretical_loss": 3.5663633873903606,
      "tokens_seen": 1279979520
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003091875626880642,
      "loss": 2.7873,
      "theoretical_loss": 3.566346794250751,
      "tokens_seen": 1280045056
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003091775325977934,
      "loss": 2.9169,
      "theoretical_loss": 3.5663302021985155,
      "tokens_seen": 1280110592
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003091675025075226,
      "loss": 3.0449,
      "theoretical_loss": 3.5663136112335274,
      "tokens_seen": 1280176128
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003091574724172518,
      "loss": 3.1691,
      "theoretical_loss": 3.566297021355659,
      "tokens_seen": 1280241664
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030914744232698094,
      "loss": 2.9161,
      "theoretical_loss": 3.5662804325647848,
      "tokens_seen": 1280307200
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030913741223671017,
      "loss": 2.8993,
      "theoretical_loss": 3.5662638448607766,
      "tokens_seen": 1280372736
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003091273821464393,
      "loss": 3.1612,
      "theoretical_loss": 3.566247258243508,
      "tokens_seen": 1280438272
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030911735205616853,
      "loss": 2.7969,
      "theoretical_loss": 3.5662306727128525,
      "tokens_seen": 1280503808
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003091073219658977,
      "loss": 2.802,
      "theoretical_loss": 3.5662140882686835,
      "tokens_seen": 1280569344
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003090972918756269,
      "loss": 2.9697,
      "theoretical_loss": 3.5661975049108734,
      "tokens_seen": 1280634880
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003090872617853561,
      "loss": 2.8291,
      "theoretical_loss": 3.5661809226392958,
      "tokens_seen": 1280700416
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030907723169508526,
      "loss": 2.9147,
      "theoretical_loss": 3.566164341453824,
      "tokens_seen": 1280765952
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030906720160481444,
      "loss": 3.0401,
      "theoretical_loss": 3.5661477613543315,
      "tokens_seen": 1280831488
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003090571715145437,
      "loss": 3.0241,
      "theoretical_loss": 3.5661311823406914,
      "tokens_seen": 1280897024
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003090471414242728,
      "loss": 2.6549,
      "theoretical_loss": 3.5661146044127774,
      "tokens_seen": 1280962560
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030903711133400204,
      "loss": 2.9967,
      "theoretical_loss": 3.566098027570462,
      "tokens_seen": 1281028096
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030902708124373116,
      "loss": 2.6381,
      "theoretical_loss": 3.5660814518136195,
      "tokens_seen": 1281093632
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003090170511534604,
      "loss": 2.6968,
      "theoretical_loss": 3.566064877142123,
      "tokens_seen": 1281159168
    },
    {
      "epoch": 4.03,
      "objective/train/docs_used": 1454920,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6659634113311768,
      "objective/train/theoretical_loss": 3.5660483035558457,
      "objective/train/tokens_used": 1301684704,
      "theoretical_loss": 3.5660483035558457,
      "tokens_seen": 1281224704
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003090070210631896,
      "loss": 2.4981,
      "theoretical_loss": 3.5660483035558457,
      "tokens_seen": 1281224704
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030899699097291876,
      "loss": 2.7639,
      "theoretical_loss": 3.566031731054661,
      "tokens_seen": 1281290240
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030898696088264794,
      "loss": 2.7521,
      "theoretical_loss": 3.566015159638442,
      "tokens_seen": 1281355776
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003089769307923772,
      "loss": 2.7621,
      "theoretical_loss": 3.565998589307064,
      "tokens_seen": 1281421312
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003089669007021063,
      "loss": 3.0341,
      "theoretical_loss": 3.5659820200603987,
      "tokens_seen": 1281486848
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030895687061183554,
      "loss": 2.8547,
      "theoretical_loss": 3.56596545189832,
      "tokens_seen": 1281552384
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030894684052156467,
      "loss": 2.8013,
      "theoretical_loss": 3.5659488848207013,
      "tokens_seen": 1281617920
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003089368104312939,
      "loss": 2.8409,
      "theoretical_loss": 3.565932318827417,
      "tokens_seen": 1281683456
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003089267803410231,
      "loss": 2.8523,
      "theoretical_loss": 3.5659157539183397,
      "tokens_seen": 1281748992
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030891675025075227,
      "loss": 3.1319,
      "theoretical_loss": 3.5658991900933437,
      "tokens_seen": 1281814528
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030890672016048145,
      "loss": 2.5981,
      "theoretical_loss": 3.5658826273523028,
      "tokens_seen": 1281880064
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030889669007021063,
      "loss": 2.8412,
      "theoretical_loss": 3.56586606569509,
      "tokens_seen": 1281945600
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003088866599799398,
      "loss": 2.7512,
      "theoretical_loss": 3.565849505121579,
      "tokens_seen": 1282011136
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030887662988966904,
      "loss": 2.7909,
      "theoretical_loss": 3.5658329456316435,
      "tokens_seen": 1282076672
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030886659979939817,
      "loss": 2.9902,
      "theoretical_loss": 3.5658163872251576,
      "tokens_seen": 1282142208
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003088565697091274,
      "loss": 3.0112,
      "theoretical_loss": 3.5657998299019953,
      "tokens_seen": 1282207744
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030884653961885653,
      "loss": 2.8936,
      "theoretical_loss": 3.5657832736620296,
      "tokens_seen": 1282273280
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030883650952858577,
      "loss": 2.9248,
      "theoretical_loss": 3.5657667185051354,
      "tokens_seen": 1282338816
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030882647943831495,
      "loss": 2.8855,
      "theoretical_loss": 3.5657501644311846,
      "tokens_seen": 1282404352
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030881644934804413,
      "loss": 2.9276,
      "theoretical_loss": 3.5657336114400526,
      "tokens_seen": 1282469888
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003088064192577733,
      "loss": 2.8715,
      "theoretical_loss": 3.5657170595316128,
      "tokens_seen": 1282535424
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030879638916750255,
      "loss": 2.7902,
      "theoretical_loss": 3.565700508705739,
      "tokens_seen": 1282600960
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003087863590772317,
      "loss": 2.809,
      "theoretical_loss": 3.565683958962305,
      "tokens_seen": 1282666496
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003087763289869609,
      "loss": 2.7277,
      "theoretical_loss": 3.5656674103011854,
      "tokens_seen": 1282732032
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030876629889669004,
      "loss": 2.9848,
      "theoretical_loss": 3.5656508627222534,
      "tokens_seen": 1282797568
    },
    {
      "epoch": 4.03,
      "objective/train/docs_used": 1455734,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.738227367401123,
      "objective/train/theoretical_loss": 3.5656343162253825,
      "objective/train/tokens_used": 1303323104,
      "theoretical_loss": 3.5656343162253825,
      "tokens_seen": 1282863104
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030875626880641927,
      "loss": 2.8359,
      "theoretical_loss": 3.5656343162253825,
      "tokens_seen": 1282863104
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030874623871614845,
      "loss": 2.641,
      "theoretical_loss": 3.565617770810448,
      "tokens_seen": 1282928640
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030873620862587763,
      "loss": 2.9164,
      "theoretical_loss": 3.5656012264773227,
      "tokens_seen": 1282994176
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003087261785356068,
      "loss": 2.7244,
      "theoretical_loss": 3.5655846832258815,
      "tokens_seen": 1283059712
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.000308716148445336,
      "loss": 3.0738,
      "theoretical_loss": 3.565568141055998,
      "tokens_seen": 1283125248
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003087061183550652,
      "loss": 2.7018,
      "theoretical_loss": 3.565551599967546,
      "tokens_seen": 1283190784
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003086960882647944,
      "loss": 2.5522,
      "theoretical_loss": 3.5655350599604,
      "tokens_seen": 1283256320
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030868605817452354,
      "loss": 2.6565,
      "theoretical_loss": 3.5655185210344333,
      "tokens_seen": 1283321856
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003086760280842528,
      "loss": 2.8782,
      "theoretical_loss": 3.5655019831895216,
      "tokens_seen": 1283387392
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003086659979939819,
      "loss": 2.9636,
      "theoretical_loss": 3.5654854464255377,
      "tokens_seen": 1283452928
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030865596790371114,
      "loss": 2.8601,
      "theoretical_loss": 3.5654689107423563,
      "tokens_seen": 1283518464
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003086459378134403,
      "loss": 2.5827,
      "theoretical_loss": 3.5654523761398513,
      "tokens_seen": 1283584000
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003086359077231695,
      "loss": 3.0475,
      "theoretical_loss": 3.565435842617897,
      "tokens_seen": 1283649536
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003086258776328987,
      "loss": 2.9249,
      "theoretical_loss": 3.5654193101763685,
      "tokens_seen": 1283715072
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003086158475426279,
      "loss": 2.855,
      "theoretical_loss": 3.5654027788151383,
      "tokens_seen": 1283780608
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030860581745235704,
      "loss": 3.0105,
      "theoretical_loss": 3.5653862485340824,
      "tokens_seen": 1283846144
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003085957873620863,
      "loss": 2.9229,
      "theoretical_loss": 3.5653697193330736,
      "tokens_seen": 1283911680
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003085857572718154,
      "loss": 2.8418,
      "theoretical_loss": 3.565353191211987,
      "tokens_seen": 1283977216
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030857572718154464,
      "loss": 2.8892,
      "theoretical_loss": 3.565336664170697,
      "tokens_seen": 1284042752
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003085656970912738,
      "loss": 2.7801,
      "theoretical_loss": 3.565320138209078,
      "tokens_seen": 1284108288
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.000308555667001003,
      "loss": 2.8704,
      "theoretical_loss": 3.5653036133270035,
      "tokens_seen": 1284173824
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030854563691073224,
      "loss": 2.884,
      "theoretical_loss": 3.565287089524349,
      "tokens_seen": 1284239360
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030853560682046137,
      "loss": 2.8085,
      "theoretical_loss": 3.5652705668009883,
      "tokens_seen": 1284304896
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003085255767301906,
      "loss": 2.5732,
      "theoretical_loss": 3.565254045156796,
      "tokens_seen": 1284370432
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003085155466399198,
      "loss": 2.8816,
      "theoretical_loss": 3.5652375245916463,
      "tokens_seen": 1284435968
    },
    {
      "epoch": 4.03,
      "objective/train/docs_used": 1456993,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8912301063537598,
      "objective/train/theoretical_loss": 3.565221005105414,
      "objective/train/tokens_used": 1304961504,
      "theoretical_loss": 3.565221005105414,
      "tokens_seen": 1284501504
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030850551654964896,
      "loss": 2.9473,
      "theoretical_loss": 3.565221005105414,
      "tokens_seen": 1284501504
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030849548645937814,
      "loss": 2.8586,
      "theoretical_loss": 3.5652044866979735,
      "tokens_seen": 1284567040
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003084854563691074,
      "loss": 2.6153,
      "theoretical_loss": 3.5651879693692,
      "tokens_seen": 1284632576
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003084754262788365,
      "loss": 2.9459,
      "theoretical_loss": 3.565171453118966,
      "tokens_seen": 1284698112
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030846539618856574,
      "loss": 2.7047,
      "theoretical_loss": 3.5651549379471486,
      "tokens_seen": 1284763648
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030845536609829487,
      "loss": 2.8943,
      "theoretical_loss": 3.565138423853621,
      "tokens_seen": 1284829184
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003084453360080241,
      "loss": 2.6728,
      "theoretical_loss": 3.5651219108382577,
      "tokens_seen": 1284894720
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003084353059177533,
      "loss": 2.8053,
      "theoretical_loss": 3.5651053989009336,
      "tokens_seen": 1284960256
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030842527582748247,
      "loss": 2.7473,
      "theoretical_loss": 3.5650888880415237,
      "tokens_seen": 1285025792
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030841524573721165,
      "loss": 2.9395,
      "theoretical_loss": 3.5650723782599023,
      "tokens_seen": 1285091328
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030840521564694083,
      "loss": 2.6765,
      "theoretical_loss": 3.565055869555944,
      "tokens_seen": 1285156864
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030839518555667,
      "loss": 2.8818,
      "theoretical_loss": 3.5650393619295233,
      "tokens_seen": 1285222400
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030838515546639924,
      "loss": 2.7072,
      "theoretical_loss": 3.5650228553805157,
      "tokens_seen": 1285287936
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030837512537612837,
      "loss": 2.7556,
      "theoretical_loss": 3.5650063499087956,
      "tokens_seen": 1285353472
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003083650952858576,
      "loss": 3.0278,
      "theoretical_loss": 3.564989845514237,
      "tokens_seen": 1285419008
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030835506519558673,
      "loss": 2.9658,
      "theoretical_loss": 3.564973342196716,
      "tokens_seen": 1285484544
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030834503510531597,
      "loss": 3.0249,
      "theoretical_loss": 3.5649568399561065,
      "tokens_seen": 1285550080
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030833500501504515,
      "loss": 2.7389,
      "theoretical_loss": 3.564940338792284,
      "tokens_seen": 1285615616
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030832497492477433,
      "loss": 2.589,
      "theoretical_loss": 3.5649238387051225,
      "tokens_seen": 1285681152
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003083149448345035,
      "loss": 2.8527,
      "theoretical_loss": 3.5649073396944972,
      "tokens_seen": 1285746688
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030830491474423275,
      "loss": 2.6517,
      "theoretical_loss": 3.564890841760284,
      "tokens_seen": 1285812224
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003082948846539619,
      "loss": 2.85,
      "theoretical_loss": 3.564874344902356,
      "tokens_seen": 1285877760
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003082848545636911,
      "loss": 2.8013,
      "theoretical_loss": 3.56485784912059,
      "tokens_seen": 1285943296
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030827482447342024,
      "loss": 2.8096,
      "theoretical_loss": 3.564841354414859,
      "tokens_seen": 1286008832
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030826479438314947,
      "loss": 2.8349,
      "theoretical_loss": 3.5648248607850395,
      "tokens_seen": 1286074368
    },
    {
      "epoch": 4.03,
      "objective/train/docs_used": 1457561,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.149444818496704,
      "objective/train/theoretical_loss": 3.5648083682310063,
      "objective/train/tokens_used": 1306599904,
      "theoretical_loss": 3.5648083682310063,
      "tokens_seen": 1286139904
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030825476429287865,
      "loss": 2.9877,
      "theoretical_loss": 3.5648083682310063,
      "tokens_seen": 1286139904
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030824473420260783,
      "loss": 2.9448,
      "theoretical_loss": 3.5647918767526336,
      "tokens_seen": 1286205440
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.000308234704112337,
      "loss": 2.8879,
      "theoretical_loss": 3.5647753863497975,
      "tokens_seen": 1286270976
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003082246740220662,
      "loss": 2.8146,
      "theoretical_loss": 3.5647588970223723,
      "tokens_seen": 1286336512
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003082146439317954,
      "loss": 2.9556,
      "theoretical_loss": 3.5647424087702335,
      "tokens_seen": 1286402048
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003082046138415246,
      "loss": 2.9042,
      "theoretical_loss": 3.564725921593256,
      "tokens_seen": 1286467584
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030819458375125374,
      "loss": 2.7621,
      "theoretical_loss": 3.5647094354913147,
      "tokens_seen": 1286533120
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.000308184553660983,
      "loss": 2.7691,
      "theoretical_loss": 3.564692950464285,
      "tokens_seen": 1286598656
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003081745235707121,
      "loss": 2.7588,
      "theoretical_loss": 3.5646764665120427,
      "tokens_seen": 1286664192
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030816449348044134,
      "loss": 2.8401,
      "theoretical_loss": 3.5646599836344617,
      "tokens_seen": 1286729728
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003081544633901705,
      "loss": 2.8366,
      "theoretical_loss": 3.564643501831418,
      "tokens_seen": 1286795264
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003081444332998997,
      "loss": 2.7475,
      "theoretical_loss": 3.564627021102787,
      "tokens_seen": 1286860800
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003081344032096289,
      "loss": 2.7572,
      "theoretical_loss": 3.5646105414484435,
      "tokens_seen": 1286926336
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003081243731193581,
      "loss": 2.779,
      "theoretical_loss": 3.5645940628682626,
      "tokens_seen": 1286991872
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030811434302908724,
      "loss": 2.9851,
      "theoretical_loss": 3.5645775853621204,
      "tokens_seen": 1287057408
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003081043129388165,
      "loss": 2.7908,
      "theoretical_loss": 3.564561108929891,
      "tokens_seen": 1287122944
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003080942828485456,
      "loss": 2.9733,
      "theoretical_loss": 3.564544633571451,
      "tokens_seen": 1287188480
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030808425275827484,
      "loss": 2.7524,
      "theoretical_loss": 3.5645281592866755,
      "tokens_seen": 1287254016
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.000308074222668004,
      "loss": 2.7138,
      "theoretical_loss": 3.5645116860754387,
      "tokens_seen": 1287319552
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003080641925777332,
      "loss": 2.8349,
      "theoretical_loss": 3.564495213937618,
      "tokens_seen": 1287385088
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003080541624874624,
      "loss": 2.5351,
      "theoretical_loss": 3.564478742873087,
      "tokens_seen": 1287450624
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030804413239719157,
      "loss": 2.5995,
      "theoretical_loss": 3.5644622728817215,
      "tokens_seen": 1287516160
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030803410230692075,
      "loss": 2.8158,
      "theoretical_loss": 3.564445803963398,
      "tokens_seen": 1287581696
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030802407221665,
      "loss": 2.7517,
      "theoretical_loss": 3.564429336117991,
      "tokens_seen": 1287647232
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003080140421263791,
      "loss": 2.9679,
      "theoretical_loss": 3.564412869345376,
      "tokens_seen": 1287712768
    },
    {
      "epoch": 4.03,
      "objective/train/docs_used": 1459012,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.3446168899536133,
      "objective/train/theoretical_loss": 3.5643964036454294,
      "objective/train/tokens_used": 1308238304,
      "theoretical_loss": 3.5643964036454294,
      "tokens_seen": 1287778304
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030800401203610834,
      "loss": 2.78,
      "theoretical_loss": 3.5643964036454294,
      "tokens_seen": 1287778304
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003079939819458375,
      "loss": 2.8933,
      "theoretical_loss": 3.5643799390180257,
      "tokens_seen": 1287843840
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003079839518555667,
      "loss": 2.6891,
      "theoretical_loss": 3.564363475463041,
      "tokens_seen": 1287909376
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003079739217652959,
      "loss": 2.7915,
      "theoretical_loss": 3.5643470129803507,
      "tokens_seen": 1287974912
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030796389167502507,
      "loss": 2.6668,
      "theoretical_loss": 3.5643305515698307,
      "tokens_seen": 1288040448
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030795386158475425,
      "loss": 2.6905,
      "theoretical_loss": 3.5643140912313562,
      "tokens_seen": 1288105984
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003079438314944835,
      "loss": 2.9587,
      "theoretical_loss": 3.5642976319648034,
      "tokens_seen": 1288171520
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003079338014042126,
      "loss": 2.9989,
      "theoretical_loss": 3.5642811737700475,
      "tokens_seen": 1288237056
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030792377131394185,
      "loss": 2.8329,
      "theoretical_loss": 3.5642647166469636,
      "tokens_seen": 1288302592
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.000307913741223671,
      "loss": 2.7683,
      "theoretical_loss": 3.564248260595429,
      "tokens_seen": 1288368128
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003079037111334002,
      "loss": 2.8708,
      "theoretical_loss": 3.5642318056153184,
      "tokens_seen": 1288433664
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003078936810431294,
      "loss": 2.6001,
      "theoretical_loss": 3.5642153517065074,
      "tokens_seen": 1288499200
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030788365095285857,
      "loss": 2.8064,
      "theoretical_loss": 3.564198898868872,
      "tokens_seen": 1288564736
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030787362086258775,
      "loss": 2.7796,
      "theoretical_loss": 3.5641824471022883,
      "tokens_seen": 1288630272
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030786359077231693,
      "loss": 2.6772,
      "theoretical_loss": 3.564165996406632,
      "tokens_seen": 1288695808
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003078535606820461,
      "loss": 3.1331,
      "theoretical_loss": 3.5641495467817785,
      "tokens_seen": 1288761344
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030784353059177535,
      "loss": 2.9419,
      "theoretical_loss": 3.564133098227604,
      "tokens_seen": 1288826880
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003078335005015045,
      "loss": 2.7909,
      "theoretical_loss": 3.5641166507439843,
      "tokens_seen": 1288892416
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003078234704112337,
      "loss": 2.5918,
      "theoretical_loss": 3.5641002043307957,
      "tokens_seen": 1288957952
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003078134403209629,
      "loss": 2.8002,
      "theoretical_loss": 3.564083758987913,
      "tokens_seen": 1289023488
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003078034102306921,
      "loss": 3.0434,
      "theoretical_loss": 3.5640673147152135,
      "tokens_seen": 1289089024
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003077933801404213,
      "loss": 2.5849,
      "theoretical_loss": 3.5640508715125723,
      "tokens_seen": 1289154560
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030778335005015044,
      "loss": 2.9531,
      "theoretical_loss": 3.5640344293798654,
      "tokens_seen": 1289220096
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030777331995987967,
      "loss": 2.8258,
      "theoretical_loss": 3.5640179883169694,
      "tokens_seen": 1289285632
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030776328986960885,
      "loss": 2.7369,
      "theoretical_loss": 3.5640015483237595,
      "tokens_seen": 1289351168
    },
    {
      "epoch": 4.03,
      "objective/train/docs_used": 1461873,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.972330093383789,
      "objective/train/theoretical_loss": 3.5639851094001127,
      "objective/train/tokens_used": 1309876704,
      "theoretical_loss": 3.5639851094001127,
      "tokens_seen": 1289416704
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030775325977933803,
      "loss": 2.8289,
      "theoretical_loss": 3.5639851094001127,
      "tokens_seen": 1289416704
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003077432296890672,
      "loss": 2.6794,
      "theoretical_loss": 3.5639686715459042,
      "tokens_seen": 1289482240
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003077331995987964,
      "loss": 2.9203,
      "theoretical_loss": 3.56395223476101,
      "tokens_seen": 1289547776
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003077231695085256,
      "loss": 2.7811,
      "theoretical_loss": 3.563935799045307,
      "tokens_seen": 1289613312
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003077131394182548,
      "loss": 2.8663,
      "theoretical_loss": 3.5639193643986706,
      "tokens_seen": 1289678848
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030770310932798394,
      "loss": 2.8262,
      "theoretical_loss": 3.5639029308209778,
      "tokens_seen": 1289744384
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003076930792377132,
      "loss": 2.7288,
      "theoretical_loss": 3.5638864983121037,
      "tokens_seen": 1289809920
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003076830491474423,
      "loss": 2.7878,
      "theoretical_loss": 3.5638700668719254,
      "tokens_seen": 1289875456
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030767301905717154,
      "loss": 3.0314,
      "theoretical_loss": 3.5638536365003186,
      "tokens_seen": 1289940992
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003076629889669007,
      "loss": 2.5894,
      "theoretical_loss": 3.5638372071971594,
      "tokens_seen": 1290006528
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003076529588766299,
      "loss": 2.6758,
      "theoretical_loss": 3.5638207789623246,
      "tokens_seen": 1290072064
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003076429287863591,
      "loss": 2.7737,
      "theoretical_loss": 3.5638043517956897,
      "tokens_seen": 1290137600
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003076328986960883,
      "loss": 2.7539,
      "theoretical_loss": 3.563787925697132,
      "tokens_seen": 1290203136
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030762286860581744,
      "loss": 2.8386,
      "theoretical_loss": 3.563771500666527,
      "tokens_seen": 1290268672
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003076128385155467,
      "loss": 2.6868,
      "theoretical_loss": 3.5637550767037514,
      "tokens_seen": 1290334208
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003076028084252758,
      "loss": 2.832,
      "theoretical_loss": 3.563738653808681,
      "tokens_seen": 1290399744
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030759277833500504,
      "loss": 2.8009,
      "theoretical_loss": 3.5637222319811928,
      "tokens_seen": 1290465280
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003075827482447342,
      "loss": 2.6638,
      "theoretical_loss": 3.563705811221163,
      "tokens_seen": 1290530816
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003075727181544634,
      "loss": 2.754,
      "theoretical_loss": 3.5636893915284675,
      "tokens_seen": 1290596352
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003075626880641926,
      "loss": 2.8253,
      "theoretical_loss": 3.563672972902984,
      "tokens_seen": 1290661888
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030755265797392177,
      "loss": 2.7206,
      "theoretical_loss": 3.563656555344588,
      "tokens_seen": 1290727424
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030754262788365095,
      "loss": 2.8326,
      "theoretical_loss": 3.563640138853156,
      "tokens_seen": 1290792960
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003075325977933802,
      "loss": 2.8658,
      "theoretical_loss": 3.5636237234285644,
      "tokens_seen": 1290858496
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003075225677031093,
      "loss": 2.9978,
      "theoretical_loss": 3.56360730907069,
      "tokens_seen": 1290924032
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030751253761283854,
      "loss": 2.8136,
      "theoretical_loss": 3.5635908957794094,
      "tokens_seen": 1290989568
    },
    {
      "epoch": 4.03,
      "objective/train/docs_used": 1467150,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8942673206329346,
      "objective/train/theoretical_loss": 3.563574483554599,
      "objective/train/tokens_used": 1311515104,
      "theoretical_loss": 3.563574483554599,
      "tokens_seen": 1291055104
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003075025075225677,
      "loss": 2.7327,
      "theoretical_loss": 3.563574483554599,
      "tokens_seen": 1291055104
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003074924774322969,
      "loss": 2.8427,
      "theoretical_loss": 3.5635580723961353,
      "tokens_seen": 1291120640
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003074824473420261,
      "loss": 2.8505,
      "theoretical_loss": 3.563541662303895,
      "tokens_seen": 1291186176
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030747241725175527,
      "loss": 2.923,
      "theoretical_loss": 3.563525253277755,
      "tokens_seen": 1291251712
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030746238716148445,
      "loss": 2.7578,
      "theoretical_loss": 3.5635088453175916,
      "tokens_seen": 1291317248
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003074523570712137,
      "loss": 2.9013,
      "theoretical_loss": 3.563492438423281,
      "tokens_seen": 1291382784
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003074423269809428,
      "loss": 2.6727,
      "theoretical_loss": 3.563476032594701,
      "tokens_seen": 1291448320
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030743229689067205,
      "loss": 2.8296,
      "theoretical_loss": 3.563459627831727,
      "tokens_seen": 1291513856
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003074222668004012,
      "loss": 2.6678,
      "theoretical_loss": 3.563443224134237,
      "tokens_seen": 1291579392
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003074122367101304,
      "loss": 2.6133,
      "theoretical_loss": 3.563426821502107,
      "tokens_seen": 1291644928
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003074022066198596,
      "loss": 2.7212,
      "theoretical_loss": 3.563410419935214,
      "tokens_seen": 1291710464
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030739217652958877,
      "loss": 2.7894,
      "theoretical_loss": 3.563394019433434,
      "tokens_seen": 1291776000
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030738214643931795,
      "loss": 2.9254,
      "theoretical_loss": 3.563377619996645,
      "tokens_seen": 1291841536
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030737211634904713,
      "loss": 3.0763,
      "theoretical_loss": 3.5633612216247235,
      "tokens_seen": 1291907072
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003073620862587763,
      "loss": 2.9024,
      "theoretical_loss": 3.563344824317546,
      "tokens_seen": 1291972608
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030735205616850555,
      "loss": 3.1284,
      "theoretical_loss": 3.5633284280749895,
      "tokens_seen": 1292038144
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003073420260782347,
      "loss": 2.679,
      "theoretical_loss": 3.563312032896931,
      "tokens_seen": 1292103680
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003073319959879639,
      "loss": 2.7707,
      "theoretical_loss": 3.563295638783247,
      "tokens_seen": 1292169216
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003073219658976931,
      "loss": 2.778,
      "theoretical_loss": 3.5632792457338147,
      "tokens_seen": 1292234752
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003073119358074223,
      "loss": 2.8644,
      "theoretical_loss": 3.563262853748511,
      "tokens_seen": 1292300288
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030730190571715146,
      "loss": 2.8502,
      "theoretical_loss": 3.5632464628272134,
      "tokens_seen": 1292365824
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030729187562688064,
      "loss": 2.6877,
      "theoretical_loss": 3.563230072969798,
      "tokens_seen": 1292431360
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003072818455366098,
      "loss": 2.8643,
      "theoretical_loss": 3.563213684176142,
      "tokens_seen": 1292496896
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030727181544633905,
      "loss": 2.8738,
      "theoretical_loss": 3.5631972964461234,
      "tokens_seen": 1292562432
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003072617853560682,
      "loss": 2.7507,
      "theoretical_loss": 3.563180909779618,
      "tokens_seen": 1292627968
    },
    {
      "epoch": 4.03,
      "objective/train/docs_used": 1472020,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7531261444091797,
      "objective/train/theoretical_loss": 3.5631645241765026,
      "objective/train/tokens_used": 1313153504,
      "theoretical_loss": 3.5631645241765026,
      "tokens_seen": 1292693504
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003072517552657974,
      "loss": 2.6529,
      "theoretical_loss": 3.5631645241765026,
      "tokens_seen": 1292693504
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030724172517552654,
      "loss": 2.8054,
      "theoretical_loss": 3.563148139636656,
      "tokens_seen": 1292759040
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003072316950852558,
      "loss": 2.904,
      "theoretical_loss": 3.5631317561599545,
      "tokens_seen": 1292824576
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030722166499498496,
      "loss": 2.7999,
      "theoretical_loss": 3.5631153737462746,
      "tokens_seen": 1292890112
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030721163490471414,
      "loss": 2.7538,
      "theoretical_loss": 3.563098992395494,
      "tokens_seen": 1292955648
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003072016048144433,
      "loss": 2.7512,
      "theoretical_loss": 3.5630826121074897,
      "tokens_seen": 1293021184
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003071915747241725,
      "loss": 2.4014,
      "theoretical_loss": 3.5630662328821385,
      "tokens_seen": 1293086720
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003071815446339017,
      "loss": 2.5846,
      "theoretical_loss": 3.563049854719319,
      "tokens_seen": 1293152256
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003071715145436309,
      "loss": 2.75,
      "theoretical_loss": 3.563033477618907,
      "tokens_seen": 1293217792
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030716148445336005,
      "loss": 2.8021,
      "theoretical_loss": 3.5630171015807806,
      "tokens_seen": 1293283328
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003071514543630893,
      "loss": 2.7755,
      "theoretical_loss": 3.563000726604816,
      "tokens_seen": 1293348864
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030714142427281846,
      "loss": 2.8792,
      "theoretical_loss": 3.5629843526908918,
      "tokens_seen": 1293414400
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030713139418254764,
      "loss": 2.8276,
      "theoretical_loss": 3.562967979838885,
      "tokens_seen": 1293479936
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003071213640922768,
      "loss": 2.8245,
      "theoretical_loss": 3.562951608048672,
      "tokens_seen": 1293545472
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.000307111334002006,
      "loss": 2.7646,
      "theoretical_loss": 3.562935237320131,
      "tokens_seen": 1293611008
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003071013039117352,
      "loss": 2.6934,
      "theoretical_loss": 3.5629188676531394,
      "tokens_seen": 1293676544
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003070912738214644,
      "loss": 2.8052,
      "theoretical_loss": 3.5629024990475746,
      "tokens_seen": 1293742080
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030708124373119355,
      "loss": 2.7833,
      "theoretical_loss": 3.562886131503314,
      "tokens_seen": 1293807616
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003070712136409228,
      "loss": 2.8311,
      "theoretical_loss": 3.562869765020234,
      "tokens_seen": 1293873152
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003070611835506519,
      "loss": 2.8201,
      "theoretical_loss": 3.562853399598213,
      "tokens_seen": 1293938688
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030705115346038115,
      "loss": 2.7745,
      "theoretical_loss": 3.562837035237129,
      "tokens_seen": 1294004224
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003070411233701104,
      "loss": 3.03,
      "theoretical_loss": 3.5628206719368585,
      "tokens_seen": 1294069760
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003070310932798395,
      "loss": 2.877,
      "theoretical_loss": 3.5628043096972792,
      "tokens_seen": 1294135296
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030702106318956874,
      "loss": 2.8951,
      "theoretical_loss": 3.5627879485182685,
      "tokens_seen": 1294200832
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003070110330992979,
      "loss": 2.7721,
      "theoretical_loss": 3.562771588399705,
      "tokens_seen": 1294266368
    },
    {
      "epoch": 4.03,
      "objective/train/docs_used": 1477198,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.093344211578369,
      "objective/train/theoretical_loss": 3.562755229341465,
      "objective/train/tokens_used": 1314791904,
      "theoretical_loss": 3.562755229341465,
      "tokens_seen": 1294331904
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003070010030090271,
      "loss": 2.9489,
      "theoretical_loss": 3.562755229341465,
      "tokens_seen": 1294331904
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003069909729187563,
      "loss": 2.6401,
      "theoretical_loss": 3.562738871343427,
      "tokens_seen": 1294397440
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030698094282848547,
      "loss": 2.7694,
      "theoretical_loss": 3.5627225144054684,
      "tokens_seen": 1294462976
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030697091273821465,
      "loss": 2.7897,
      "theoretical_loss": 3.562706158527466,
      "tokens_seen": 1294528512
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003069608826479439,
      "loss": 2.6688,
      "theoretical_loss": 3.562689803709299,
      "tokens_seen": 1294594048
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.000306950852557673,
      "loss": 2.8461,
      "theoretical_loss": 3.562673449950844,
      "tokens_seen": 1294659584
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030694082246740225,
      "loss": 3.0205,
      "theoretical_loss": 3.5626570972519787,
      "tokens_seen": 1294725120
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003069307923771314,
      "loss": 2.9597,
      "theoretical_loss": 3.5626407456125806,
      "tokens_seen": 1294790656
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003069207622868606,
      "loss": 2.7252,
      "theoretical_loss": 3.562624395032528,
      "tokens_seen": 1294856192
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003069107321965898,
      "loss": 2.7873,
      "theoretical_loss": 3.5626080455116993,
      "tokens_seen": 1294921728
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030690070210631897,
      "loss": 2.7035,
      "theoretical_loss": 3.562591697049971,
      "tokens_seen": 1294987264
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030689067201604815,
      "loss": 2.8921,
      "theoretical_loss": 3.5625753496472212,
      "tokens_seen": 1295052800
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030688064192577733,
      "loss": 2.6662,
      "theoretical_loss": 3.5625590033033285,
      "tokens_seen": 1295118336
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003068706118355065,
      "loss": 2.9771,
      "theoretical_loss": 3.5625426580181703,
      "tokens_seen": 1295183872
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030686058174523575,
      "loss": 2.6901,
      "theoretical_loss": 3.562526313791624,
      "tokens_seen": 1295249408
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003068505516549649,
      "loss": 2.9169,
      "theoretical_loss": 3.5625099706235677,
      "tokens_seen": 1295314944
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003068405215646941,
      "loss": 2.6764,
      "theoretical_loss": 3.562493628513879,
      "tokens_seen": 1295380480
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003068304914744233,
      "loss": 2.6589,
      "theoretical_loss": 3.5624772874624373,
      "tokens_seen": 1295446016
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003068204613841525,
      "loss": 2.3541,
      "theoretical_loss": 3.5624609474691185,
      "tokens_seen": 1295511552
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030681043129388166,
      "loss": 2.6614,
      "theoretical_loss": 3.562444608533802,
      "tokens_seen": 1295577088
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030680040120361084,
      "loss": 2.6918,
      "theoretical_loss": 3.5624282706563655,
      "tokens_seen": 1295642624
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030679037111334,
      "loss": 2.4951,
      "theoretical_loss": 3.562411933836686,
      "tokens_seen": 1295708160
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030678034102306925,
      "loss": 2.7433,
      "theoretical_loss": 3.5623955980746436,
      "tokens_seen": 1295773696
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003067703109327984,
      "loss": 2.8738,
      "theoretical_loss": 3.562379263370114,
      "tokens_seen": 1295839232
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003067602808425276,
      "loss": 2.6308,
      "theoretical_loss": 3.562362929722977,
      "tokens_seen": 1295904768
    },
    {
      "epoch": 4.03,
      "objective/train/docs_used": 1480267,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.763779640197754,
      "objective/train/theoretical_loss": 3.56234659713311,
      "objective/train/tokens_used": 1316430304,
      "theoretical_loss": 3.56234659713311,
      "tokens_seen": 1295970304
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030675025075225674,
      "loss": 2.5947,
      "theoretical_loss": 3.56234659713311,
      "tokens_seen": 1295970304
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.000306740220661986,
      "loss": 2.6591,
      "theoretical_loss": 3.562330265600391,
      "tokens_seen": 1296035840
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030673019057171516,
      "loss": 2.8535,
      "theoretical_loss": 3.562313935124698,
      "tokens_seen": 1296101376
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030672016048144434,
      "loss": 2.7508,
      "theoretical_loss": 3.56229760570591,
      "tokens_seen": 1296166912
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003067101303911735,
      "loss": 2.7883,
      "theoretical_loss": 3.562281277343904,
      "tokens_seen": 1296232448
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003067001003009027,
      "loss": 2.8509,
      "theoretical_loss": 3.5622649500385593,
      "tokens_seen": 1296297984
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003066900702106319,
      "loss": 2.8038,
      "theoretical_loss": 3.562248623789753,
      "tokens_seen": 1296363520
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003066800401203611,
      "loss": 2.7852,
      "theoretical_loss": 3.5622322985973645,
      "tokens_seen": 1296429056
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030667001003009025,
      "loss": 2.7057,
      "theoretical_loss": 3.562215974461271,
      "tokens_seen": 1296494592
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003066599799398195,
      "loss": 2.6778,
      "theoretical_loss": 3.562199651381351,
      "tokens_seen": 1296560128
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030664994984954866,
      "loss": 2.7529,
      "theoretical_loss": 3.5621833293574836,
      "tokens_seen": 1296625664
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030663991975927784,
      "loss": 2.5541,
      "theoretical_loss": 3.562167008389546,
      "tokens_seen": 1296691200
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.000306629889669007,
      "loss": 2.7384,
      "theoretical_loss": 3.5621506884774172,
      "tokens_seen": 1296756736
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003066198595787362,
      "loss": 2.9487,
      "theoretical_loss": 3.562134369620975,
      "tokens_seen": 1296822272
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003066098294884654,
      "loss": 2.71,
      "theoretical_loss": 3.5621180518200988,
      "tokens_seen": 1296887808
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003065997993981946,
      "loss": 2.8281,
      "theoretical_loss": 3.562101735074666,
      "tokens_seen": 1296953344
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030658976930792375,
      "loss": 2.6336,
      "theoretical_loss": 3.5620854193845553,
      "tokens_seen": 1297018880
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.000306579739217653,
      "loss": 2.8688,
      "theoretical_loss": 3.562069104749645,
      "tokens_seen": 1297084416
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003065697091273821,
      "loss": 2.8225,
      "theoretical_loss": 3.5620527911698137,
      "tokens_seen": 1297149952
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030655967903711135,
      "loss": 2.664,
      "theoretical_loss": 3.56203647864494,
      "tokens_seen": 1297215488
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030654964894684053,
      "loss": 2.7885,
      "theoretical_loss": 3.5620201671749023,
      "tokens_seen": 1297281024
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003065396188565697,
      "loss": 2.8634,
      "theoretical_loss": 3.562003856759579,
      "tokens_seen": 1297346560
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003065295887662989,
      "loss": 2.7129,
      "theoretical_loss": 3.5619875473988487,
      "tokens_seen": 1297412096
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003065195586760281,
      "loss": 2.7556,
      "theoretical_loss": 3.56197123909259,
      "tokens_seen": 1297477632
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030650952858575725,
      "loss": 2.7793,
      "theoretical_loss": 3.561954931840681,
      "tokens_seen": 1297543168
    },
    {
      "epoch": 4.03,
      "objective/train/docs_used": 1480743,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7817656993865967,
      "objective/train/theoretical_loss": 3.561938625643001,
      "objective/train/tokens_used": 1318068704,
      "theoretical_loss": 3.561938625643001,
      "tokens_seen": 1297608704
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003064994984954865,
      "loss": 2.6356,
      "theoretical_loss": 3.561938625643001,
      "tokens_seen": 1297608704
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003064894684052156,
      "loss": 2.5468,
      "theoretical_loss": 3.5619223204994284,
      "tokens_seen": 1297674240
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030647943831494485,
      "loss": 2.9011,
      "theoretical_loss": 3.5619060164098415,
      "tokens_seen": 1297739776
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030646940822467403,
      "loss": 3.0608,
      "theoretical_loss": 3.5618897133741196,
      "tokens_seen": 1297805312
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003064593781344032,
      "loss": 2.73,
      "theoretical_loss": 3.56187341139214,
      "tokens_seen": 1297870848
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003064493480441324,
      "loss": 2.7114,
      "theoretical_loss": 3.5618571104637833,
      "tokens_seen": 1297936384
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003064393179538616,
      "loss": 2.8433,
      "theoretical_loss": 3.5618408105889268,
      "tokens_seen": 1298001920
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030642928786359076,
      "loss": 2.943,
      "theoretical_loss": 3.5618245117674494,
      "tokens_seen": 1298067456
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030641925777332,
      "loss": 2.9349,
      "theoretical_loss": 3.5618082139992304,
      "tokens_seen": 1298132992
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003064092276830491,
      "loss": 2.8906,
      "theoretical_loss": 3.561791917284148,
      "tokens_seen": 1298198528
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030639919759277835,
      "loss": 2.8165,
      "theoretical_loss": 3.5617756216220817,
      "tokens_seen": 1298264064
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003063891675025075,
      "loss": 2.6757,
      "theoretical_loss": 3.5617593270129095,
      "tokens_seen": 1298329600
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003063791374122367,
      "loss": 2.7742,
      "theoretical_loss": 3.561743033456511,
      "tokens_seen": 1298395136
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003063691073219659,
      "loss": 2.8762,
      "theoretical_loss": 3.561726740952764,
      "tokens_seen": 1298460672
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003063590772316951,
      "loss": 2.4558,
      "theoretical_loss": 3.5617104495015486,
      "tokens_seen": 1298526208
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030634904714142426,
      "loss": 2.8145,
      "theoretical_loss": 3.561694159102743,
      "tokens_seen": 1298591744
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003063390170511535,
      "loss": 2.7078,
      "theoretical_loss": 3.561677869756226,
      "tokens_seen": 1298657280
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003063289869608826,
      "loss": 2.8624,
      "theoretical_loss": 3.561661581461877,
      "tokens_seen": 1298722816
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030631895687061186,
      "loss": 2.8161,
      "theoretical_loss": 3.5616452942195744,
      "tokens_seen": 1298788352
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.000306308926780341,
      "loss": 2.606,
      "theoretical_loss": 3.5616290080291977,
      "tokens_seen": 1298853888
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003062988966900702,
      "loss": 2.7127,
      "theoretical_loss": 3.561612722890626,
      "tokens_seen": 1298919424
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030628886659979945,
      "loss": 2.884,
      "theoretical_loss": 3.561596438803737,
      "tokens_seen": 1298984960
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003062788365095286,
      "loss": 2.7747,
      "theoretical_loss": 3.5615801557684117,
      "tokens_seen": 1299050496
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003062688064192578,
      "loss": 2.6148,
      "theoretical_loss": 3.5615638737845274,
      "tokens_seen": 1299116032
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030625877632898694,
      "loss": 2.895,
      "theoretical_loss": 3.561547592851964,
      "tokens_seen": 1299181568
    },
    {
      "epoch": 4.03,
      "objective/train/docs_used": 1482030,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.784689426422119,
      "objective/train/theoretical_loss": 3.5615313129706005,
      "objective/train/tokens_used": 1319707104,
      "theoretical_loss": 3.5615313129706005,
      "tokens_seen": 1299247104
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003062487462387162,
      "loss": 2.6131,
      "theoretical_loss": 3.5615313129706005,
      "tokens_seen": 1299247104
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030623871614844536,
      "loss": 2.7551,
      "theoretical_loss": 3.5615150341403163,
      "tokens_seen": 1299312640
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030622868605817454,
      "loss": 2.7255,
      "theoretical_loss": 3.56149875636099,
      "tokens_seen": 1299378176
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003062186559679037,
      "loss": 2.6936,
      "theoretical_loss": 3.5614824796325006,
      "tokens_seen": 1299443712
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003062086258776329,
      "loss": 2.6426,
      "theoretical_loss": 3.5614662039547285,
      "tokens_seen": 1299509248
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003061985957873621,
      "loss": 2.7086,
      "theoretical_loss": 3.5614499293275514,
      "tokens_seen": 1299574784
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003061885656970913,
      "loss": 2.8353,
      "theoretical_loss": 3.5614336557508492,
      "tokens_seen": 1299640320
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030617853560682045,
      "loss": 2.8087,
      "theoretical_loss": 3.5614173832245015,
      "tokens_seen": 1299705856
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003061685055165497,
      "loss": 2.8922,
      "theoretical_loss": 3.5614011117483866,
      "tokens_seen": 1299771392
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030615847542627886,
      "loss": 2.7872,
      "theoretical_loss": 3.5613848413223845,
      "tokens_seen": 1299836928
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030614844533600804,
      "loss": 2.7919,
      "theoretical_loss": 3.5613685719463746,
      "tokens_seen": 1299902464
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003061384152457372,
      "loss": 2.6497,
      "theoretical_loss": 3.5613523036202355,
      "tokens_seen": 1299968000
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003061283851554664,
      "loss": 2.6103,
      "theoretical_loss": 3.561336036343847,
      "tokens_seen": 1300033536
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003061183550651956,
      "loss": 2.468,
      "theoretical_loss": 3.5613197701170884,
      "tokens_seen": 1300099072
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003061083249749248,
      "loss": 2.9235,
      "theoretical_loss": 3.561303504939839,
      "tokens_seen": 1300164608
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030609829488465395,
      "loss": 2.8431,
      "theoretical_loss": 3.5612872408119784,
      "tokens_seen": 1300230144
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003060882647943832,
      "loss": 2.9829,
      "theoretical_loss": 3.5612709777333853,
      "tokens_seen": 1300295680
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003060782347041123,
      "loss": 2.7384,
      "theoretical_loss": 3.5612547157039405,
      "tokens_seen": 1300361216
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030606820461384155,
      "loss": 2.9586,
      "theoretical_loss": 3.561238454723522,
      "tokens_seen": 1300426752
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030605817452357073,
      "loss": 2.8804,
      "theoretical_loss": 3.56122219479201,
      "tokens_seen": 1300492288
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003060481444332999,
      "loss": 2.764,
      "theoretical_loss": 3.5612059359092836,
      "tokens_seen": 1300557824
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003060381143430291,
      "loss": 2.8745,
      "theoretical_loss": 3.5611896780752232,
      "tokens_seen": 1300623360
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003060280842527583,
      "loss": 2.8413,
      "theoretical_loss": 3.5611734212897073,
      "tokens_seen": 1300688896
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030601805416248745,
      "loss": 2.6772,
      "theoretical_loss": 3.5611571655526157,
      "tokens_seen": 1300754432
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003060080240722167,
      "loss": 2.6801,
      "theoretical_loss": 3.5611409108638283,
      "tokens_seen": 1300819968
    },
    {
      "epoch": 4.03,
      "objective/train/docs_used": 1482698,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8064839839935303,
      "objective/train/theoretical_loss": 3.5611246572232242,
      "objective/train/tokens_used": 1321345504,
      "theoretical_loss": 3.5611246572232242,
      "tokens_seen": 1300885504
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003059979939819458,
      "loss": 2.7375,
      "theoretical_loss": 3.5611246572232242,
      "tokens_seen": 1300885504
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030598796389167505,
      "loss": 2.8112,
      "theoretical_loss": 3.5611084046306836,
      "tokens_seen": 1300951040
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030597793380140423,
      "loss": 2.9242,
      "theoretical_loss": 3.561092153086086,
      "tokens_seen": 1301016576
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003059679037111334,
      "loss": 2.624,
      "theoretical_loss": 3.5610759025893106,
      "tokens_seen": 1301082112
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003059578736208626,
      "loss": 2.7119,
      "theoretical_loss": 3.5610596531402376,
      "tokens_seen": 1301147648
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003059478435305918,
      "loss": 2.8299,
      "theoretical_loss": 3.5610434047387463,
      "tokens_seen": 1301213184
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030593781344032096,
      "loss": 2.7453,
      "theoretical_loss": 3.5610271573847165,
      "tokens_seen": 1301278720
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003059277833500502,
      "loss": 2.6769,
      "theoretical_loss": 3.561010911078028,
      "tokens_seen": 1301344256
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003059177532597793,
      "loss": 2.5735,
      "theoretical_loss": 3.5609946658185607,
      "tokens_seen": 1301409792
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030590772316950855,
      "loss": 2.814,
      "theoretical_loss": 3.560978421606194,
      "tokens_seen": 1301475328
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003058976930792377,
      "loss": 2.7553,
      "theoretical_loss": 3.560962178440808,
      "tokens_seen": 1301540864
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003058876629889669,
      "loss": 3.0141,
      "theoretical_loss": 3.560945936322282,
      "tokens_seen": 1301606400
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003058776328986961,
      "loss": 2.7595,
      "theoretical_loss": 3.560929695250497,
      "tokens_seen": 1301671936
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003058676028084253,
      "loss": 2.778,
      "theoretical_loss": 3.5609134552253314,
      "tokens_seen": 1301737472
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030585757271815446,
      "loss": 2.8669,
      "theoretical_loss": 3.5608972162466657,
      "tokens_seen": 1301803008
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003058475426278837,
      "loss": 2.7928,
      "theoretical_loss": 3.5608809783143798,
      "tokens_seen": 1301868544
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003058375125376128,
      "loss": 2.6076,
      "theoretical_loss": 3.560864741428354,
      "tokens_seen": 1301934080
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030582748244734206,
      "loss": 2.8176,
      "theoretical_loss": 3.560848505588468,
      "tokens_seen": 1301999616
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003058174523570712,
      "loss": 2.7657,
      "theoretical_loss": 3.560832270794601,
      "tokens_seen": 1302065152
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003058074222668004,
      "loss": 2.6445,
      "theoretical_loss": 3.5608160370466333,
      "tokens_seen": 1302130688
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003057973921765296,
      "loss": 2.8125,
      "theoretical_loss": 3.560799804344446,
      "tokens_seen": 1302196224
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003057873620862588,
      "loss": 2.5955,
      "theoretical_loss": 3.5607835726879173,
      "tokens_seen": 1302261760
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030577733199598796,
      "loss": 2.7631,
      "theoretical_loss": 3.560767342076929,
      "tokens_seen": 1302327296
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030576730190571714,
      "loss": 2.9111,
      "theoretical_loss": 3.56075111251136,
      "tokens_seen": 1302392832
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0003057572718154463,
      "loss": 2.7766,
      "theoretical_loss": 3.5607348839910906,
      "tokens_seen": 1302458368
    },
    {
      "epoch": 4.03,
      "objective/train/docs_used": 1483979,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.209777355194092,
      "objective/train/theoretical_loss": 3.560718656516001,
      "objective/train/tokens_used": 1322983904,
      "theoretical_loss": 3.560718656516001,
      "tokens_seen": 1302523904
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.00030574724172517556,
      "loss": 2.6928,
      "theoretical_loss": 3.560718656516001,
      "tokens_seen": 1302523904
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003057372116349047,
      "loss": 2.7563,
      "theoretical_loss": 3.5607024300859713,
      "tokens_seen": 1302589440
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003057271815446339,
      "loss": 2.7126,
      "theoretical_loss": 3.5606862047008816,
      "tokens_seen": 1302654976
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030571715145436305,
      "loss": 2.735,
      "theoretical_loss": 3.5606699803606117,
      "tokens_seen": 1302720512
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003057071213640923,
      "loss": 2.7698,
      "theoretical_loss": 3.560653757065042,
      "tokens_seen": 1302786048
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030569709127382147,
      "loss": 2.9375,
      "theoretical_loss": 3.5606375348140533,
      "tokens_seen": 1302851584
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030568706118355065,
      "loss": 2.8591,
      "theoretical_loss": 3.560621313607525,
      "tokens_seen": 1302917120
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030567703109327983,
      "loss": 2.5649,
      "theoretical_loss": 3.560605093445338,
      "tokens_seen": 1302982656
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030566700100300906,
      "loss": 2.9389,
      "theoretical_loss": 3.5605888743273715,
      "tokens_seen": 1303048192
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003056569709127382,
      "loss": 2.5524,
      "theoretical_loss": 3.5605726562535067,
      "tokens_seen": 1303113728
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003056469408224674,
      "loss": 2.9435,
      "theoretical_loss": 3.5605564392236237,
      "tokens_seen": 1303179264
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030563691073219655,
      "loss": 2.9062,
      "theoretical_loss": 3.560540223237602,
      "tokens_seen": 1303244800
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003056268806419258,
      "loss": 2.752,
      "theoretical_loss": 3.5605240082953236,
      "tokens_seen": 1303310336
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030561685055165497,
      "loss": 2.8643,
      "theoretical_loss": 3.5605077943966674,
      "tokens_seen": 1303375872
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030560682046138415,
      "loss": 2.9482,
      "theoretical_loss": 3.5604915815415143,
      "tokens_seen": 1303441408
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030559679037111333,
      "loss": 2.8014,
      "theoretical_loss": 3.560475369729744,
      "tokens_seen": 1303506944
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003055867602808425,
      "loss": 2.4184,
      "theoretical_loss": 3.560459158961238,
      "tokens_seen": 1303572480
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003055767301905717,
      "loss": 2.7457,
      "theoretical_loss": 3.5604429492358767,
      "tokens_seen": 1303638016
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030556670010030093,
      "loss": 2.69,
      "theoretical_loss": 3.560426740553539,
      "tokens_seen": 1303703552
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030555667001003006,
      "loss": 2.5995,
      "theoretical_loss": 3.5604105329141067,
      "tokens_seen": 1303769088
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003055466399197593,
      "loss": 2.9198,
      "theoretical_loss": 3.5603943263174607,
      "tokens_seen": 1303834624
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003055366098294885,
      "loss": 2.5802,
      "theoretical_loss": 3.56037812076348,
      "tokens_seen": 1303900160
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030552657973921765,
      "loss": 2.797,
      "theoretical_loss": 3.560361916252046,
      "tokens_seen": 1303965696
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003055165496489469,
      "loss": 2.8847,
      "theoretical_loss": 3.5603457127830396,
      "tokens_seen": 1304031232
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.000305506519558676,
      "loss": 2.694,
      "theoretical_loss": 3.5603295103563406,
      "tokens_seen": 1304096768
    },
    {
      "epoch": 4.04,
      "objective/train/docs_used": 1484701,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.840122699737549,
      "objective/train/theoretical_loss": 3.5603133089718293,
      "objective/train/tokens_used": 1324622304,
      "theoretical_loss": 3.5603133089718293,
      "tokens_seen": 1304162304
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030549648946840525,
      "loss": 2.7326,
      "theoretical_loss": 3.5603133089718293,
      "tokens_seen": 1304162304
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030548645937813443,
      "loss": 2.8274,
      "theoretical_loss": 3.5602971086293875,
      "tokens_seen": 1304227840
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003054764292878636,
      "loss": 2.5019,
      "theoretical_loss": 3.5602809093288954,
      "tokens_seen": 1304293376
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003054663991975928,
      "loss": 2.6509,
      "theoretical_loss": 3.560264711070233,
      "tokens_seen": 1304358912
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.000305456369107322,
      "loss": 2.5558,
      "theoretical_loss": 3.5602485138532813,
      "tokens_seen": 1304424448
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030544633901705116,
      "loss": 2.5316,
      "theoretical_loss": 3.560232317677921,
      "tokens_seen": 1304489984
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003054363089267804,
      "loss": 2.8007,
      "theoretical_loss": 3.560216122544033,
      "tokens_seen": 1304555520
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003054262788365095,
      "loss": 2.9023,
      "theoretical_loss": 3.560199928451498,
      "tokens_seen": 1304621056
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030541624874623875,
      "loss": 2.6966,
      "theoretical_loss": 3.560183735400196,
      "tokens_seen": 1304686592
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003054062186559679,
      "loss": 2.8067,
      "theoretical_loss": 3.5601675433900093,
      "tokens_seen": 1304752128
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003053961885656971,
      "loss": 2.7268,
      "theoretical_loss": 3.5601513524208173,
      "tokens_seen": 1304817664
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003053861584754263,
      "loss": 2.7663,
      "theoretical_loss": 3.560135162492501,
      "tokens_seen": 1304883200
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003053761283851555,
      "loss": 2.6635,
      "theoretical_loss": 3.5601189736049417,
      "tokens_seen": 1304948736
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030536609829488466,
      "loss": 2.8489,
      "theoretical_loss": 3.5601027857580196,
      "tokens_seen": 1305014272
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003053560682046139,
      "loss": 2.8466,
      "theoretical_loss": 3.5600865989516164,
      "tokens_seen": 1305079808
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.000305346038114343,
      "loss": 2.8473,
      "theoretical_loss": 3.5600704131856125,
      "tokens_seen": 1305145344
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030533600802407226,
      "loss": 2.6758,
      "theoretical_loss": 3.5600542284598884,
      "tokens_seen": 1305210880
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003053259779338014,
      "loss": 2.598,
      "theoretical_loss": 3.560038044774325,
      "tokens_seen": 1305276416
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003053159478435306,
      "loss": 2.6272,
      "theoretical_loss": 3.5600218621288047,
      "tokens_seen": 1305341952
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003053059177532598,
      "loss": 2.7393,
      "theoretical_loss": 3.5600056805232065,
      "tokens_seen": 1305407488
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.000305295887662989,
      "loss": 2.8424,
      "theoretical_loss": 3.5599894999574127,
      "tokens_seen": 1305473024
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030528585757271816,
      "loss": 2.7768,
      "theoretical_loss": 3.559973320431304,
      "tokens_seen": 1305538560
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030527582748244734,
      "loss": 2.6083,
      "theoretical_loss": 3.5599571419447607,
      "tokens_seen": 1305604096
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003052657973921765,
      "loss": 2.8214,
      "theoretical_loss": 3.559940964497664,
      "tokens_seen": 1305669632
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030525576730190576,
      "loss": 2.8903,
      "theoretical_loss": 3.559924788089896,
      "tokens_seen": 1305735168
    },
    {
      "epoch": 4.04,
      "objective/train/docs_used": 1485880,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7324297428131104,
      "objective/train/theoretical_loss": 3.5599086127213373,
      "objective/train/tokens_used": 1326260704,
      "theoretical_loss": 3.5599086127213373,
      "tokens_seen": 1305800704
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003052457372116349,
      "loss": 2.7493,
      "theoretical_loss": 3.5599086127213373,
      "tokens_seen": 1305800704
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003052357071213641,
      "loss": 2.8219,
      "theoretical_loss": 3.559892438391868,
      "tokens_seen": 1305866240
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030522567703109325,
      "loss": 3.01,
      "theoretical_loss": 3.559876265101371,
      "tokens_seen": 1305931776
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003052156469408225,
      "loss": 2.8747,
      "theoretical_loss": 3.5598600928497257,
      "tokens_seen": 1305997312
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030520561685055167,
      "loss": 2.5935,
      "theoretical_loss": 3.559843921636814,
      "tokens_seen": 1306062848
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030519558676028085,
      "loss": 2.9397,
      "theoretical_loss": 3.5598277514625174,
      "tokens_seen": 1306128384
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030518555667001003,
      "loss": 2.9605,
      "theoretical_loss": 3.5598115823267165,
      "tokens_seen": 1306193920
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030517552657973926,
      "loss": 2.6671,
      "theoretical_loss": 3.5597954142292925,
      "tokens_seen": 1306259456
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003051654964894684,
      "loss": 2.8636,
      "theoretical_loss": 3.559779247170127,
      "tokens_seen": 1306324992
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003051554663991976,
      "loss": 2.6677,
      "theoretical_loss": 3.559763081149101,
      "tokens_seen": 1306390528
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030514543630892675,
      "loss": 2.8608,
      "theoretical_loss": 3.5597469161660955,
      "tokens_seen": 1306456064
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.000305135406218656,
      "loss": 2.7773,
      "theoretical_loss": 3.559730752220993,
      "tokens_seen": 1306521600
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030512537612838517,
      "loss": 2.5519,
      "theoretical_loss": 3.559714589313673,
      "tokens_seen": 1306587136
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030511534603811435,
      "loss": 2.752,
      "theoretical_loss": 3.5596984274440184,
      "tokens_seen": 1306652672
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030510531594784353,
      "loss": 2.8003,
      "theoretical_loss": 3.55968226661191,
      "tokens_seen": 1306718208
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003050952858575727,
      "loss": 2.7637,
      "theoretical_loss": 3.5596661068172284,
      "tokens_seen": 1306783744
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003050852557673019,
      "loss": 2.5725,
      "theoretical_loss": 3.559649948059856,
      "tokens_seen": 1306849280
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030507522567703113,
      "loss": 2.9491,
      "theoretical_loss": 3.5596337903396735,
      "tokens_seen": 1306914816
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030506519558676026,
      "loss": 3.0898,
      "theoretical_loss": 3.559617633656563,
      "tokens_seen": 1306980352
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003050551654964895,
      "loss": 2.7713,
      "theoretical_loss": 3.559601478010406,
      "tokens_seen": 1307045888
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003050451354062186,
      "loss": 2.734,
      "theoretical_loss": 3.559585323401083,
      "tokens_seen": 1307111424
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030503510531594785,
      "loss": 2.5877,
      "theoretical_loss": 3.559569169828476,
      "tokens_seen": 1307176960
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030502507522567703,
      "loss": 2.5828,
      "theoretical_loss": 3.5595530172924663,
      "tokens_seen": 1307242496
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003050150451354062,
      "loss": 2.6721,
      "theoretical_loss": 3.5595368657929365,
      "tokens_seen": 1307308032
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003050050150451354,
      "loss": 2.7474,
      "theoretical_loss": 3.559520715329767,
      "tokens_seen": 1307373568
    },
    {
      "epoch": 4.04,
      "objective/train/docs_used": 1486544,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7242846488952637,
      "objective/train/theoretical_loss": 3.559504565902839,
      "objective/train/tokens_used": 1327899104,
      "theoretical_loss": 3.559504565902839,
      "tokens_seen": 1307439104
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030499498495486463,
      "loss": 2.6497,
      "theoretical_loss": 3.559504565902839,
      "tokens_seen": 1307439104
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030498495486459376,
      "loss": 2.7098,
      "theoretical_loss": 3.559488417512035,
      "tokens_seen": 1307504640
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.000304974924774323,
      "loss": 2.7243,
      "theoretical_loss": 3.5594722701572366,
      "tokens_seen": 1307570176
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003049648946840521,
      "loss": 2.7246,
      "theoretical_loss": 3.559456123838325,
      "tokens_seen": 1307635712
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030495486459378136,
      "loss": 2.7058,
      "theoretical_loss": 3.559439978555182,
      "tokens_seen": 1307701248
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030494483450351054,
      "loss": 2.8295,
      "theoretical_loss": 3.5594238343076894,
      "tokens_seen": 1307766784
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003049348044132397,
      "loss": 2.8496,
      "theoretical_loss": 3.5594076910957284,
      "tokens_seen": 1307832320
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003049247743229689,
      "loss": 2.6376,
      "theoretical_loss": 3.559391548919181,
      "tokens_seen": 1307897856
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003049147442326981,
      "loss": 2.869,
      "theoretical_loss": 3.559375407777929,
      "tokens_seen": 1307963392
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030490471414242726,
      "loss": 2.6927,
      "theoretical_loss": 3.5593592676718533,
      "tokens_seen": 1308028928
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003048946840521565,
      "loss": 2.7517,
      "theoretical_loss": 3.5593431286008372,
      "tokens_seen": 1308094464
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003048846539618856,
      "loss": 2.6785,
      "theoretical_loss": 3.5593269905647613,
      "tokens_seen": 1308160000
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030487462387161486,
      "loss": 2.5399,
      "theoretical_loss": 3.559310853563508,
      "tokens_seen": 1308225536
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.000304864593781344,
      "loss": 2.8535,
      "theoretical_loss": 3.5592947175969587,
      "tokens_seen": 1308291072
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003048545636910732,
      "loss": 2.7204,
      "theoretical_loss": 3.5592785826649953,
      "tokens_seen": 1308356608
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003048445336008024,
      "loss": 2.5783,
      "theoretical_loss": 3.5592624487675,
      "tokens_seen": 1308422144
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003048345035105316,
      "loss": 2.5697,
      "theoretical_loss": 3.559246315904354,
      "tokens_seen": 1308487680
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030482447342026076,
      "loss": 2.7008,
      "theoretical_loss": 3.55923018407544,
      "tokens_seen": 1308553216
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030481444332999,
      "loss": 2.6793,
      "theoretical_loss": 3.5592140532806393,
      "tokens_seen": 1308618752
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003048044132397192,
      "loss": 2.6712,
      "theoretical_loss": 3.559197923519834,
      "tokens_seen": 1308684288
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030479438314944836,
      "loss": 2.9295,
      "theoretical_loss": 3.559181794792906,
      "tokens_seen": 1308749824
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030478435305917754,
      "loss": 2.8828,
      "theoretical_loss": 3.5591656670997374,
      "tokens_seen": 1308815360
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003047743229689067,
      "loss": 2.6114,
      "theoretical_loss": 3.5591495404402096,
      "tokens_seen": 1308880896
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030476429287863596,
      "loss": 2.6058,
      "theoretical_loss": 3.5591334148142058,
      "tokens_seen": 1308946432
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003047542627883651,
      "loss": 2.7359,
      "theoretical_loss": 3.559117290221607,
      "tokens_seen": 1309011968
    },
    {
      "epoch": 4.04,
      "objective/train/docs_used": 1488092,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8345699310302734,
      "objective/train/theoretical_loss": 3.5591011666622956,
      "objective/train/tokens_used": 1329537504,
      "theoretical_loss": 3.5591011666622956,
      "tokens_seen": 1309077504
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003047442326980943,
      "loss": 2.831,
      "theoretical_loss": 3.5591011666622956,
      "tokens_seen": 1309077504
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030473420260782345,
      "loss": 2.7527,
      "theoretical_loss": 3.5590850441361535,
      "tokens_seen": 1309143040
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003047241725175527,
      "loss": 2.9376,
      "theoretical_loss": 3.559068922643063,
      "tokens_seen": 1309208576
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030471414242728187,
      "loss": 2.8627,
      "theoretical_loss": 3.559052802182906,
      "tokens_seen": 1309274112
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030470411233701105,
      "loss": 2.8548,
      "theoretical_loss": 3.559036682755565,
      "tokens_seen": 1309339648
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030469408224674023,
      "loss": 2.5759,
      "theoretical_loss": 3.559020564360922,
      "tokens_seen": 1309405184
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030468405215646946,
      "loss": 2.7288,
      "theoretical_loss": 3.559004446998858,
      "tokens_seen": 1309470720
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003046740220661986,
      "loss": 2.814,
      "theoretical_loss": 3.5589883306692576,
      "tokens_seen": 1309536256
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003046639919759278,
      "loss": 2.7851,
      "theoretical_loss": 3.558972215372001,
      "tokens_seen": 1309601792
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030465396188565695,
      "loss": 2.7324,
      "theoretical_loss": 3.5589561011069706,
      "tokens_seen": 1309667328
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003046439317953862,
      "loss": 2.5386,
      "theoretical_loss": 3.5589399878740493,
      "tokens_seen": 1309732864
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030463390170511537,
      "loss": 2.7966,
      "theoretical_loss": 3.558923875673119,
      "tokens_seen": 1309798400
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030462387161484455,
      "loss": 2.8575,
      "theoretical_loss": 3.5589077645040623,
      "tokens_seen": 1309863936
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030461384152457373,
      "loss": 2.7993,
      "theoretical_loss": 3.5588916543667617,
      "tokens_seen": 1309929472
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003046038114343029,
      "loss": 2.8267,
      "theoretical_loss": 3.558875545261098,
      "tokens_seen": 1309995008
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003045937813440321,
      "loss": 2.6302,
      "theoretical_loss": 3.558859437186956,
      "tokens_seen": 1310060544
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030458375125376133,
      "loss": 2.8977,
      "theoretical_loss": 3.5588433301442155,
      "tokens_seen": 1310126080
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030457372116349046,
      "loss": 2.5146,
      "theoretical_loss": 3.5588272241327603,
      "tokens_seen": 1310191616
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003045636910732197,
      "loss": 2.5984,
      "theoretical_loss": 3.5588111191524727,
      "tokens_seen": 1310257152
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003045536609829488,
      "loss": 2.7796,
      "theoretical_loss": 3.558795015203235,
      "tokens_seen": 1310322688
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030454363089267805,
      "loss": 2.5794,
      "theoretical_loss": 3.5587789122849296,
      "tokens_seen": 1310388224
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030453360080240723,
      "loss": 2.601,
      "theoretical_loss": 3.5587628103974387,
      "tokens_seen": 1310453760
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003045235707121364,
      "loss": 2.7363,
      "theoretical_loss": 3.558746709540645,
      "tokens_seen": 1310519296
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003045135406218656,
      "loss": 2.8367,
      "theoretical_loss": 3.5587306097144307,
      "tokens_seen": 1310584832
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030450351053159483,
      "loss": 2.9193,
      "theoretical_loss": 3.558714510918679,
      "tokens_seen": 1310650368
    },
    {
      "debugging/Self-BLEU-5": 0.5335497100843765,
      "debugging/distinct-1-grams": 0.7520819374414864,
      "debugging/distinct-2-grams": 0.9491809766666045,
      "debugging/entropy-1-grams": 5.943300109412918,
      "debugging/entropy-2-grams": 7.027315581762769,
      "debugging/length": 477.3888888888889,
      "debugging/num_segments": 18,
      "debugging/score": 0.0018735238636605285,
      "debugging/score_std": 0.0038943973989434763,
      "epoch": 4.04,
      "objective/train/docs_used": 1488756,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.642608642578125,
      "objective/train/theoretical_loss": 3.558698413153272,
      "objective/train/tokens_used": 1331175904,
      "theoretical_loss": 3.558698413153272,
      "tokens_seen": 1310715904
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030449348044132396,
      "loss": 2.7676,
      "theoretical_loss": 3.558698413153272,
      "tokens_seen": 1310715904
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003044834503510532,
      "loss": 2.6587,
      "theoretical_loss": 3.558682316418092,
      "tokens_seen": 1310781440
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003044734202607823,
      "loss": 2.7785,
      "theoretical_loss": 3.5586662207130217,
      "tokens_seen": 1310846976
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030446339017051156,
      "loss": 2.6989,
      "theoretical_loss": 3.558650126037944,
      "tokens_seen": 1310912512
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030445336008024074,
      "loss": 2.6195,
      "theoretical_loss": 3.5586340323927415,
      "tokens_seen": 1310978048
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003044433299899699,
      "loss": 2.8218,
      "theoretical_loss": 3.558617939777297,
      "tokens_seen": 1311043584
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003044332998996991,
      "loss": 2.6592,
      "theoretical_loss": 3.5586018481914916,
      "tokens_seen": 1311109120
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003044232698094283,
      "loss": 2.3855,
      "theoretical_loss": 3.5585857576352096,
      "tokens_seen": 1311174656
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030441323971915746,
      "loss": 2.6516,
      "theoretical_loss": 3.5585696681083334,
      "tokens_seen": 1311240192
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003044032096288867,
      "loss": 2.6568,
      "theoretical_loss": 3.558553579610745,
      "tokens_seen": 1311305728
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003043931795386158,
      "loss": 2.5956,
      "theoretical_loss": 3.558537492142328,
      "tokens_seen": 1311371264
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030438314944834506,
      "loss": 2.7413,
      "theoretical_loss": 3.558521405702965,
      "tokens_seen": 1311436800
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003043731193580742,
      "loss": 2.4865,
      "theoretical_loss": 3.558505320292538,
      "tokens_seen": 1311502336
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003043630892678034,
      "loss": 2.5645,
      "theoretical_loss": 3.5584892359109306,
      "tokens_seen": 1311567872
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003043530591775326,
      "loss": 2.7924,
      "theoretical_loss": 3.558473152558025,
      "tokens_seen": 1311633408
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003043430290872618,
      "loss": 2.6965,
      "theoretical_loss": 3.558457070233704,
      "tokens_seen": 1311698944
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030433299899699097,
      "loss": 2.5144,
      "theoretical_loss": 3.558440988937851,
      "tokens_seen": 1311764480
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003043229689067202,
      "loss": 2.6407,
      "theoretical_loss": 3.558424908670349,
      "tokens_seen": 1311830016
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030431293881644933,
      "loss": 2.6782,
      "theoretical_loss": 3.5584088294310803,
      "tokens_seen": 1311895552
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030430290872617856,
      "loss": 2.6771,
      "theoretical_loss": 3.558392751219927,
      "tokens_seen": 1311961088
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003042928786359077,
      "loss": 2.7834,
      "theoretical_loss": 3.558376674036774,
      "tokens_seen": 1312026624
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003042828485456369,
      "loss": 2.7159,
      "theoretical_loss": 3.558360597881503,
      "tokens_seen": 1312092160
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003042728184553661,
      "loss": 2.6537,
      "theoretical_loss": 3.5583445227539965,
      "tokens_seen": 1312157696
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003042627883650953,
      "loss": 2.79,
      "theoretical_loss": 3.5583284486541387,
      "tokens_seen": 1312223232
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030425275827482447,
      "loss": 2.9484,
      "theoretical_loss": 3.558312375581812,
      "tokens_seen": 1312288768
    },
    {
      "epoch": 4.04,
      "objective/train/docs_used": 1490224,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.775263786315918,
      "objective/train/theoretical_loss": 3.5582963035368986,
      "objective/train/tokens_used": 1332814304,
      "theoretical_loss": 3.5582963035368986,
      "tokens_seen": 1312354304
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030424272818455365,
      "loss": 2.5093,
      "theoretical_loss": 3.5582963035368986,
      "tokens_seen": 1312354304
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030423269809428283,
      "loss": 2.9231,
      "theoretical_loss": 3.558280232519283,
      "tokens_seen": 1312419840
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030422266800401207,
      "loss": 2.684,
      "theoretical_loss": 3.558264162528847,
      "tokens_seen": 1312485376
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003042126379137412,
      "loss": 2.7024,
      "theoretical_loss": 3.5582480935654743,
      "tokens_seen": 1312550912
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030420260782347043,
      "loss": 2.7182,
      "theoretical_loss": 3.5582320256290485,
      "tokens_seen": 1312616448
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003041925777331996,
      "loss": 2.6307,
      "theoretical_loss": 3.5582159587194515,
      "tokens_seen": 1312681984
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003041825476429288,
      "loss": 2.626,
      "theoretical_loss": 3.5581998928365675,
      "tokens_seen": 1312747520
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030417251755265797,
      "loss": 2.7163,
      "theoretical_loss": 3.5581838279802787,
      "tokens_seen": 1312813056
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030416248746238715,
      "loss": 3.0121,
      "theoretical_loss": 3.5581677641504683,
      "tokens_seen": 1312878592
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030415245737211633,
      "loss": 2.7012,
      "theoretical_loss": 3.558151701347021,
      "tokens_seen": 1312944128
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030414242728184557,
      "loss": 2.6484,
      "theoretical_loss": 3.5581356395698176,
      "tokens_seen": 1313009664
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003041323971915747,
      "loss": 2.6568,
      "theoretical_loss": 3.5581195788187436,
      "tokens_seen": 1313075200
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030412236710130393,
      "loss": 2.6755,
      "theoretical_loss": 3.558103519093681,
      "tokens_seen": 1313140736
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030411233701103306,
      "loss": 2.6254,
      "theoretical_loss": 3.558087460394513,
      "tokens_seen": 1313206272
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003041023069207623,
      "loss": 2.6578,
      "theoretical_loss": 3.558071402721123,
      "tokens_seen": 1313271808
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003040922768304915,
      "loss": 2.8261,
      "theoretical_loss": 3.558055346073395,
      "tokens_seen": 1313337344
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030408224674022066,
      "loss": 2.8016,
      "theoretical_loss": 3.5580392904512115,
      "tokens_seen": 1313402880
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030407221664994984,
      "loss": 2.7408,
      "theoretical_loss": 3.558023235854456,
      "tokens_seen": 1313468416
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.000304062186559679,
      "loss": 2.6432,
      "theoretical_loss": 3.5580071822830117,
      "tokens_seen": 1313533952
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030405215646940825,
      "loss": 2.6584,
      "theoretical_loss": 3.5579911297367626,
      "tokens_seen": 1313599488
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030404212637913743,
      "loss": 2.5022,
      "theoretical_loss": 3.5579750782155917,
      "tokens_seen": 1313665024
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003040320962888666,
      "loss": 2.6291,
      "theoretical_loss": 3.557959027719382,
      "tokens_seen": 1313730560
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003040220661985958,
      "loss": 2.7724,
      "theoretical_loss": 3.557942978248018,
      "tokens_seen": 1313796096
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030401203610832503,
      "loss": 2.5832,
      "theoretical_loss": 3.557926929801382,
      "tokens_seen": 1313861632
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030400200601805416,
      "loss": 2.4701,
      "theoretical_loss": 3.5579108823793577,
      "tokens_seen": 1313927168
    },
    {
      "epoch": 4.04,
      "objective/train/docs_used": 1490984,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.301514148712158,
      "objective/train/theoretical_loss": 3.5578948359818297,
      "objective/train/tokens_used": 1334452704,
      "theoretical_loss": 3.5578948359818297,
      "tokens_seen": 1313992704
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003039919759277834,
      "loss": 2.5106,
      "theoretical_loss": 3.5578948359818297,
      "tokens_seen": 1313992704
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003039819458375125,
      "loss": 2.7581,
      "theoretical_loss": 3.5578787906086795,
      "tokens_seen": 1314058240
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030397191574724176,
      "loss": 2.7226,
      "theoretical_loss": 3.5578627462597927,
      "tokens_seen": 1314123776
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030396188565697094,
      "loss": 2.9072,
      "theoretical_loss": 3.557846702935051,
      "tokens_seen": 1314189312
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003039518555667001,
      "loss": 2.5841,
      "theoretical_loss": 3.5578306606343393,
      "tokens_seen": 1314254848
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003039418254764293,
      "loss": 2.3466,
      "theoretical_loss": 3.557814619357541,
      "tokens_seen": 1314320384
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003039317953861585,
      "loss": 2.5913,
      "theoretical_loss": 3.557798579104539,
      "tokens_seen": 1314385920
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030392176529588766,
      "loss": 2.6581,
      "theoretical_loss": 3.5577825398752174,
      "tokens_seen": 1314451456
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003039117352056169,
      "loss": 2.61,
      "theoretical_loss": 3.5577665016694597,
      "tokens_seen": 1314516992
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.000303901705115346,
      "loss": 2.9948,
      "theoretical_loss": 3.5577504644871496,
      "tokens_seen": 1314582528
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030389167502507526,
      "loss": 2.6986,
      "theoretical_loss": 3.557734428328171,
      "tokens_seen": 1314648064
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003038816449348044,
      "loss": 2.808,
      "theoretical_loss": 3.557718393192407,
      "tokens_seen": 1314713600
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003038716148445336,
      "loss": 2.6255,
      "theoretical_loss": 3.5577023590797423,
      "tokens_seen": 1314779136
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003038615847542628,
      "loss": 2.5779,
      "theoretical_loss": 3.5576863259900597,
      "tokens_seen": 1314844672
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.000303851554663992,
      "loss": 2.6617,
      "theoretical_loss": 3.5576702939232434,
      "tokens_seen": 1314910208
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030384152457372117,
      "loss": 2.8716,
      "theoretical_loss": 3.5576542628791765,
      "tokens_seen": 1314975744
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003038314944834504,
      "loss": 2.8262,
      "theoretical_loss": 3.557638232857744,
      "tokens_seen": 1315041280
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030382146439317953,
      "loss": 2.7641,
      "theoretical_loss": 3.5576222038588288,
      "tokens_seen": 1315106816
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030381143430290876,
      "loss": 2.8674,
      "theoretical_loss": 3.557606175882315,
      "tokens_seen": 1315172352
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003038014042126379,
      "loss": 2.8339,
      "theoretical_loss": 3.5575901489280866,
      "tokens_seen": 1315237888
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003037913741223671,
      "loss": 2.6716,
      "theoretical_loss": 3.557574122996027,
      "tokens_seen": 1315303424
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003037813440320963,
      "loss": 2.8439,
      "theoretical_loss": 3.5575580980860204,
      "tokens_seen": 1315368960
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003037713139418255,
      "loss": 2.7056,
      "theoretical_loss": 3.5575420741979507,
      "tokens_seen": 1315434496
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030376128385155467,
      "loss": 2.7797,
      "theoretical_loss": 3.5575260513317017,
      "tokens_seen": 1315500032
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030375125376128385,
      "loss": 2.5237,
      "theoretical_loss": 3.5575100294871573,
      "tokens_seen": 1315565568
    },
    {
      "epoch": 4.04,
      "objective/train/docs_used": 1492577,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.998716354370117,
      "objective/train/theoretical_loss": 3.557494008664202,
      "objective/train/tokens_used": 1336091104,
      "theoretical_loss": 3.557494008664202,
      "tokens_seen": 1315631104
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030374122367101303,
      "loss": 2.6291,
      "theoretical_loss": 3.557494008664202,
      "tokens_seen": 1315631104
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030373119358074227,
      "loss": 2.5316,
      "theoretical_loss": 3.557477988862719,
      "tokens_seen": 1315696640
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003037211634904714,
      "loss": 2.768,
      "theoretical_loss": 3.5574619700825925,
      "tokens_seen": 1315762176
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030371113340020063,
      "loss": 2.7363,
      "theoretical_loss": 3.557445952323707,
      "tokens_seen": 1315827712
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003037011033099298,
      "loss": 2.671,
      "theoretical_loss": 3.557429935585946,
      "tokens_seen": 1315893248
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.000303691073219659,
      "loss": 2.6929,
      "theoretical_loss": 3.5574139198691936,
      "tokens_seen": 1315958784
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030368104312938817,
      "loss": 2.8983,
      "theoretical_loss": 3.5573979051733344,
      "tokens_seen": 1316024320
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030367101303911735,
      "loss": 2.7451,
      "theoretical_loss": 3.5573818914982525,
      "tokens_seen": 1316089856
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030366098294884653,
      "loss": 2.506,
      "theoretical_loss": 3.557365878843831,
      "tokens_seen": 1316155392
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030365095285857577,
      "loss": 2.5843,
      "theoretical_loss": 3.557349867209955,
      "tokens_seen": 1316220928
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003036409227683049,
      "loss": 2.8426,
      "theoretical_loss": 3.5573338565965074,
      "tokens_seen": 1316286464
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030363089267803413,
      "loss": 2.8309,
      "theoretical_loss": 3.5573178470033744,
      "tokens_seen": 1316352000
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030362086258776326,
      "loss": 2.8832,
      "theoretical_loss": 3.5573018384304382,
      "tokens_seen": 1316417536
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003036108324974925,
      "loss": 2.7057,
      "theoretical_loss": 3.5572858308775843,
      "tokens_seen": 1316483072
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003036008024072217,
      "loss": 2.7707,
      "theoretical_loss": 3.5572698243446963,
      "tokens_seen": 1316548608
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030359077231695086,
      "loss": 2.3975,
      "theoretical_loss": 3.5572538188316587,
      "tokens_seen": 1316614144
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030358074222668004,
      "loss": 2.8102,
      "theoretical_loss": 3.5572378143383556,
      "tokens_seen": 1316679680
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003035707121364092,
      "loss": 2.6688,
      "theoretical_loss": 3.5572218108646716,
      "tokens_seen": 1316745216
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003035606820461384,
      "loss": 2.969,
      "theoretical_loss": 3.5572058084104907,
      "tokens_seen": 1316810752
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030355065195586763,
      "loss": 2.5797,
      "theoretical_loss": 3.557189806975697,
      "tokens_seen": 1316876288
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030354062186559676,
      "loss": 2.4623,
      "theoretical_loss": 3.5571738065601757,
      "tokens_seen": 1316941824
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.000303530591775326,
      "loss": 2.8467,
      "theoretical_loss": 3.55715780716381,
      "tokens_seen": 1317007360
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003035205616850552,
      "loss": 2.8461,
      "theoretical_loss": 3.5571418087864854,
      "tokens_seen": 1317072896
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030351053159478436,
      "loss": 2.6863,
      "theoretical_loss": 3.5571258114280853,
      "tokens_seen": 1317138432
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030350050150451354,
      "loss": 2.8625,
      "theoretical_loss": 3.557109815088495,
      "tokens_seen": 1317203968
    },
    {
      "epoch": 4.04,
      "objective/train/docs_used": 1493455,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.698178291320801,
      "objective/train/theoretical_loss": 3.557093819767598,
      "objective/train/tokens_used": 1337729504,
      "theoretical_loss": 3.557093819767598,
      "tokens_seen": 1317269504
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003034904714142427,
      "loss": 2.6329,
      "theoretical_loss": 3.557093819767598,
      "tokens_seen": 1317269504
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003034804413239719,
      "loss": 2.5847,
      "theoretical_loss": 3.5570778254652793,
      "tokens_seen": 1317335040
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030347041123370114,
      "loss": 2.5977,
      "theoretical_loss": 3.5570618321814234,
      "tokens_seen": 1317400576
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030346038114343026,
      "loss": 2.711,
      "theoretical_loss": 3.5570458399159146,
      "tokens_seen": 1317466112
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003034503510531595,
      "loss": 2.7197,
      "theoretical_loss": 3.5570298486686376,
      "tokens_seen": 1317531648
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003034403209628886,
      "loss": 2.5191,
      "theoretical_loss": 3.557013858439477,
      "tokens_seen": 1317597184
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030343029087261786,
      "loss": 2.568,
      "theoretical_loss": 3.556997869228317,
      "tokens_seen": 1317662720
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030342026078234704,
      "loss": 2.6477,
      "theoretical_loss": 3.5569818810350418,
      "tokens_seen": 1317728256
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003034102306920762,
      "loss": 2.7502,
      "theoretical_loss": 3.5569658938595374,
      "tokens_seen": 1317793792
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003034002006018054,
      "loss": 2.7715,
      "theoretical_loss": 3.5569499077016866,
      "tokens_seen": 1317859328
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003033901705115346,
      "loss": 2.7943,
      "theoretical_loss": 3.5569339225613756,
      "tokens_seen": 1317924864
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030338014042126377,
      "loss": 2.6835,
      "theoretical_loss": 3.556917938438488,
      "tokens_seen": 1317990400
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.000303370110330993,
      "loss": 2.6136,
      "theoretical_loss": 3.5569019553329095,
      "tokens_seen": 1318055936
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030336008024072213,
      "loss": 2.7573,
      "theoretical_loss": 3.5568859732445235,
      "tokens_seen": 1318121472
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030335005015045137,
      "loss": 2.6267,
      "theoretical_loss": 3.5568699921732154,
      "tokens_seen": 1318187008
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030334002006018055,
      "loss": 2.652,
      "theoretical_loss": 3.5568540121188703,
      "tokens_seen": 1318252544
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030332998996990973,
      "loss": 2.6998,
      "theoretical_loss": 3.556838033081372,
      "tokens_seen": 1318318080
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003033199598796389,
      "loss": 2.7622,
      "theoretical_loss": 3.5568220550606053,
      "tokens_seen": 1318383616
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003033099297893681,
      "loss": 2.6623,
      "theoretical_loss": 3.5568060780564563,
      "tokens_seen": 1318449152
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003032998996990973,
      "loss": 2.5731,
      "theoretical_loss": 3.556790102068808,
      "tokens_seen": 1318514688
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003032898696088265,
      "loss": 2.7561,
      "theoretical_loss": 3.5567741270975466,
      "tokens_seen": 1318580224
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003032798395185557,
      "loss": 2.7818,
      "theoretical_loss": 3.5567581531425563,
      "tokens_seen": 1318645760
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030326980942828487,
      "loss": 2.5645,
      "theoretical_loss": 3.556742180203722,
      "tokens_seen": 1318711296
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030325977933801405,
      "loss": 2.8654,
      "theoretical_loss": 3.5567262082809283,
      "tokens_seen": 1318776832
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030324974924774323,
      "loss": 2.4856,
      "theoretical_loss": 3.5567102373740607,
      "tokens_seen": 1318842368
    },
    {
      "epoch": 4.04,
      "objective/train/docs_used": 1494825,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.679054021835327,
      "objective/train/theoretical_loss": 3.5566942674830035,
      "objective/train/tokens_used": 1339367904,
      "theoretical_loss": 3.5566942674830035,
      "tokens_seen": 1318907904
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030323971915747247,
      "loss": 2.5742,
      "theoretical_loss": 3.5566942674830035,
      "tokens_seen": 1318907904
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003032296890672016,
      "loss": 2.674,
      "theoretical_loss": 3.5566782986076424,
      "tokens_seen": 1318973440
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030321965897693083,
      "loss": 2.6931,
      "theoretical_loss": 3.5566623307478613,
      "tokens_seen": 1319038976
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030320962888666,
      "loss": 2.5451,
      "theoretical_loss": 3.556646363903546,
      "tokens_seen": 1319104512
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003031995987963892,
      "loss": 2.7607,
      "theoretical_loss": 3.556630398074581,
      "tokens_seen": 1319170048
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030318956870611837,
      "loss": 2.6781,
      "theoretical_loss": 3.556614433260852,
      "tokens_seen": 1319235584
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030317953861584755,
      "loss": 2.9642,
      "theoretical_loss": 3.5565984694622426,
      "tokens_seen": 1319301120
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030316950852557673,
      "loss": 2.8698,
      "theoretical_loss": 3.556582506678639,
      "tokens_seen": 1319366656
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030315947843530597,
      "loss": 2.4901,
      "theoretical_loss": 3.556566544909926,
      "tokens_seen": 1319432192
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003031494483450351,
      "loss": 2.6169,
      "theoretical_loss": 3.556550584155989,
      "tokens_seen": 1319497728
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030313941825476433,
      "loss": 2.7552,
      "theoretical_loss": 3.5565346244167126,
      "tokens_seen": 1319563264
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030312938816449346,
      "loss": 2.5759,
      "theoretical_loss": 3.556518665691982,
      "tokens_seen": 1319628800
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003031193580742227,
      "loss": 2.8975,
      "theoretical_loss": 3.556502707981682,
      "tokens_seen": 1319694336
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003031093279839519,
      "loss": 2.7001,
      "theoretical_loss": 3.556486751285699,
      "tokens_seen": 1319759872
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030309929789368106,
      "loss": 2.5405,
      "theoretical_loss": 3.556470795603916,
      "tokens_seen": 1319825408
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030308926780341024,
      "loss": 2.7051,
      "theoretical_loss": 3.5564548409362207,
      "tokens_seen": 1319890944
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030308926780341024,
      "loss": 2.6966,
      "theoretical_loss": 3.5564388872824964,
      "tokens_seen": 1319956480
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003030792377131394,
      "loss": 2.593,
      "theoretical_loss": 3.556422934642629,
      "tokens_seen": 1320022016
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003030692076228686,
      "loss": 2.5576,
      "theoretical_loss": 3.5564069830165037,
      "tokens_seen": 1320087552
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030305917753259783,
      "loss": 2.7134,
      "theoretical_loss": 3.5563910324040062,
      "tokens_seen": 1320153088
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030304914744232696,
      "loss": 2.7155,
      "theoretical_loss": 3.5563750828050207,
      "tokens_seen": 1320218624
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003030391173520562,
      "loss": 2.4824,
      "theoretical_loss": 3.5563591342194334,
      "tokens_seen": 1320284160
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003030290872617854,
      "loss": 2.7657,
      "theoretical_loss": 3.5563431866471293,
      "tokens_seen": 1320349696
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030301905717151456,
      "loss": 2.6831,
      "theoretical_loss": 3.556327240087994,
      "tokens_seen": 1320415232
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030300902708124374,
      "loss": 2.77,
      "theoretical_loss": 3.5563112945419126,
      "tokens_seen": 1320480768
    },
    {
      "epoch": 4.04,
      "objective/train/docs_used": 1495445,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9724247455596924,
      "objective/train/theoretical_loss": 3.55629535000877,
      "objective/train/tokens_used": 1341006304,
      "theoretical_loss": 3.55629535000877,
      "tokens_seen": 1320546304
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003029989969909729,
      "loss": 2.9013,
      "theoretical_loss": 3.55629535000877,
      "tokens_seen": 1320546304
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003029889669007021,
      "loss": 2.7225,
      "theoretical_loss": 3.5562794064884526,
      "tokens_seen": 1320611840
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030297893681043134,
      "loss": 2.4485,
      "theoretical_loss": 3.5562634639808453,
      "tokens_seen": 1320677376
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030296890672016046,
      "loss": 2.7318,
      "theoretical_loss": 3.556247522485833,
      "tokens_seen": 1320742912
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003029588766298897,
      "loss": 2.7806,
      "theoretical_loss": 3.556231582003303,
      "tokens_seen": 1320808448
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003029488465396188,
      "loss": 2.6104,
      "theoretical_loss": 3.5562156425331377,
      "tokens_seen": 1320873984
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030293881644934806,
      "loss": 2.7759,
      "theoretical_loss": 3.5561997040752256,
      "tokens_seen": 1320939520
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030292878635907724,
      "loss": 2.5906,
      "theoretical_loss": 3.5561837666294505,
      "tokens_seen": 1321005056
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003029187562688064,
      "loss": 2.8147,
      "theoretical_loss": 3.5561678301956983,
      "tokens_seen": 1321070592
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003029087261785356,
      "loss": 2.6234,
      "theoretical_loss": 3.556151894773854,
      "tokens_seen": 1321136128
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003028986960882648,
      "loss": 2.8506,
      "theoretical_loss": 3.5561359603638047,
      "tokens_seen": 1321201664
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030288866599799397,
      "loss": 2.7474,
      "theoretical_loss": 3.5561200269654343,
      "tokens_seen": 1321267200
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003028786359077232,
      "loss": 2.8021,
      "theoretical_loss": 3.5561040945786293,
      "tokens_seen": 1321332736
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030286860581745233,
      "loss": 2.818,
      "theoretical_loss": 3.556088163203275,
      "tokens_seen": 1321398272
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030285857572718157,
      "loss": 2.7076,
      "theoretical_loss": 3.5560722328392576,
      "tokens_seen": 1321463808
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030284854563691075,
      "loss": 2.6369,
      "theoretical_loss": 3.5560563034864616,
      "tokens_seen": 1321529344
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030283851554663993,
      "loss": 2.5183,
      "theoretical_loss": 3.556040375144774,
      "tokens_seen": 1321594880
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003028284854563691,
      "loss": 2.545,
      "theoretical_loss": 3.556024447814079,
      "tokens_seen": 1321660416
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003028184553660983,
      "loss": 2.4651,
      "theoretical_loss": 3.556008521494264,
      "tokens_seen": 1321725952
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030280842527582747,
      "loss": 2.6089,
      "theoretical_loss": 3.555992596185213,
      "tokens_seen": 1321791488
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003027983951855567,
      "loss": 2.6557,
      "theoretical_loss": 3.555976671886813,
      "tokens_seen": 1321857024
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030278836509528583,
      "loss": 2.6763,
      "theoretical_loss": 3.5559607485989497,
      "tokens_seen": 1321922560
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030277833500501507,
      "loss": 2.6129,
      "theoretical_loss": 3.5559448263215083,
      "tokens_seen": 1321988096
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003027683049147442,
      "loss": 2.5999,
      "theoretical_loss": 3.555928905054375,
      "tokens_seen": 1322053632
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030275827482447343,
      "loss": 2.923,
      "theoretical_loss": 3.555912984797435,
      "tokens_seen": 1322119168
    },
    {
      "epoch": 4.04,
      "objective/train/docs_used": 1496938,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7434074878692627,
      "objective/train/theoretical_loss": 3.555897065550575,
      "objective/train/tokens_used": 1342644704,
      "theoretical_loss": 3.555897065550575,
      "tokens_seen": 1322184704
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003027482447342026,
      "loss": 2.7297,
      "theoretical_loss": 3.555897065550575,
      "tokens_seen": 1322184704
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003027382146439318,
      "loss": 2.804,
      "theoretical_loss": 3.5558811473136807,
      "tokens_seen": 1322250240
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.000302728184553661,
      "loss": 2.3758,
      "theoretical_loss": 3.5558652300866367,
      "tokens_seen": 1322315776
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003027181544633902,
      "loss": 2.5547,
      "theoretical_loss": 3.555849313869331,
      "tokens_seen": 1322381312
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030270812437311934,
      "loss": 2.663,
      "theoretical_loss": 3.5558333986616475,
      "tokens_seen": 1322446848
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030269809428284857,
      "loss": 2.5285,
      "theoretical_loss": 3.555817484463474,
      "tokens_seen": 1322512384
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003026880641925777,
      "loss": 2.6559,
      "theoretical_loss": 3.555801571274695,
      "tokens_seen": 1322577920
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030267803410230693,
      "loss": 2.7535,
      "theoretical_loss": 3.5557856590951964,
      "tokens_seen": 1322643456
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003026680040120361,
      "loss": 2.8497,
      "theoretical_loss": 3.5557697479248653,
      "tokens_seen": 1322708992
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003026579739217653,
      "loss": 2.6702,
      "theoretical_loss": 3.555753837763587,
      "tokens_seen": 1322774528
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003026479438314945,
      "loss": 2.6117,
      "theoretical_loss": 3.555737928611248,
      "tokens_seen": 1322840064
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030263791374122366,
      "loss": 2.5529,
      "theoretical_loss": 3.5557220204677336,
      "tokens_seen": 1322905600
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030262788365095284,
      "loss": 2.6863,
      "theoretical_loss": 3.5557061133329304,
      "tokens_seen": 1322971136
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003026178535606821,
      "loss": 2.7612,
      "theoretical_loss": 3.555690207206724,
      "tokens_seen": 1323036672
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003026078234704112,
      "loss": 2.8795,
      "theoretical_loss": 3.5556743020890007,
      "tokens_seen": 1323102208
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030259779338014044,
      "loss": 2.7482,
      "theoretical_loss": 3.5556583979796477,
      "tokens_seen": 1323167744
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030258776328986956,
      "loss": 2.6481,
      "theoretical_loss": 3.5556424948785494,
      "tokens_seen": 1323233280
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003025777331995988,
      "loss": 2.561,
      "theoretical_loss": 3.555626592785593,
      "tokens_seen": 1323298816
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.000302567703109328,
      "loss": 2.6666,
      "theoretical_loss": 3.555610691700664,
      "tokens_seen": 1323364352
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030255767301905716,
      "loss": 2.6272,
      "theoretical_loss": 3.5555947916236494,
      "tokens_seen": 1323429888
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003025476429287864,
      "loss": 2.755,
      "theoretical_loss": 3.5555788925544345,
      "tokens_seen": 1323495424
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003025376128385156,
      "loss": 2.7486,
      "theoretical_loss": 3.5555629944929064,
      "tokens_seen": 1323560960
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030252758274824476,
      "loss": 2.768,
      "theoretical_loss": 3.555547097438951,
      "tokens_seen": 1323626496
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030251755265797394,
      "loss": 2.7912,
      "theoretical_loss": 3.555531201392454,
      "tokens_seen": 1323692032
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003025075225677031,
      "loss": 2.7348,
      "theoretical_loss": 3.5555153063533025,
      "tokens_seen": 1323757568
    },
    {
      "epoch": 4.04,
      "objective/train/docs_used": 1497393,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.852140426635742,
      "objective/train/theoretical_loss": 3.5554994123213826,
      "objective/train/tokens_used": 1344283104,
      "theoretical_loss": 3.5554994123213826,
      "tokens_seen": 1323823104
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003024974924774323,
      "loss": 2.8627,
      "theoretical_loss": 3.5554994123213826,
      "tokens_seen": 1323823104
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030248746238716154,
      "loss": 2.9006,
      "theoretical_loss": 3.55548351929658,
      "tokens_seen": 1323888640
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030247743229689066,
      "loss": 2.5953,
      "theoretical_loss": 3.555467627278782,
      "tokens_seen": 1323954176
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003024674022066199,
      "loss": 2.5436,
      "theoretical_loss": 3.555451736267874,
      "tokens_seen": 1324019712
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030245737211634903,
      "loss": 2.6353,
      "theoretical_loss": 3.5554358462637436,
      "tokens_seen": 1324085248
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030244734202607826,
      "loss": 2.6937,
      "theoretical_loss": 3.5554199572662757,
      "tokens_seen": 1324150784
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030243731193580744,
      "loss": 2.849,
      "theoretical_loss": 3.5554040692753577,
      "tokens_seen": 1324216320
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003024272818455366,
      "loss": 2.7065,
      "theoretical_loss": 3.555388182290876,
      "tokens_seen": 1324281856
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003024172517552658,
      "loss": 2.8507,
      "theoretical_loss": 3.555372296312717,
      "tokens_seen": 1324347392
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.000302407221664995,
      "loss": 2.6622,
      "theoretical_loss": 3.555356411340766,
      "tokens_seen": 1324412928
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030239719157472417,
      "loss": 2.7269,
      "theoretical_loss": 3.555340527374911,
      "tokens_seen": 1324478464
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003023871614844534,
      "loss": 2.6154,
      "theoretical_loss": 3.5553246444150384,
      "tokens_seen": 1324544000
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030237713139418253,
      "loss": 2.501,
      "theoretical_loss": 3.5553087624610336,
      "tokens_seen": 1324609536
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030236710130391177,
      "loss": 2.6562,
      "theoretical_loss": 3.555292881512784,
      "tokens_seen": 1324675072
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030235707121364095,
      "loss": 2.5376,
      "theoretical_loss": 3.5552770015701762,
      "tokens_seen": 1324740608
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030234704112337013,
      "loss": 2.8098,
      "theoretical_loss": 3.5552611226330963,
      "tokens_seen": 1324806144
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003023370110330993,
      "loss": 2.767,
      "theoretical_loss": 3.555245244701431,
      "tokens_seen": 1324871680
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003023269809428285,
      "loss": 2.5934,
      "theoretical_loss": 3.5552293677750675,
      "tokens_seen": 1324937216
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030231695085255767,
      "loss": 2.7604,
      "theoretical_loss": 3.5552134918538916,
      "tokens_seen": 1325002752
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003023069207622869,
      "loss": 2.7105,
      "theoretical_loss": 3.5551976169377904,
      "tokens_seen": 1325068288
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030229689067201603,
      "loss": 2.5927,
      "theoretical_loss": 3.5551817430266506,
      "tokens_seen": 1325133824
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030228686058174527,
      "loss": 2.7396,
      "theoretical_loss": 3.5551658701203586,
      "tokens_seen": 1325199360
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003022768304914744,
      "loss": 2.7238,
      "theoretical_loss": 3.5551499982188015,
      "tokens_seen": 1325264896
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030226680040120363,
      "loss": 2.7457,
      "theoretical_loss": 3.5551341273218657,
      "tokens_seen": 1325330432
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003022567703109328,
      "loss": 2.7258,
      "theoretical_loss": 3.555118257429438,
      "tokens_seen": 1325395968
    },
    {
      "epoch": 4.04,
      "objective/train/docs_used": 1498967,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8756134510040283,
      "objective/train/theoretical_loss": 3.555102388541405,
      "objective/train/tokens_used": 1345921504,
      "theoretical_loss": 3.555102388541405,
      "tokens_seen": 1325461504
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.000302246740220662,
      "loss": 2.9374,
      "theoretical_loss": 3.555102388541405,
      "tokens_seen": 1325461504
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003022367101303912,
      "loss": 2.7265,
      "theoretical_loss": 3.555086520657654,
      "tokens_seen": 1325527040
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003022266800401204,
      "loss": 2.7696,
      "theoretical_loss": 3.5550706537780714,
      "tokens_seen": 1325592576
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030221664994984954,
      "loss": 2.6507,
      "theoretical_loss": 3.5550547879025443,
      "tokens_seen": 1325658112
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030220661985957877,
      "loss": 2.7139,
      "theoretical_loss": 3.555038923030959,
      "tokens_seen": 1325723648
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003021965897693079,
      "loss": 2.8103,
      "theoretical_loss": 3.5550230591632026,
      "tokens_seen": 1325789184
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030218655967903713,
      "loss": 2.656,
      "theoretical_loss": 3.5550071962991616,
      "tokens_seen": 1325854720
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003021765295887663,
      "loss": 2.6585,
      "theoretical_loss": 3.554991334438724,
      "tokens_seen": 1325920256
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003021664994984955,
      "loss": 2.7376,
      "theoretical_loss": 3.554975473581776,
      "tokens_seen": 1325985792
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003021564694082247,
      "loss": 2.6548,
      "theoretical_loss": 3.5549596137282045,
      "tokens_seen": 1326051328
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030214643931795386,
      "loss": 2.7974,
      "theoretical_loss": 3.554943754877896,
      "tokens_seen": 1326116864
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030213640922768304,
      "loss": 2.6282,
      "theoretical_loss": 3.5549278970307383,
      "tokens_seen": 1326182400
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003021263791374123,
      "loss": 2.695,
      "theoretical_loss": 3.554912040186618,
      "tokens_seen": 1326247936
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003021163490471414,
      "loss": 2.7362,
      "theoretical_loss": 3.554896184345422,
      "tokens_seen": 1326313472
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030210631895687064,
      "loss": 2.5531,
      "theoretical_loss": 3.5548803295070375,
      "tokens_seen": 1326379008
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030209628886659976,
      "loss": 2.6168,
      "theoretical_loss": 3.5548644756713514,
      "tokens_seen": 1326444544
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.000302086258776329,
      "loss": 2.9134,
      "theoretical_loss": 3.554848622838251,
      "tokens_seen": 1326510080
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003020762286860582,
      "loss": 2.7424,
      "theoretical_loss": 3.5548327710076233,
      "tokens_seen": 1326575616
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030206619859578736,
      "loss": 2.648,
      "theoretical_loss": 3.5548169201793547,
      "tokens_seen": 1326641152
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030205616850551654,
      "loss": 2.8463,
      "theoretical_loss": 3.5548010703533333,
      "tokens_seen": 1326706688
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003020461384152458,
      "loss": 2.5416,
      "theoretical_loss": 3.5547852215294453,
      "tokens_seen": 1326772224
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003020361083249749,
      "loss": 2.6082,
      "theoretical_loss": 3.554769373707579,
      "tokens_seen": 1326837760
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030202607823470414,
      "loss": 2.6569,
      "theoretical_loss": 3.5547535268876205,
      "tokens_seen": 1326903296
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030201604814443327,
      "loss": 2.6231,
      "theoretical_loss": 3.554737681069458,
      "tokens_seen": 1326968832
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003020060180541625,
      "loss": 2.7764,
      "theoretical_loss": 3.5547218362529773,
      "tokens_seen": 1327034368
    },
    {
      "epoch": 4.04,
      "objective/train/docs_used": 1499694,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.877124547958374,
      "objective/train/theoretical_loss": 3.5547059924380666,
      "objective/train/tokens_used": 1347559904,
      "theoretical_loss": 3.5547059924380666,
      "tokens_seen": 1327099904
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003019959879638917,
      "loss": 2.8643,
      "theoretical_loss": 3.5547059924380666,
      "tokens_seen": 1327099904
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030198595787362087,
      "loss": 2.7271,
      "theoretical_loss": 3.554690149624613,
      "tokens_seen": 1327165440
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030197592778335005,
      "loss": 2.7456,
      "theoretical_loss": 3.554674307812504,
      "tokens_seen": 1327230976
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030196589769307923,
      "loss": 2.6898,
      "theoretical_loss": 3.5546584670016257,
      "tokens_seen": 1327296512
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003019558676028084,
      "loss": 2.5976,
      "theoretical_loss": 3.5546426271918667,
      "tokens_seen": 1327362048
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030194583751253764,
      "loss": 2.7791,
      "theoretical_loss": 3.554626788383114,
      "tokens_seen": 1327427584
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030193580742226677,
      "loss": 2.7436,
      "theoretical_loss": 3.554610950575255,
      "tokens_seen": 1327493120
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.000301925777331996,
      "loss": 2.3796,
      "theoretical_loss": 3.554595113768176,
      "tokens_seen": 1327558656
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030191574724172513,
      "loss": 2.3177,
      "theoretical_loss": 3.554579277961766,
      "tokens_seen": 1327624192
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030190571715145437,
      "loss": 2.7672,
      "theoretical_loss": 3.554563443155911,
      "tokens_seen": 1327689728
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030189568706118355,
      "loss": 2.8033,
      "theoretical_loss": 3.5545476093504993,
      "tokens_seen": 1327755264
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030188565697091273,
      "loss": 2.5326,
      "theoretical_loss": 3.554531776545418,
      "tokens_seen": 1327820800
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003018756268806419,
      "loss": 2.7771,
      "theoretical_loss": 3.554515944740554,
      "tokens_seen": 1327886336
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030186559679037115,
      "loss": 2.9061,
      "theoretical_loss": 3.554500113935796,
      "tokens_seen": 1327951872
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003018555667001003,
      "loss": 2.6614,
      "theoretical_loss": 3.5544842841310302,
      "tokens_seen": 1328017408
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003018455366098295,
      "loss": 2.7331,
      "theoretical_loss": 3.5544684553261447,
      "tokens_seen": 1328082944
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030183550651955864,
      "loss": 2.5361,
      "theoretical_loss": 3.5544526275210266,
      "tokens_seen": 1328148480
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030182547642928787,
      "loss": 2.7823,
      "theoretical_loss": 3.554436800715564,
      "tokens_seen": 1328214016
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030181544633901705,
      "loss": 2.5981,
      "theoretical_loss": 3.5544209749096445,
      "tokens_seen": 1328279552
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030180541624874623,
      "loss": 2.6959,
      "theoretical_loss": 3.554405150103155,
      "tokens_seen": 1328345088
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030179538615847547,
      "loss": 2.7439,
      "theoretical_loss": 3.5543893262959836,
      "tokens_seen": 1328410624
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003017853560682046,
      "loss": 2.6138,
      "theoretical_loss": 3.5543735034880175,
      "tokens_seen": 1328476160
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030177532597793383,
      "loss": 2.5255,
      "theoretical_loss": 3.5543576816791447,
      "tokens_seen": 1328541696
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.000301765295887663,
      "loss": 2.6184,
      "theoretical_loss": 3.554341860869253,
      "tokens_seen": 1328607232
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003017552657973922,
      "loss": 2.557,
      "theoretical_loss": 3.5543260410582294,
      "tokens_seen": 1328672768
    },
    {
      "epoch": 4.04,
      "objective/train/docs_used": 1500360,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.750994920730591,
      "objective/train/theoretical_loss": 3.554310222245962,
      "objective/train/tokens_used": 1349198304,
      "theoretical_loss": 3.554310222245962,
      "tokens_seen": 1328738304
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003017452357071214,
      "loss": 2.6728,
      "theoretical_loss": 3.554310222245962,
      "tokens_seen": 1328738304
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003017352056168506,
      "loss": 2.6421,
      "theoretical_loss": 3.5542944044323384,
      "tokens_seen": 1328803840
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030172517552657974,
      "loss": 2.787,
      "theoretical_loss": 3.5542785876172465,
      "tokens_seen": 1328869376
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030171514543630897,
      "loss": 2.7608,
      "theoretical_loss": 3.5542627718005733,
      "tokens_seen": 1328934912
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003017051153460381,
      "loss": 2.5684,
      "theoretical_loss": 3.5542469569822073,
      "tokens_seen": 1329000448
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030169508525576733,
      "loss": 2.539,
      "theoretical_loss": 3.554231143162036,
      "tokens_seen": 1329065984
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003016850551654965,
      "loss": 2.5904,
      "theoretical_loss": 3.5542153303399475,
      "tokens_seen": 1329131520
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003016750250752257,
      "loss": 2.5785,
      "theoretical_loss": 3.554199518515829,
      "tokens_seen": 1329197056
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003016649949849549,
      "loss": 2.8468,
      "theoretical_loss": 3.554183707689569,
      "tokens_seen": 1329262592
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030165496489468406,
      "loss": 2.7968,
      "theoretical_loss": 3.554167897861055,
      "tokens_seen": 1329328128
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030164493480441324,
      "loss": 2.4569,
      "theoretical_loss": 3.5541520890301745,
      "tokens_seen": 1329393664
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003016349047141425,
      "loss": 2.6965,
      "theoretical_loss": 3.554136281196816,
      "tokens_seen": 1329459200
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003016248746238716,
      "loss": 2.555,
      "theoretical_loss": 3.5541204743608663,
      "tokens_seen": 1329524736
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030161484453360084,
      "loss": 2.7436,
      "theoretical_loss": 3.5541046685222146,
      "tokens_seen": 1329590272
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030160481444332996,
      "loss": 2.76,
      "theoretical_loss": 3.5540888636807484,
      "tokens_seen": 1329655808
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003015947843530592,
      "loss": 2.6591,
      "theoretical_loss": 3.5540730598363552,
      "tokens_seen": 1329721344
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003015847542627884,
      "loss": 2.5531,
      "theoretical_loss": 3.5540572569889237,
      "tokens_seen": 1329786880
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030157472417251756,
      "loss": 2.7457,
      "theoretical_loss": 3.554041455138342,
      "tokens_seen": 1329852416
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030156469408224674,
      "loss": 2.6176,
      "theoretical_loss": 3.5540256542844966,
      "tokens_seen": 1329917952
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.000301554663991976,
      "loss": 2.508,
      "theoretical_loss": 3.554009854427277,
      "tokens_seen": 1329983488
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003015446339017051,
      "loss": 2.8276,
      "theoretical_loss": 3.5539940555665703,
      "tokens_seen": 1330049024
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030153460381143434,
      "loss": 2.7321,
      "theoretical_loss": 3.5539782577022656,
      "tokens_seen": 1330114560
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030152457372116347,
      "loss": 2.5992,
      "theoretical_loss": 3.55396246083425,
      "tokens_seen": 1330180096
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003015145436308927,
      "loss": 2.5204,
      "theoretical_loss": 3.553946664962412,
      "tokens_seen": 1330245632
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003015045135406219,
      "loss": 2.8735,
      "theoretical_loss": 3.5539308700866394,
      "tokens_seen": 1330311168
    },
    {
      "epoch": 4.04,
      "objective/train/docs_used": 1501699,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.85687518119812,
      "objective/train/theoretical_loss": 3.5539150762068212,
      "objective/train/tokens_used": 1350836704,
      "theoretical_loss": 3.5539150762068212,
      "tokens_seen": 1330376704
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030149448345035107,
      "loss": 2.6929,
      "theoretical_loss": 3.5539150762068212,
      "tokens_seen": 1330376704
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030148445336008025,
      "loss": 2.5348,
      "theoretical_loss": 3.5538992833228447,
      "tokens_seen": 1330442240
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030147442326980943,
      "loss": 2.8463,
      "theoretical_loss": 3.553883491434598,
      "tokens_seen": 1330507776
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003014643931795386,
      "loss": 2.6109,
      "theoretical_loss": 3.5538677005419697,
      "tokens_seen": 1330573312
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030145436308926784,
      "loss": 2.8107,
      "theoretical_loss": 3.5538519106448483,
      "tokens_seen": 1330638848
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030144433299899697,
      "loss": 2.5365,
      "theoretical_loss": 3.553836121743121,
      "tokens_seen": 1330704384
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003014343029087262,
      "loss": 2.7058,
      "theoretical_loss": 3.5538203338366765,
      "tokens_seen": 1330769920
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030142427281845533,
      "loss": 2.6364,
      "theoretical_loss": 3.553804546925404,
      "tokens_seen": 1330835456
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030141424272818457,
      "loss": 2.2382,
      "theoretical_loss": 3.5537887610091903,
      "tokens_seen": 1330900992
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030140421263791375,
      "loss": 2.6748,
      "theoretical_loss": 3.553772976087924,
      "tokens_seen": 1330966528
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030139418254764293,
      "loss": 2.7097,
      "theoretical_loss": 3.5537571921614943,
      "tokens_seen": 1331032064
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003013841524573721,
      "loss": 2.6677,
      "theoretical_loss": 3.5537414092297888,
      "tokens_seen": 1331097600
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030137412236710135,
      "loss": 2.5247,
      "theoretical_loss": 3.553725627292696,
      "tokens_seen": 1331163136
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003013640922768305,
      "loss": 2.4289,
      "theoretical_loss": 3.5537098463501042,
      "tokens_seen": 1331228672
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003013540621865597,
      "loss": 2.6032,
      "theoretical_loss": 3.553694066401902,
      "tokens_seen": 1331294208
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030134403209628884,
      "loss": 2.7124,
      "theoretical_loss": 3.553678287447977,
      "tokens_seen": 1331359744
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030133400200601807,
      "loss": 2.6053,
      "theoretical_loss": 3.553662509488219,
      "tokens_seen": 1331425280
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030132397191574725,
      "loss": 2.6833,
      "theoretical_loss": 3.553646732522515,
      "tokens_seen": 1331490816
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030131394182547643,
      "loss": 2.7973,
      "theoretical_loss": 3.5536309565507542,
      "tokens_seen": 1331556352
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003013039117352056,
      "loss": 2.5321,
      "theoretical_loss": 3.5536151815728254,
      "tokens_seen": 1331621888
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003012938816449348,
      "loss": 2.7239,
      "theoretical_loss": 3.5535994075886164,
      "tokens_seen": 1331687424
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.000301283851554664,
      "loss": 2.8171,
      "theoretical_loss": 3.5535836345980165,
      "tokens_seen": 1331752960
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003012738214643932,
      "loss": 2.7051,
      "theoretical_loss": 3.5535678626009126,
      "tokens_seen": 1331818496
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030126379137412234,
      "loss": 2.6428,
      "theoretical_loss": 3.5535520915971954,
      "tokens_seen": 1331884032
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003012537612838516,
      "loss": 2.857,
      "theoretical_loss": 3.5535363215867517,
      "tokens_seen": 1331949568
    },
    {
      "epoch": 4.04,
      "objective/train/docs_used": 1502342,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4824728965759277,
      "objective/train/theoretical_loss": 3.553520552569471,
      "objective/train/tokens_used": 1352475104,
      "theoretical_loss": 3.553520552569471,
      "tokens_seen": 1332015104
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003012437311935807,
      "loss": 2.8718,
      "theoretical_loss": 3.553520552569471,
      "tokens_seen": 1332015104
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030123370110330994,
      "loss": 2.7301,
      "theoretical_loss": 3.5535047845452414,
      "tokens_seen": 1332080640
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003012236710130391,
      "loss": 2.6574,
      "theoretical_loss": 3.553489017513952,
      "tokens_seen": 1332146176
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003012136409227683,
      "loss": 2.6469,
      "theoretical_loss": 3.5534732514754914,
      "tokens_seen": 1332211712
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003012036108324975,
      "loss": 2.8646,
      "theoretical_loss": 3.5534574864297475,
      "tokens_seen": 1332277248
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003011935807422267,
      "loss": 2.8386,
      "theoretical_loss": 3.5534417223766095,
      "tokens_seen": 1332342784
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030118355065195584,
      "loss": 2.7916,
      "theoretical_loss": 3.553425959315967,
      "tokens_seen": 1332408320
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003011735205616851,
      "loss": 2.7231,
      "theoretical_loss": 3.5534101972477066,
      "tokens_seen": 1332473856
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003011634904714142,
      "loss": 2.6288,
      "theoretical_loss": 3.553394436171719,
      "tokens_seen": 1332539392
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030115346038114344,
      "loss": 2.6364,
      "theoretical_loss": 3.553378676087892,
      "tokens_seen": 1332604928
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003011434302908726,
      "loss": 2.5164,
      "theoretical_loss": 3.553362916996114,
      "tokens_seen": 1332670464
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003011334002006018,
      "loss": 2.5531,
      "theoretical_loss": 3.553347158896275,
      "tokens_seen": 1332736000
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.000301123370110331,
      "loss": 2.8268,
      "theoretical_loss": 3.5533314017882622,
      "tokens_seen": 1332801536
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030111334002006016,
      "loss": 2.658,
      "theoretical_loss": 3.553315645671966,
      "tokens_seen": 1332867072
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030110330992978935,
      "loss": 2.6564,
      "theoretical_loss": 3.553299890547274,
      "tokens_seen": 1332932608
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003010932798395186,
      "loss": 2.4595,
      "theoretical_loss": 3.553284136414076,
      "tokens_seen": 1332998144
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003010832497492477,
      "loss": 2.5884,
      "theoretical_loss": 3.5532683832722602,
      "tokens_seen": 1333063680
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030107321965897694,
      "loss": 2.6104,
      "theoretical_loss": 3.553252631121716,
      "tokens_seen": 1333129216
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030106318956870607,
      "loss": 2.8035,
      "theoretical_loss": 3.5532368799623315,
      "tokens_seen": 1333194752
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003010531594784353,
      "loss": 2.641,
      "theoretical_loss": 3.553221129793996,
      "tokens_seen": 1333260288
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030104312938816454,
      "loss": 2.4992,
      "theoretical_loss": 3.553205380616599,
      "tokens_seen": 1333325824
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030103309929789367,
      "loss": 2.7852,
      "theoretical_loss": 3.5531896324300285,
      "tokens_seen": 1333391360
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003010230692076229,
      "loss": 2.8122,
      "theoretical_loss": 3.553173885234174,
      "tokens_seen": 1333456896
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003010130391173521,
      "loss": 2.769,
      "theoretical_loss": 3.553158139028925,
      "tokens_seen": 1333522432
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030100300902708127,
      "loss": 2.7795,
      "theoretical_loss": 3.5531423938141695,
      "tokens_seen": 1333587968
    },
    {
      "epoch": 4.04,
      "objective/train/docs_used": 1503973,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.984467029571533,
      "objective/train/theoretical_loss": 3.553126649589797,
      "objective/train/tokens_used": 1354113504,
      "theoretical_loss": 3.553126649589797,
      "tokens_seen": 1333653504
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030099297893681045,
      "loss": 2.9471,
      "theoretical_loss": 3.553126649589797,
      "tokens_seen": 1333653504
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030098294884653963,
      "loss": 2.8916,
      "theoretical_loss": 3.5531109063556965,
      "tokens_seen": 1333719040
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003009729187562688,
      "loss": 2.3241,
      "theoretical_loss": 3.553095164111757,
      "tokens_seen": 1333784576
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030096288866599804,
      "loss": 2.926,
      "theoretical_loss": 3.553079422857868,
      "tokens_seen": 1333850112
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030095285857572717,
      "loss": 2.8648,
      "theoretical_loss": 3.553063682593918,
      "tokens_seen": 1333915648
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003009428284854564,
      "loss": 2.6997,
      "theoretical_loss": 3.5530479433197963,
      "tokens_seen": 1333981184
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030093279839518553,
      "loss": 2.6958,
      "theoretical_loss": 3.553032205035392,
      "tokens_seen": 1334046720
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030092276830491477,
      "loss": 2.6967,
      "theoretical_loss": 3.553016467740594,
      "tokens_seen": 1334112256
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030091273821464395,
      "loss": 2.5381,
      "theoretical_loss": 3.5530007314352927,
      "tokens_seen": 1334177792
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030090270812437313,
      "loss": 2.7095,
      "theoretical_loss": 3.552984996119376,
      "tokens_seen": 1334243328
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003008926780341023,
      "loss": 2.9267,
      "theoretical_loss": 3.5529692617927333,
      "tokens_seen": 1334308864
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030088264794383155,
      "loss": 2.9963,
      "theoretical_loss": 3.5529535284552543,
      "tokens_seen": 1334374400
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003008726178535607,
      "loss": 2.7728,
      "theoretical_loss": 3.5529377961068276,
      "tokens_seen": 1334439936
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003008625877632899,
      "loss": 2.7666,
      "theoretical_loss": 3.5529220647473427,
      "tokens_seen": 1334505472
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030085255767301904,
      "loss": 2.7262,
      "theoretical_loss": 3.552906334376689,
      "tokens_seen": 1334571008
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030084252758274827,
      "loss": 2.5032,
      "theoretical_loss": 3.552890604994756,
      "tokens_seen": 1334636544
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030083249749247745,
      "loss": 2.8243,
      "theoretical_loss": 3.5528748766014333,
      "tokens_seen": 1334702080
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030082246740220663,
      "loss": 2.6245,
      "theoretical_loss": 3.5528591491966086,
      "tokens_seen": 1334767616
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003008124373119358,
      "loss": 2.6308,
      "theoretical_loss": 3.552843422780173,
      "tokens_seen": 1334833152
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.000300802407221665,
      "loss": 2.6874,
      "theoretical_loss": 3.552827697352015,
      "tokens_seen": 1334898688
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003007923771313942,
      "loss": 2.753,
      "theoretical_loss": 3.5528119729120244,
      "tokens_seen": 1334964224
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003007823470411234,
      "loss": 2.6477,
      "theoretical_loss": 3.5527962494600906,
      "tokens_seen": 1335029760
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030077231695085254,
      "loss": 2.7341,
      "theoretical_loss": 3.552780526996102,
      "tokens_seen": 1335095296
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003007622868605818,
      "loss": 2.6776,
      "theoretical_loss": 3.552764805519949,
      "tokens_seen": 1335160832
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003007522567703109,
      "loss": 2.6165,
      "theoretical_loss": 3.5527490850315218,
      "tokens_seen": 1335226368
    },
    {
      "epoch": 4.04,
      "objective/train/docs_used": 1504757,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.535205125808716,
      "objective/train/theoretical_loss": 3.552733365530708,
      "objective/train/tokens_used": 1355751904,
      "theoretical_loss": 3.552733365530708,
      "tokens_seen": 1335291904
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.00030074222668004014,
      "loss": 2.5726,
      "theoretical_loss": 3.552733365530708,
      "tokens_seen": 1335291904
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003007321965897693,
      "loss": 2.7941,
      "theoretical_loss": 3.5527176470173982,
      "tokens_seen": 1335357440
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003007221664994985,
      "loss": 2.77,
      "theoretical_loss": 3.552701929491482,
      "tokens_seen": 1335422976
    },
    {
      "epoch": 4.04,
      "learning_rate": 0.0003007121364092277,
      "loss": 2.4994,
      "theoretical_loss": 3.5526862129528487,
      "tokens_seen": 1335488512
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0003007021063189569,
      "loss": 2.6934,
      "theoretical_loss": 3.5526704974013876,
      "tokens_seen": 1335554048
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00030069207622868604,
      "loss": 2.7728,
      "theoretical_loss": 3.5526547828369885,
      "tokens_seen": 1335619584
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0003006820461384153,
      "loss": 2.7509,
      "theoretical_loss": 3.5526390692595413,
      "tokens_seen": 1335685120
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0003006720160481444,
      "loss": 2.7018,
      "theoretical_loss": 3.552623356668935,
      "tokens_seen": 1335750656
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00030066198595787364,
      "loss": 2.6671,
      "theoretical_loss": 3.552607645065059,
      "tokens_seen": 1335816192
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0003006519558676028,
      "loss": 2.7556,
      "theoretical_loss": 3.552591934447804,
      "tokens_seen": 1335881728
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.000300641925777332,
      "loss": 2.8379,
      "theoretical_loss": 3.5525762248170585,
      "tokens_seen": 1335947264
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0003006318956870612,
      "loss": 2.5638,
      "theoretical_loss": 3.552560516172713,
      "tokens_seen": 1336012800
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00030062186559679036,
      "loss": 2.7779,
      "theoretical_loss": 3.552544808514657,
      "tokens_seen": 1336078336
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00030061183550651955,
      "loss": 2.6395,
      "theoretical_loss": 3.5525291018427803,
      "tokens_seen": 1336143872
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0003006018054162488,
      "loss": 2.7103,
      "theoretical_loss": 3.5525133961569715,
      "tokens_seen": 1336209408
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0003005917753259779,
      "loss": 2.5746,
      "theoretical_loss": 3.5524976914571216,
      "tokens_seen": 1336274944
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00030058174523570714,
      "loss": 2.5816,
      "theoretical_loss": 3.5524819877431204,
      "tokens_seen": 1336340480
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00030057171514543627,
      "loss": 2.7479,
      "theoretical_loss": 3.552466285014857,
      "tokens_seen": 1336406016
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0003005616850551655,
      "loss": 2.7497,
      "theoretical_loss": 3.5524505832722215,
      "tokens_seen": 1336471552
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0003005516549648947,
      "loss": 2.7455,
      "theoretical_loss": 3.5524348825151035,
      "tokens_seen": 1336537088
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00030054162487462387,
      "loss": 2.5357,
      "theoretical_loss": 3.5524191827433933,
      "tokens_seen": 1336602624
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00030053159478435305,
      "loss": 2.6763,
      "theoretical_loss": 3.5524034839569802,
      "tokens_seen": 1336668160
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0003005215646940823,
      "loss": 2.6913,
      "theoretical_loss": 3.5523877861557542,
      "tokens_seen": 1336733696
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0003005115346038114,
      "loss": 2.693,
      "theoretical_loss": 3.552372089339605,
      "tokens_seen": 1336799232
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00030050150451354065,
      "loss": 2.6484,
      "theoretical_loss": 3.5523563935084232,
      "tokens_seen": 1336864768
    },
    {
      "epoch": 4.05,
      "objective/train/docs_used": 1506128,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.197422742843628,
      "objective/train/theoretical_loss": 3.552340698662098,
      "objective/train/tokens_used": 1357390304,
      "theoretical_loss": 3.552340698662098,
      "tokens_seen": 1336930304
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0003004914744232698,
      "loss": 2.6837,
      "theoretical_loss": 3.552340698662098,
      "tokens_seen": 1336930304
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.000300481444332999,
      "loss": 2.4243,
      "theoretical_loss": 3.5523250048005197,
      "tokens_seen": 1336995840
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0003004714142427282,
      "loss": 2.6001,
      "theoretical_loss": 3.5523093119235782,
      "tokens_seen": 1337061376
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00030046138415245737,
      "loss": 2.7931,
      "theoretical_loss": 3.5522936200311634,
      "tokens_seen": 1337126912
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00030045135406218655,
      "loss": 2.5818,
      "theoretical_loss": 3.552277929123165,
      "tokens_seen": 1337192448
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00030044132397191573,
      "loss": 2.7045,
      "theoretical_loss": 3.5522622391994734,
      "tokens_seen": 1337257984
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0003004312938816449,
      "loss": 2.6828,
      "theoretical_loss": 3.552246550259979,
      "tokens_seen": 1337323520
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00030042126379137415,
      "loss": 2.4758,
      "theoretical_loss": 3.552230862304571,
      "tokens_seen": 1337389056
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0003004112337011033,
      "loss": 2.7366,
      "theoretical_loss": 3.55221517533314,
      "tokens_seen": 1337454592
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0003004012036108325,
      "loss": 2.8237,
      "theoretical_loss": 3.5521994893455755,
      "tokens_seen": 1337520128
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0003003911735205617,
      "loss": 2.7176,
      "theoretical_loss": 3.5521838043417686,
      "tokens_seen": 1337585664
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0003003811434302909,
      "loss": 2.6029,
      "theoretical_loss": 3.552168120321608,
      "tokens_seen": 1337651200
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00030037111334002006,
      "loss": 2.7121,
      "theoretical_loss": 3.5521524372849846,
      "tokens_seen": 1337716736
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00030036108324974924,
      "loss": 2.8973,
      "theoretical_loss": 3.552136755231789,
      "tokens_seen": 1337782272
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0003003510531594784,
      "loss": 2.8893,
      "theoretical_loss": 3.552121074161911,
      "tokens_seen": 1337847808
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00030034102306920765,
      "loss": 2.7755,
      "theoretical_loss": 3.5521053940752405,
      "tokens_seen": 1337913344
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0003003309929789368,
      "loss": 2.8843,
      "theoretical_loss": 3.552089714971668,
      "tokens_seen": 1337978880
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.000300320962888666,
      "loss": 2.6604,
      "theoretical_loss": 3.5520740368510832,
      "tokens_seen": 1338044416
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00030031093279839514,
      "loss": 2.7421,
      "theoretical_loss": 3.552058359713377,
      "tokens_seen": 1338109952
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0003003009027081244,
      "loss": 2.6681,
      "theoretical_loss": 3.5520426835584393,
      "tokens_seen": 1338175488
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0003002908726178536,
      "loss": 2.7552,
      "theoretical_loss": 3.552027008386161,
      "tokens_seen": 1338241024
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00030028084252758274,
      "loss": 2.633,
      "theoretical_loss": 3.552011334196431,
      "tokens_seen": 1338306560
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.000300270812437312,
      "loss": 2.6558,
      "theoretical_loss": 3.5519956609891405,
      "tokens_seen": 1338372096
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0003002607823470411,
      "loss": 2.635,
      "theoretical_loss": 3.55197998876418,
      "tokens_seen": 1338437632
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00030025075225677034,
      "loss": 2.8198,
      "theoretical_loss": 3.5519643175214397,
      "tokens_seen": 1338503168
    },
    {
      "epoch": 4.05,
      "objective/train/docs_used": 1506680,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8979873657226562,
      "objective/train/theoretical_loss": 3.5519486472608097,
      "objective/train/tokens_used": 1359028704,
      "theoretical_loss": 3.5519486472608097,
      "tokens_seen": 1338568704
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0003002407221664995,
      "loss": 3.014,
      "theoretical_loss": 3.5519486472608097,
      "tokens_seen": 1338568704
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0003002306920762287,
      "loss": 2.7011,
      "theoretical_loss": 3.5519329779821804,
      "tokens_seen": 1338634240
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0003002206619859579,
      "loss": 2.7054,
      "theoretical_loss": 3.5519173096854426,
      "tokens_seen": 1338699776
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0003002106318956871,
      "loss": 2.6003,
      "theoretical_loss": 3.551901642370486,
      "tokens_seen": 1338765312
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00030020060180541624,
      "loss": 2.5723,
      "theoretical_loss": 3.5518859760372017,
      "tokens_seen": 1338830848
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0003001905717151455,
      "loss": 2.5742,
      "theoretical_loss": 3.55187031068548,
      "tokens_seen": 1338896384
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0003001805416248746,
      "loss": 2.903,
      "theoretical_loss": 3.5518546463152108,
      "tokens_seen": 1338961920
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00030017051153460384,
      "loss": 2.753,
      "theoretical_loss": 3.5518389829262853,
      "tokens_seen": 1339027456
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.000300160481444333,
      "loss": 2.5019,
      "theoretical_loss": 3.5518233205185936,
      "tokens_seen": 1339092992
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0003001504513540622,
      "loss": 2.7562,
      "theoretical_loss": 3.5518076590920264,
      "tokens_seen": 1339158528
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0003001404212637914,
      "loss": 2.602,
      "theoretical_loss": 3.5517919986464745,
      "tokens_seen": 1339224064
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00030013039117352056,
      "loss": 2.6566,
      "theoretical_loss": 3.551776339181828,
      "tokens_seen": 1339289600
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00030012036108324975,
      "loss": 2.5779,
      "theoretical_loss": 3.551760680697977,
      "tokens_seen": 1339355136
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.000300110330992979,
      "loss": 2.7822,
      "theoretical_loss": 3.5517450231948127,
      "tokens_seen": 1339420672
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0003001003009027081,
      "loss": 2.9438,
      "theoretical_loss": 3.551729366672226,
      "tokens_seen": 1339486208
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00030009027081243734,
      "loss": 2.7497,
      "theoretical_loss": 3.551713711130107,
      "tokens_seen": 1339551744
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00030008024072216647,
      "loss": 2.8555,
      "theoretical_loss": 3.5516980565683465,
      "tokens_seen": 1339617280
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0003000702106318957,
      "loss": 2.4172,
      "theoretical_loss": 3.5516824029868355,
      "tokens_seen": 1339682816
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0003000601805416249,
      "loss": 2.5359,
      "theoretical_loss": 3.551666750385464,
      "tokens_seen": 1339748352
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00030005015045135407,
      "loss": 2.7497,
      "theoretical_loss": 3.5516510987641228,
      "tokens_seen": 1339813888
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00030004012036108325,
      "loss": 2.9051,
      "theoretical_loss": 3.5516354481227035,
      "tokens_seen": 1339879424
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0003000300902708125,
      "loss": 2.5491,
      "theoretical_loss": 3.551619798461095,
      "tokens_seen": 1339944960
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0003000200601805416,
      "loss": 2.9152,
      "theoretical_loss": 3.55160414977919,
      "tokens_seen": 1340010496
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00030001003009027085,
      "loss": 2.7579,
      "theoretical_loss": 3.5515885020768785,
      "tokens_seen": 1340076032
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0003,
      "loss": 2.7868,
      "theoretical_loss": 3.551572855354051,
      "tokens_seen": 1340141568
    },
    {
      "epoch": 4.05,
      "objective/train/docs_used": 1506682,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5435173511505127,
      "objective/train/theoretical_loss": 3.551557209610599,
      "objective/train/tokens_used": 1360667104,
      "theoretical_loss": 3.551557209610599,
      "tokens_seen": 1340207104
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002999899699097292,
      "loss": 2.5745,
      "theoretical_loss": 3.551557209610599,
      "tokens_seen": 1340207104
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002999799398194584,
      "loss": 2.315,
      "theoretical_loss": 3.551541564846412,
      "tokens_seen": 1340272640
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029996990972918757,
      "loss": 2.8512,
      "theoretical_loss": 3.551525921061382,
      "tokens_seen": 1340338176
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029995987963891675,
      "loss": 2.7299,
      "theoretical_loss": 3.5515102782553996,
      "tokens_seen": 1340403712
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029994984954864593,
      "loss": 2.7094,
      "theoretical_loss": 3.5514946364283557,
      "tokens_seen": 1340469248
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002999398194583751,
      "loss": 2.6636,
      "theoretical_loss": 3.5514789955801405,
      "tokens_seen": 1340534784
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029992978936810435,
      "loss": 2.7232,
      "theoretical_loss": 3.551463355710646,
      "tokens_seen": 1340600320
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002999197592778335,
      "loss": 2.6671,
      "theoretical_loss": 3.5514477168197622,
      "tokens_seen": 1340665856
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002999097291875627,
      "loss": 2.8143,
      "theoretical_loss": 3.5514320789073808,
      "tokens_seen": 1340731392
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002998996990972919,
      "loss": 2.4735,
      "theoretical_loss": 3.551416441973392,
      "tokens_seen": 1340796928
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002998896690070211,
      "loss": 2.7152,
      "theoretical_loss": 3.5514008060176874,
      "tokens_seen": 1340862464
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029987963891675026,
      "loss": 2.6733,
      "theoretical_loss": 3.5513851710401574,
      "tokens_seen": 1340928000
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029986960882647944,
      "loss": 2.8356,
      "theoretical_loss": 3.551369537040693,
      "tokens_seen": 1340993536
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002998595787362086,
      "loss": 2.7239,
      "theoretical_loss": 3.5513539040191864,
      "tokens_seen": 1341059072
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029984954864593785,
      "loss": 2.6926,
      "theoretical_loss": 3.5513382719755273,
      "tokens_seen": 1341124608
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.000299839518555667,
      "loss": 2.537,
      "theoretical_loss": 3.551322640909607,
      "tokens_seen": 1341190144
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002998294884653962,
      "loss": 2.826,
      "theoretical_loss": 3.551307010821317,
      "tokens_seen": 1341255680
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029981945837512534,
      "loss": 2.705,
      "theoretical_loss": 3.5512913817105485,
      "tokens_seen": 1341321216
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002998094282848546,
      "loss": 2.6646,
      "theoretical_loss": 3.5512757535771917,
      "tokens_seen": 1341386752
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029979939819458376,
      "loss": 2.6163,
      "theoretical_loss": 3.5512601264211385,
      "tokens_seen": 1341452288
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029978936810431294,
      "loss": 2.6182,
      "theoretical_loss": 3.55124450024228,
      "tokens_seen": 1341517824
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002997793380140421,
      "loss": 2.6714,
      "theoretical_loss": 3.551228875040507,
      "tokens_seen": 1341583360
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002997693079237713,
      "loss": 2.7248,
      "theoretical_loss": 3.551213250815711,
      "tokens_seen": 1341648896
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002997592778335005,
      "loss": 2.832,
      "theoretical_loss": 3.5511976275677837,
      "tokens_seen": 1341714432
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002997492477432297,
      "loss": 2.7048,
      "theoretical_loss": 3.551182005296615,
      "tokens_seen": 1341779968
    },
    {
      "epoch": 4.05,
      "objective/train/docs_used": 1507446,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8023407459259033,
      "objective/train/theoretical_loss": 3.551166384002097,
      "objective/train/tokens_used": 1362305504,
      "theoretical_loss": 3.551166384002097,
      "tokens_seen": 1341845504
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029973921765295885,
      "loss": 2.6677,
      "theoretical_loss": 3.551166384002097,
      "tokens_seen": 1341845504
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002997291875626881,
      "loss": 2.635,
      "theoretical_loss": 3.5511507636841206,
      "tokens_seen": 1341911040
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029971915747241726,
      "loss": 2.8049,
      "theoretical_loss": 3.5511351443425774,
      "tokens_seen": 1341976576
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029970912738214644,
      "loss": 2.6777,
      "theoretical_loss": 3.5511195259773585,
      "tokens_seen": 1342042112
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002996990972918756,
      "loss": 2.8039,
      "theoretical_loss": 3.551103908588355,
      "tokens_seen": 1342107648
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002996890672016048,
      "loss": 2.7576,
      "theoretical_loss": 3.551088292175459,
      "tokens_seen": 1342173184
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.000299679037111334,
      "loss": 2.5265,
      "theoretical_loss": 3.551072676738561,
      "tokens_seen": 1342238720
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002996690070210632,
      "loss": 2.6238,
      "theoretical_loss": 3.551057062277553,
      "tokens_seen": 1342304256
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029965897693079235,
      "loss": 2.8529,
      "theoretical_loss": 3.5510414487923256,
      "tokens_seen": 1342369792
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002996489468405216,
      "loss": 2.7188,
      "theoretical_loss": 3.5510258362827707,
      "tokens_seen": 1342435328
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002996389167502507,
      "loss": 2.469,
      "theoretical_loss": 3.5510102247487794,
      "tokens_seen": 1342500864
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029962888665997995,
      "loss": 2.737,
      "theoretical_loss": 3.5509946141902433,
      "tokens_seen": 1342566400
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029961885656970913,
      "loss": 2.7072,
      "theoretical_loss": 3.550979004607054,
      "tokens_seen": 1342631936
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002996088264794383,
      "loss": 2.8882,
      "theoretical_loss": 3.550963395999103,
      "tokens_seen": 1342697472
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002995987963891675,
      "loss": 2.6852,
      "theoretical_loss": 3.5509477883662814,
      "tokens_seen": 1342763008
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029958876629889667,
      "loss": 2.8004,
      "theoretical_loss": 3.550932181708481,
      "tokens_seen": 1342828544
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029957873620862585,
      "loss": 2.566,
      "theoretical_loss": 3.5509165760255934,
      "tokens_seen": 1342894080
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002995687061183551,
      "loss": 2.4521,
      "theoretical_loss": 3.5509009713175095,
      "tokens_seen": 1342959616
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002995586760280842,
      "loss": 2.9367,
      "theoretical_loss": 3.5508853675841214,
      "tokens_seen": 1343025152
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029954864593781345,
      "loss": 2.634,
      "theoretical_loss": 3.5508697648253205,
      "tokens_seen": 1343090688
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002995386158475427,
      "loss": 2.6086,
      "theoretical_loss": 3.5508541630409987,
      "tokens_seen": 1343156224
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002995285857572718,
      "loss": 2.542,
      "theoretical_loss": 3.5508385622310468,
      "tokens_seen": 1343221760
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029951855566700105,
      "loss": 2.6771,
      "theoretical_loss": 3.550822962395357,
      "tokens_seen": 1343287296
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002995085255767302,
      "loss": 2.6823,
      "theoretical_loss": 3.5508073635338206,
      "tokens_seen": 1343352832
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002994984954864594,
      "loss": 2.4627,
      "theoretical_loss": 3.55079176564633,
      "tokens_seen": 1343418368
    },
    {
      "epoch": 4.05,
      "objective/train/docs_used": 1508187,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9692583084106445,
      "objective/train/theoretical_loss": 3.550776168732776,
      "objective/train/tokens_used": 1363943904,
      "theoretical_loss": 3.550776168732776,
      "tokens_seen": 1343483904
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002994884653961886,
      "loss": 2.5772,
      "theoretical_loss": 3.550776168732776,
      "tokens_seen": 1343483904
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029947843530591777,
      "loss": 2.6239,
      "theoretical_loss": 3.550760572793051,
      "tokens_seen": 1343549440
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029946840521564695,
      "loss": 2.4927,
      "theoretical_loss": 3.5507449778270463,
      "tokens_seen": 1343614976
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029945837512537613,
      "loss": 2.5679,
      "theoretical_loss": 3.5507293838346534,
      "tokens_seen": 1343680512
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002994483450351053,
      "loss": 2.9101,
      "theoretical_loss": 3.550713790815765,
      "tokens_seen": 1343746048
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029943831494483455,
      "loss": 2.6945,
      "theoretical_loss": 3.550698198770271,
      "tokens_seen": 1343811584
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002994282848545637,
      "loss": 2.7543,
      "theoretical_loss": 3.550682607698065,
      "tokens_seen": 1343877120
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002994182547642929,
      "loss": 2.6513,
      "theoretical_loss": 3.5506670175990376,
      "tokens_seen": 1343942656
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002994082246740221,
      "loss": 2.7973,
      "theoretical_loss": 3.5506514284730812,
      "tokens_seen": 1344008192
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002993981945837513,
      "loss": 2.803,
      "theoretical_loss": 3.550635840320088,
      "tokens_seen": 1344073728
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029938816449348046,
      "loss": 2.5209,
      "theoretical_loss": 3.5506202531399493,
      "tokens_seen": 1344139264
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029937813440320964,
      "loss": 2.601,
      "theoretical_loss": 3.5506046669325566,
      "tokens_seen": 1344204800
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002993681043129388,
      "loss": 2.5979,
      "theoretical_loss": 3.5505890816978027,
      "tokens_seen": 1344270336
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029935807422266805,
      "loss": 2.7865,
      "theoretical_loss": 3.550573497435578,
      "tokens_seen": 1344335872
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002993480441323972,
      "loss": 2.6167,
      "theoretical_loss": 3.5505579141457764,
      "tokens_seen": 1344401408
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002993380140421264,
      "loss": 2.5938,
      "theoretical_loss": 3.5505423318282885,
      "tokens_seen": 1344466944
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029932798395185554,
      "loss": 2.6116,
      "theoretical_loss": 3.5505267504830065,
      "tokens_seen": 1344532480
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002993179538615848,
      "loss": 2.6737,
      "theoretical_loss": 3.5505111701098224,
      "tokens_seen": 1344598016
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029930792377131396,
      "loss": 2.6748,
      "theoretical_loss": 3.550495590708628,
      "tokens_seen": 1344663552
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029929789368104314,
      "loss": 2.6336,
      "theoretical_loss": 3.5504800122793156,
      "tokens_seen": 1344729088
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002992878635907723,
      "loss": 2.6307,
      "theoretical_loss": 3.550464434821777,
      "tokens_seen": 1344794624
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002992778335005015,
      "loss": 2.6766,
      "theoretical_loss": 3.5504488583359044,
      "tokens_seen": 1344860160
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002992678034102307,
      "loss": 2.5858,
      "theoretical_loss": 3.55043328282159,
      "tokens_seen": 1344925696
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002992577733199599,
      "loss": 2.8844,
      "theoretical_loss": 3.5504177082787254,
      "tokens_seen": 1344991232
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029924774322968905,
      "loss": 2.7651,
      "theoretical_loss": 3.550402134707203,
      "tokens_seen": 1345056768
    },
    {
      "epoch": 4.05,
      "objective/train/docs_used": 1509426,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.554690361022949,
      "objective/train/theoretical_loss": 3.5503865621069144,
      "objective/train/tokens_used": 1365582304,
      "theoretical_loss": 3.5503865621069144,
      "tokens_seen": 1345122304
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002992377131394183,
      "loss": 2.6236,
      "theoretical_loss": 3.5503865621069144,
      "tokens_seen": 1345122304
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029922768304914746,
      "loss": 2.7252,
      "theoretical_loss": 3.5503709904777523,
      "tokens_seen": 1345187840
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029921765295887664,
      "loss": 2.8637,
      "theoretical_loss": 3.550355419819609,
      "tokens_seen": 1345253376
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002992076228686058,
      "loss": 2.5164,
      "theoretical_loss": 3.550339850132376,
      "tokens_seen": 1345318912
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.000299197592778335,
      "loss": 2.7611,
      "theoretical_loss": 3.550324281415946,
      "tokens_seen": 1345384448
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002991875626880642,
      "loss": 2.8538,
      "theoretical_loss": 3.5503087136702103,
      "tokens_seen": 1345449984
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002991775325977934,
      "loss": 2.5655,
      "theoretical_loss": 3.5502931468950623,
      "tokens_seen": 1345515520
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029916750250752255,
      "loss": 2.8136,
      "theoretical_loss": 3.550277581090394,
      "tokens_seen": 1345581056
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002991574724172518,
      "loss": 2.8745,
      "theoretical_loss": 3.5502620162560965,
      "tokens_seen": 1345646592
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002991474423269809,
      "loss": 2.728,
      "theoretical_loss": 3.550246452392063,
      "tokens_seen": 1345712128
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029913741223671015,
      "loss": 2.765,
      "theoretical_loss": 3.550230889498186,
      "tokens_seen": 1345777664
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029912738214643933,
      "loss": 2.8291,
      "theoretical_loss": 3.550215327574357,
      "tokens_seen": 1345843200
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002991173520561685,
      "loss": 2.4863,
      "theoretical_loss": 3.550199766620469,
      "tokens_seen": 1345908736
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002991073219658977,
      "loss": 2.7966,
      "theoretical_loss": 3.550184206636414,
      "tokens_seen": 1345974272
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029909729187562687,
      "loss": 2.6954,
      "theoretical_loss": 3.550168647622084,
      "tokens_seen": 1346039808
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029908726178535605,
      "loss": 2.8217,
      "theoretical_loss": 3.550153089577372,
      "tokens_seen": 1346105344
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002990772316950853,
      "loss": 2.6436,
      "theoretical_loss": 3.5501375325021707,
      "tokens_seen": 1346170880
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002990672016048144,
      "loss": 2.8473,
      "theoretical_loss": 3.550121976396371,
      "tokens_seen": 1346236416
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029905717151454365,
      "loss": 2.7741,
      "theoretical_loss": 3.5501064212598665,
      "tokens_seen": 1346301952
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029904714142427283,
      "loss": 2.477,
      "theoretical_loss": 3.5500908670925493,
      "tokens_seen": 1346367488
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.000299037111334002,
      "loss": 2.5492,
      "theoretical_loss": 3.550075313894312,
      "tokens_seen": 1346433024
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002990270812437312,
      "loss": 2.7839,
      "theoretical_loss": 3.5500597616650467,
      "tokens_seen": 1346498560
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002990170511534604,
      "loss": 2.6982,
      "theoretical_loss": 3.5500442104046463,
      "tokens_seen": 1346564096
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029900702106318956,
      "loss": 2.6283,
      "theoretical_loss": 3.5500286601130027,
      "tokens_seen": 1346629632
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002989969909729188,
      "loss": 2.6279,
      "theoretical_loss": 3.550013110790009,
      "tokens_seen": 1346695168
    },
    {
      "epoch": 4.05,
      "objective/train/docs_used": 1510013,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7701847553253174,
      "objective/train/theoretical_loss": 3.549997562435558,
      "objective/train/tokens_used": 1367220704,
      "theoretical_loss": 3.549997562435558,
      "tokens_seen": 1346760704
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002989869608826479,
      "loss": 2.6528,
      "theoretical_loss": 3.549997562435558,
      "tokens_seen": 1346760704
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029897693079237715,
      "loss": 2.8473,
      "theoretical_loss": 3.549982015049541,
      "tokens_seen": 1346826240
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002989669007021063,
      "loss": 2.8033,
      "theoretical_loss": 3.549966468631852,
      "tokens_seen": 1346891776
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002989568706118355,
      "loss": 2.8413,
      "theoretical_loss": 3.5499509231823825,
      "tokens_seen": 1346957312
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002989468405215647,
      "loss": 2.7241,
      "theoretical_loss": 3.5499353787010257,
      "tokens_seen": 1347022848
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002989368104312939,
      "loss": 2.5255,
      "theoretical_loss": 3.5499198351876737,
      "tokens_seen": 1347088384
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029892678034102306,
      "loss": 2.782,
      "theoretical_loss": 3.5499042926422195,
      "tokens_seen": 1347153920
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002989167502507523,
      "loss": 2.6674,
      "theoretical_loss": 3.549888751064556,
      "tokens_seen": 1347219456
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002989067201604814,
      "loss": 2.7379,
      "theoretical_loss": 3.5498732104545754,
      "tokens_seen": 1347284992
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029889669007021066,
      "loss": 2.7572,
      "theoretical_loss": 3.5498576708121705,
      "tokens_seen": 1347350528
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002988866599799398,
      "loss": 2.8585,
      "theoretical_loss": 3.549842132137234,
      "tokens_seen": 1347416064
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.000298876629889669,
      "loss": 2.6141,
      "theoretical_loss": 3.5498265944296588,
      "tokens_seen": 1347481600
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002988665997993982,
      "loss": 2.6897,
      "theoretical_loss": 3.549811057689337,
      "tokens_seen": 1347547136
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002988565697091274,
      "loss": 2.9036,
      "theoretical_loss": 3.5497955219161623,
      "tokens_seen": 1347612672
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029884653961885656,
      "loss": 2.8636,
      "theoretical_loss": 3.549779987110027,
      "tokens_seen": 1347678208
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029883650952858574,
      "loss": 2.7638,
      "theoretical_loss": 3.549764453270824,
      "tokens_seen": 1347743744
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002988264794383149,
      "loss": 2.8871,
      "theoretical_loss": 3.5497489203984456,
      "tokens_seen": 1347809280
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029881644934804416,
      "loss": 2.513,
      "theoretical_loss": 3.5497333884927853,
      "tokens_seen": 1347874816
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002988064192577733,
      "loss": 2.6083,
      "theoretical_loss": 3.549717857553736,
      "tokens_seen": 1347940352
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002987963891675025,
      "loss": 2.7726,
      "theoretical_loss": 3.549702327581189,
      "tokens_seen": 1348005888
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002987863590772317,
      "loss": 2.8613,
      "theoretical_loss": 3.5496867985750393,
      "tokens_seen": 1348071424
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002987763289869609,
      "loss": 2.7175,
      "theoretical_loss": 3.5496712705351787,
      "tokens_seen": 1348136960
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002987662988966901,
      "loss": 2.8827,
      "theoretical_loss": 3.5496557434615,
      "tokens_seen": 1348202496
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029875626880641925,
      "loss": 2.7725,
      "theoretical_loss": 3.549640217353897,
      "tokens_seen": 1348268032
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002987462387161485,
      "loss": 2.8003,
      "theoretical_loss": 3.5496246922122614,
      "tokens_seen": 1348333568
    },
    {
      "epoch": 4.05,
      "objective/train/docs_used": 1511229,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.716949939727783,
      "objective/train/theoretical_loss": 3.549609168036487,
      "objective/train/tokens_used": 1368859104,
      "theoretical_loss": 3.549609168036487,
      "tokens_seen": 1348399104
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029873620862587766,
      "loss": 2.5916,
      "theoretical_loss": 3.549609168036487,
      "tokens_seen": 1348399104
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029872617853560684,
      "loss": 2.5237,
      "theoretical_loss": 3.5495936448264667,
      "tokens_seen": 1348464640
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.000298716148445336,
      "loss": 2.7218,
      "theoretical_loss": 3.549578122582093,
      "tokens_seen": 1348530176
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002987061183550652,
      "loss": 2.7952,
      "theoretical_loss": 3.549562601303259,
      "tokens_seen": 1348595712
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002986960882647944,
      "loss": 2.7038,
      "theoretical_loss": 3.5495470809898584,
      "tokens_seen": 1348661248
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002986860581745236,
      "loss": 2.7883,
      "theoretical_loss": 3.5495315616417837,
      "tokens_seen": 1348726784
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029867602808425275,
      "loss": 2.5982,
      "theoretical_loss": 3.549516043258928,
      "tokens_seen": 1348792320
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.000298665997993982,
      "loss": 2.3621,
      "theoretical_loss": 3.549500525841185,
      "tokens_seen": 1348857856
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002986559679037111,
      "loss": 2.8927,
      "theoretical_loss": 3.5494850093884462,
      "tokens_seen": 1348923392
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029864593781344035,
      "loss": 2.95,
      "theoretical_loss": 3.5494694939006064,
      "tokens_seen": 1348988928
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029863590772316953,
      "loss": 2.5967,
      "theoretical_loss": 3.549453979377558,
      "tokens_seen": 1349054464
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002986258776328987,
      "loss": 2.8127,
      "theoretical_loss": 3.549438465819194,
      "tokens_seen": 1349120000
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002986158475426279,
      "loss": 2.6664,
      "theoretical_loss": 3.549422953225408,
      "tokens_seen": 1349185536
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029860581745235707,
      "loss": 2.7774,
      "theoretical_loss": 3.5494074415960926,
      "tokens_seen": 1349251072
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029859578736208625,
      "loss": 2.5631,
      "theoretical_loss": 3.5493919309311415,
      "tokens_seen": 1349316608
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002985857572718155,
      "loss": 2.6669,
      "theoretical_loss": 3.5493764212304475,
      "tokens_seen": 1349382144
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002985757271815446,
      "loss": 2.6485,
      "theoretical_loss": 3.5493609124939045,
      "tokens_seen": 1349447680
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029856569709127385,
      "loss": 2.6763,
      "theoretical_loss": 3.549345404721405,
      "tokens_seen": 1349513216
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029855566700100303,
      "loss": 2.6191,
      "theoretical_loss": 3.549329897912843,
      "tokens_seen": 1349578752
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002985456369107322,
      "loss": 2.8153,
      "theoretical_loss": 3.5493143920681107,
      "tokens_seen": 1349644288
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002985356068204614,
      "loss": 2.9487,
      "theoretical_loss": 3.5492988871871023,
      "tokens_seen": 1349709824
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002985255767301906,
      "loss": 2.7807,
      "theoretical_loss": 3.549283383269711,
      "tokens_seen": 1349775360
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029851554663991976,
      "loss": 2.4485,
      "theoretical_loss": 3.5492678803158295,
      "tokens_seen": 1349840896
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.000298505516549649,
      "loss": 2.6957,
      "theoretical_loss": 3.5492523783253525,
      "tokens_seen": 1349906432
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002984954864593781,
      "loss": 2.695,
      "theoretical_loss": 3.549236877298172,
      "tokens_seen": 1349971968
    },
    {
      "epoch": 4.05,
      "objective/train/docs_used": 1512083,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.1851322650909424,
      "objective/train/theoretical_loss": 3.549221377234182,
      "objective/train/tokens_used": 1370497504,
      "theoretical_loss": 3.549221377234182,
      "tokens_seen": 1350037504
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029848545636910735,
      "loss": 2.5258,
      "theoretical_loss": 3.549221377234182,
      "tokens_seen": 1350037504
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002984754262788365,
      "loss": 2.8254,
      "theoretical_loss": 3.5492058781332756,
      "tokens_seen": 1350103040
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002984653961885657,
      "loss": 2.7243,
      "theoretical_loss": 3.5491903799953466,
      "tokens_seen": 1350168576
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002984553660982949,
      "loss": 2.6613,
      "theoretical_loss": 3.549174882820288,
      "tokens_seen": 1350234112
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002984453360080241,
      "loss": 2.676,
      "theoretical_loss": 3.549159386607993,
      "tokens_seen": 1350299648
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029843530591775326,
      "loss": 2.7808,
      "theoretical_loss": 3.549143891358357,
      "tokens_seen": 1350365184
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002984252758274825,
      "loss": 2.6886,
      "theoretical_loss": 3.5491283970712706,
      "tokens_seen": 1350430720
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002984152457372116,
      "loss": 2.4495,
      "theoretical_loss": 3.549112903746629,
      "tokens_seen": 1350496256
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029840521564694086,
      "loss": 2.6895,
      "theoretical_loss": 3.549097411384326,
      "tokens_seen": 1350561792
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029839518555667,
      "loss": 2.8342,
      "theoretical_loss": 3.5490819199842543,
      "tokens_seen": 1350627328
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002983851554663992,
      "loss": 2.4988,
      "theoretical_loss": 3.5490664295463077,
      "tokens_seen": 1350692864
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002983751253761284,
      "loss": 2.7996,
      "theoretical_loss": 3.549050940070379,
      "tokens_seen": 1350758400
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002983650952858576,
      "loss": 2.7887,
      "theoretical_loss": 3.549035451556364,
      "tokens_seen": 1350823936
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029835506519558676,
      "loss": 2.7817,
      "theoretical_loss": 3.549019964004154,
      "tokens_seen": 1350889472
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029834503510531594,
      "loss": 2.747,
      "theoretical_loss": 3.5490044774136433,
      "tokens_seen": 1350955008
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002983350050150451,
      "loss": 2.7483,
      "theoretical_loss": 3.548988991784726,
      "tokens_seen": 1351020544
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029832497492477436,
      "loss": 2.6019,
      "theoretical_loss": 3.548973507117296,
      "tokens_seen": 1351086080
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002983149448345035,
      "loss": 2.6762,
      "theoretical_loss": 3.5489580234112457,
      "tokens_seen": 1351151616
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002983049147442327,
      "loss": 2.8147,
      "theoretical_loss": 3.54894254066647,
      "tokens_seen": 1351217152
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029829488465396185,
      "loss": 2.6882,
      "theoretical_loss": 3.5489270588828616,
      "tokens_seen": 1351282688
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002982848545636911,
      "loss": 2.5956,
      "theoretical_loss": 3.548911578060315,
      "tokens_seen": 1351348224
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029827482447342026,
      "loss": 2.7254,
      "theoretical_loss": 3.548896098198724,
      "tokens_seen": 1351413760
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029826479438314945,
      "loss": 2.5003,
      "theoretical_loss": 3.5488806192979814,
      "tokens_seen": 1351479296
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002982547642928786,
      "loss": 2.9524,
      "theoretical_loss": 3.5488651413579824,
      "tokens_seen": 1351544832
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029824473420260786,
      "loss": 2.7396,
      "theoretical_loss": 3.5488496643786194,
      "tokens_seen": 1351610368
    },
    {
      "epoch": 4.05,
      "objective/train/docs_used": 1513423,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7058820724487305,
      "objective/train/theoretical_loss": 3.5488341883597867,
      "objective/train/tokens_used": 1372135904,
      "theoretical_loss": 3.5488341883597867,
      "tokens_seen": 1351675904
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.000298234704112337,
      "loss": 2.7002,
      "theoretical_loss": 3.5488341883597867,
      "tokens_seen": 1351675904
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002982246740220662,
      "loss": 2.4233,
      "theoretical_loss": 3.5488187133013787,
      "tokens_seen": 1351741440
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029821464393179535,
      "loss": 2.6596,
      "theoretical_loss": 3.5488032392032887,
      "tokens_seen": 1351806976
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002982046138415246,
      "loss": 2.768,
      "theoretical_loss": 3.54878776606541,
      "tokens_seen": 1351872512
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029819458375125377,
      "loss": 2.6433,
      "theoretical_loss": 3.5487722938876374,
      "tokens_seen": 1351938048
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029818455366098295,
      "loss": 2.6283,
      "theoretical_loss": 3.5487568226698647,
      "tokens_seen": 1352003584
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029817452357071213,
      "loss": 2.8337,
      "theoretical_loss": 3.5487413524119855,
      "tokens_seen": 1352069120
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002981644934804413,
      "loss": 2.7834,
      "theoretical_loss": 3.5487258831138937,
      "tokens_seen": 1352134656
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002981544633901705,
      "loss": 2.5967,
      "theoretical_loss": 3.548710414775483,
      "tokens_seen": 1352200192
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029814443329989973,
      "loss": 2.69,
      "theoretical_loss": 3.5486949473966485,
      "tokens_seen": 1352265728
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029813440320962885,
      "loss": 2.7136,
      "theoretical_loss": 3.548679480977283,
      "tokens_seen": 1352331264
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002981243731193581,
      "loss": 2.4809,
      "theoretical_loss": 3.5486640155172804,
      "tokens_seen": 1352396800
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002981143430290872,
      "loss": 2.668,
      "theoretical_loss": 3.5486485510165355,
      "tokens_seen": 1352462336
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029810431293881645,
      "loss": 2.6052,
      "theoretical_loss": 3.5486330874749425,
      "tokens_seen": 1352527872
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029809428284854563,
      "loss": 2.697,
      "theoretical_loss": 3.5486176248923944,
      "tokens_seen": 1352593408
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002980842527582748,
      "loss": 2.6051,
      "theoretical_loss": 3.548602163268786,
      "tokens_seen": 1352658944
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.000298074222668004,
      "loss": 2.6487,
      "theoretical_loss": 3.5485867026040108,
      "tokens_seen": 1352724480
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029806419257773323,
      "loss": 2.7828,
      "theoretical_loss": 3.5485712428979634,
      "tokens_seen": 1352790016
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029805416248746236,
      "loss": 2.5762,
      "theoretical_loss": 3.5485557841505377,
      "tokens_seen": 1352855552
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002980441323971916,
      "loss": 2.6515,
      "theoretical_loss": 3.548540326361628,
      "tokens_seen": 1352921088
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002980341023069208,
      "loss": 2.8252,
      "theoretical_loss": 3.5485248695311284,
      "tokens_seen": 1352986624
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029802407221664996,
      "loss": 2.6945,
      "theoretical_loss": 3.5485094136589326,
      "tokens_seen": 1353052160
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002980140421263792,
      "loss": 2.8939,
      "theoretical_loss": 3.5484939587449356,
      "tokens_seen": 1353117696
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002980040120361083,
      "loss": 2.7708,
      "theoretical_loss": 3.5484785047890304,
      "tokens_seen": 1353183232
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029799398194583755,
      "loss": 2.4756,
      "theoretical_loss": 3.5484630517911127,
      "tokens_seen": 1353248768
    },
    {
      "epoch": 4.05,
      "objective/train/docs_used": 1514073,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4900929927825928,
      "objective/train/theoretical_loss": 3.548447599751076,
      "objective/train/tokens_used": 1373774304,
      "theoretical_loss": 3.548447599751076,
      "tokens_seen": 1353314304
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002979839518555667,
      "loss": 2.7386,
      "theoretical_loss": 3.548447599751076,
      "tokens_seen": 1353314304
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002979739217652959,
      "loss": 2.6159,
      "theoretical_loss": 3.548432148668814,
      "tokens_seen": 1353379840
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002979638916750251,
      "loss": 2.7703,
      "theoretical_loss": 3.548416698544222,
      "tokens_seen": 1353445376
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002979538615847543,
      "loss": 2.6948,
      "theoretical_loss": 3.5484012493771933,
      "tokens_seen": 1353510912
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029794383149448346,
      "loss": 2.7864,
      "theoretical_loss": 3.5483858011676226,
      "tokens_seen": 1353576448
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002979338014042127,
      "loss": 2.5321,
      "theoretical_loss": 3.548370353915405,
      "tokens_seen": 1353641984
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002979237713139418,
      "loss": 2.614,
      "theoretical_loss": 3.548354907620433,
      "tokens_seen": 1353707520
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029791374122367106,
      "loss": 2.539,
      "theoretical_loss": 3.5483394622826028,
      "tokens_seen": 1353773056
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002979037111334002,
      "loss": 2.6795,
      "theoretical_loss": 3.548324017901807,
      "tokens_seen": 1353838592
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002978936810431294,
      "loss": 2.4953,
      "theoretical_loss": 3.548308574477942,
      "tokens_seen": 1353904128
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002978836509528586,
      "loss": 2.658,
      "theoretical_loss": 3.548293132010901,
      "tokens_seen": 1353969664
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002978736208625878,
      "loss": 2.5627,
      "theoretical_loss": 3.548277690500578,
      "tokens_seen": 1354035200
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029786359077231696,
      "loss": 2.6826,
      "theoretical_loss": 3.548262249946869,
      "tokens_seen": 1354100736
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029785356068204614,
      "loss": 2.5905,
      "theoretical_loss": 3.5482468103496663,
      "tokens_seen": 1354166272
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002978435305917753,
      "loss": 2.7626,
      "theoretical_loss": 3.548231371708866,
      "tokens_seen": 1354231808
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029783350050150456,
      "loss": 2.8008,
      "theoretical_loss": 3.5482159340243618,
      "tokens_seen": 1354297344
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002978234704112337,
      "loss": 2.6798,
      "theoretical_loss": 3.548200497296049,
      "tokens_seen": 1354362880
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002978134403209629,
      "loss": 2.7777,
      "theoretical_loss": 3.548185061523821,
      "tokens_seen": 1354428416
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029780341023069205,
      "loss": 2.4268,
      "theoretical_loss": 3.548169626707573,
      "tokens_seen": 1354493952
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002977933801404213,
      "loss": 2.4828,
      "theoretical_loss": 3.5481541928472,
      "tokens_seen": 1354559488
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029778335005015046,
      "loss": 2.6172,
      "theoretical_loss": 3.5481387599425953,
      "tokens_seen": 1354625024
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029777331995987965,
      "loss": 2.6898,
      "theoretical_loss": 3.5481233279936544,
      "tokens_seen": 1354690560
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029776328986960883,
      "loss": 2.8771,
      "theoretical_loss": 3.5481078970002713,
      "tokens_seen": 1354756096
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029775325977933806,
      "loss": 2.8269,
      "theoretical_loss": 3.5480924669623413,
      "tokens_seen": 1354821632
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002977432296890672,
      "loss": 2.6392,
      "theoretical_loss": 3.548077037879759,
      "tokens_seen": 1354887168
    },
    {
      "epoch": 4.05,
      "objective/train/docs_used": 1515331,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.898995876312256,
      "objective/train/theoretical_loss": 3.548061609752419,
      "objective/train/tokens_used": 1375412704,
      "theoretical_loss": 3.548061609752419,
      "tokens_seen": 1354952704
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002977331995987964,
      "loss": 2.8321,
      "theoretical_loss": 3.548061609752419,
      "tokens_seen": 1354952704
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029772316950852555,
      "loss": 2.7478,
      "theoretical_loss": 3.5480461825802148,
      "tokens_seen": 1355018240
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002977131394182548,
      "loss": 2.7196,
      "theoretical_loss": 3.548030756363042,
      "tokens_seen": 1355083776
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029770310932798397,
      "loss": 2.8133,
      "theoretical_loss": 3.548015331100796,
      "tokens_seen": 1355149312
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029769307923771315,
      "loss": 2.7607,
      "theoretical_loss": 3.5479999067933705,
      "tokens_seen": 1355214848
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029768304914744233,
      "loss": 2.879,
      "theoretical_loss": 3.5479844834406604,
      "tokens_seen": 1355280384
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002976730190571715,
      "loss": 2.8117,
      "theoretical_loss": 3.5479690610425605,
      "tokens_seen": 1355345920
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002976629889669007,
      "loss": 2.6956,
      "theoretical_loss": 3.547953639598966,
      "tokens_seen": 1355411456
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029765295887662993,
      "loss": 2.6193,
      "theoretical_loss": 3.5479382191097706,
      "tokens_seen": 1355476992
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029764292878635905,
      "loss": 2.5348,
      "theoretical_loss": 3.5479227995748706,
      "tokens_seen": 1355542528
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002976328986960883,
      "loss": 2.741,
      "theoretical_loss": 3.5479073809941593,
      "tokens_seen": 1355608064
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002976228686058174,
      "loss": 2.6227,
      "theoretical_loss": 3.5478919633675328,
      "tokens_seen": 1355673600
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029761283851554665,
      "loss": 2.7048,
      "theoretical_loss": 3.5478765466948854,
      "tokens_seen": 1355739136
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029760280842527583,
      "loss": 2.6778,
      "theoretical_loss": 3.5478611309761114,
      "tokens_seen": 1355804672
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.000297592778335005,
      "loss": 2.7497,
      "theoretical_loss": 3.5478457162111066,
      "tokens_seen": 1355870208
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002975827482447342,
      "loss": 2.9066,
      "theoretical_loss": 3.547830302399765,
      "tokens_seen": 1355935744
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029757271815446343,
      "loss": 2.7068,
      "theoretical_loss": 3.547814889541983,
      "tokens_seen": 1356001280
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029756268806419256,
      "loss": 2.4788,
      "theoretical_loss": 3.547799477637654,
      "tokens_seen": 1356066816
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002975526579739218,
      "loss": 2.5697,
      "theoretical_loss": 3.547784066686673,
      "tokens_seen": 1356132352
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002975426278836509,
      "loss": 2.4808,
      "theoretical_loss": 3.547768656688936,
      "tokens_seen": 1356197888
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029753259779338016,
      "loss": 2.8969,
      "theoretical_loss": 3.5477532476443376,
      "tokens_seen": 1356263424
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029752256770310934,
      "loss": 2.5684,
      "theoretical_loss": 3.5477378395527723,
      "tokens_seen": 1356328960
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002975125376128385,
      "loss": 2.6905,
      "theoretical_loss": 3.5477224324141354,
      "tokens_seen": 1356394496
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002975025075225677,
      "loss": 2.6665,
      "theoretical_loss": 3.547707026228322,
      "tokens_seen": 1356460032
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002974924774322969,
      "loss": 2.6812,
      "theoretical_loss": 3.547691620995227,
      "tokens_seen": 1356525568
    },
    {
      "epoch": 4.05,
      "objective/train/docs_used": 1516059,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.887085437774658,
      "objective/train/theoretical_loss": 3.5476762167147458,
      "objective/train/tokens_used": 1377051104,
      "theoretical_loss": 3.5476762167147458,
      "tokens_seen": 1356591104
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029748244734202606,
      "loss": 2.7841,
      "theoretical_loss": 3.5476762167147458,
      "tokens_seen": 1356591104
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002974724172517553,
      "loss": 2.8691,
      "theoretical_loss": 3.5476608133867735,
      "tokens_seen": 1356656640
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002974623871614844,
      "loss": 2.8379,
      "theoretical_loss": 3.547645411011205,
      "tokens_seen": 1356722176
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029745235707121366,
      "loss": 2.6562,
      "theoretical_loss": 3.5476300095879347,
      "tokens_seen": 1356787712
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002974423269809428,
      "loss": 2.8677,
      "theoretical_loss": 3.5476146091168586,
      "tokens_seen": 1356853248
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.000297432296890672,
      "loss": 2.5165,
      "theoretical_loss": 3.5475992095978715,
      "tokens_seen": 1356918784
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002974222668004012,
      "loss": 2.6703,
      "theoretical_loss": 3.547583811030869,
      "tokens_seen": 1356984320
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002974122367101304,
      "loss": 2.6626,
      "theoretical_loss": 3.5475684134157457,
      "tokens_seen": 1357049856
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029740220661985956,
      "loss": 2.6634,
      "theoretical_loss": 3.5475530167523974,
      "tokens_seen": 1357115392
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002973921765295888,
      "loss": 2.6589,
      "theoretical_loss": 3.5475376210407186,
      "tokens_seen": 1357180928
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002973821464393179,
      "loss": 2.7632,
      "theoretical_loss": 3.5475222262806048,
      "tokens_seen": 1357246464
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029737211634904716,
      "loss": 2.8114,
      "theoretical_loss": 3.5475068324719516,
      "tokens_seen": 1357312000
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002973620862587763,
      "loss": 2.6815,
      "theoretical_loss": 3.547491439614654,
      "tokens_seen": 1357377536
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002973520561685055,
      "loss": 2.6894,
      "theoretical_loss": 3.5474760477086074,
      "tokens_seen": 1357443072
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002973420260782347,
      "loss": 2.8562,
      "theoretical_loss": 3.5474606567537066,
      "tokens_seen": 1357508608
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002973319959879639,
      "loss": 2.5874,
      "theoretical_loss": 3.5474452667498477,
      "tokens_seen": 1357574144
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029732196589769307,
      "loss": 2.5583,
      "theoretical_loss": 3.547429877696925,
      "tokens_seen": 1357639680
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029731193580742225,
      "loss": 2.5643,
      "theoretical_loss": 3.5474144895948347,
      "tokens_seen": 1357705216
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029730190571715143,
      "loss": 2.6758,
      "theoretical_loss": 3.547399102443472,
      "tokens_seen": 1357770752
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029729187562688067,
      "loss": 2.7373,
      "theoretical_loss": 3.547383716242732,
      "tokens_seen": 1357836288
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029728184553660985,
      "loss": 2.7633,
      "theoretical_loss": 3.5473683309925104,
      "tokens_seen": 1357901824
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029727181544633903,
      "loss": 2.643,
      "theoretical_loss": 3.547352946692703,
      "tokens_seen": 1357967360
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029726178535606826,
      "loss": 2.4956,
      "theoretical_loss": 3.5473375633432043,
      "tokens_seen": 1358032896
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002972517552657974,
      "loss": 2.7176,
      "theoretical_loss": 3.54732218094391,
      "tokens_seen": 1358098432
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002972417251755266,
      "loss": 2.9503,
      "theoretical_loss": 3.5473067994947156,
      "tokens_seen": 1358163968
    },
    {
      "epoch": 4.05,
      "objective/train/docs_used": 1517513,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5391719341278076,
      "objective/train/theoretical_loss": 3.547291418995517,
      "objective/train/tokens_used": 1378689504,
      "theoretical_loss": 3.547291418995517,
      "tokens_seen": 1358229504
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029723169508525575,
      "loss": 2.6743,
      "theoretical_loss": 3.547291418995517,
      "tokens_seen": 1358229504
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.000297221664994985,
      "loss": 2.8176,
      "theoretical_loss": 3.547276039446209,
      "tokens_seen": 1358295040
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029721163490471417,
      "loss": 2.7804,
      "theoretical_loss": 3.547260660846688,
      "tokens_seen": 1358360576
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029720160481444335,
      "loss": 2.6297,
      "theoretical_loss": 3.5472452831968484,
      "tokens_seen": 1358426112
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029719157472417253,
      "loss": 2.7697,
      "theoretical_loss": 3.547229906496587,
      "tokens_seen": 1358491648
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002971815446339017,
      "loss": 2.6438,
      "theoretical_loss": 3.547214530745798,
      "tokens_seen": 1358557184
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002971715145436309,
      "loss": 2.8875,
      "theoretical_loss": 3.5471991559443783,
      "tokens_seen": 1358622720
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029716148445336013,
      "loss": 2.5986,
      "theoretical_loss": 3.5471837820922225,
      "tokens_seen": 1358688256
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029715145436308926,
      "loss": 2.7038,
      "theoretical_loss": 3.5471684091892266,
      "tokens_seen": 1358753792
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002971414242728185,
      "loss": 2.7325,
      "theoretical_loss": 3.5471530372352866,
      "tokens_seen": 1358819328
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002971313941825476,
      "loss": 2.5486,
      "theoretical_loss": 3.5471376662302974,
      "tokens_seen": 1358884864
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029712136409227685,
      "loss": 2.9332,
      "theoretical_loss": 3.547122296174155,
      "tokens_seen": 1358950400
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029711133400200603,
      "loss": 2.789,
      "theoretical_loss": 3.5471069270667552,
      "tokens_seen": 1359015936
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002971013039117352,
      "loss": 2.6935,
      "theoretical_loss": 3.5470915589079937,
      "tokens_seen": 1359081472
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002970912738214644,
      "loss": 2.6713,
      "theoretical_loss": 3.547076191697766,
      "tokens_seen": 1359147008
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029708124373119363,
      "loss": 2.8167,
      "theoretical_loss": 3.5470608254359672,
      "tokens_seen": 1359212544
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029707121364092276,
      "loss": 2.4853,
      "theoretical_loss": 3.5470454601224946,
      "tokens_seen": 1359278080
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.000297061183550652,
      "loss": 2.7376,
      "theoretical_loss": 3.5470300957572425,
      "tokens_seen": 1359343616
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002970511534603811,
      "loss": 2.7228,
      "theoretical_loss": 3.5470147323401076,
      "tokens_seen": 1359409152
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029704112337011036,
      "loss": 2.958,
      "theoretical_loss": 3.546999369870985,
      "tokens_seen": 1359474688
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029703109327983954,
      "loss": 2.7322,
      "theoretical_loss": 3.5469840083497717,
      "tokens_seen": 1359540224
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002970210631895687,
      "loss": 2.7878,
      "theoretical_loss": 3.546968647776362,
      "tokens_seen": 1359605760
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002970110330992979,
      "loss": 2.3531,
      "theoretical_loss": 3.5469532881506525,
      "tokens_seen": 1359671296
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002970010030090271,
      "loss": 2.8062,
      "theoretical_loss": 3.5469379294725387,
      "tokens_seen": 1359736832
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029699097291875626,
      "loss": 2.582,
      "theoretical_loss": 3.5469225717419173,
      "tokens_seen": 1359802368
    },
    {
      "epoch": 4.05,
      "objective/train/docs_used": 1518071,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.3069539070129395,
      "objective/train/theoretical_loss": 3.5469072149586833,
      "objective/train/tokens_used": 1380327904,
      "theoretical_loss": 3.5469072149586833,
      "tokens_seen": 1359867904
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002969809428284855,
      "loss": 2.5244,
      "theoretical_loss": 3.5469072149586833,
      "tokens_seen": 1359867904
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002969709127382146,
      "loss": 2.8248,
      "theoretical_loss": 3.5468918591227334,
      "tokens_seen": 1359933440
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029696088264794386,
      "loss": 2.7735,
      "theoretical_loss": 3.5468765042339623,
      "tokens_seen": 1359998976
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.000296950852557673,
      "loss": 2.522,
      "theoretical_loss": 3.5468611502922673,
      "tokens_seen": 1360064512
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002969408224674022,
      "loss": 2.7694,
      "theoretical_loss": 3.546845797297544,
      "tokens_seen": 1360130048
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002969408224674022,
      "loss": 2.6656,
      "theoretical_loss": 3.5468304452496873,
      "tokens_seen": 1360195584
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002969307923771314,
      "loss": 2.591,
      "theoretical_loss": 3.5468150941485943,
      "tokens_seen": 1360261120
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002969207622868606,
      "loss": 2.7995,
      "theoretical_loss": 3.5467997439941605,
      "tokens_seen": 1360326656
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029691073219658976,
      "loss": 2.8339,
      "theoretical_loss": 3.5467843947862825,
      "tokens_seen": 1360392192
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.000296900702106319,
      "loss": 2.5979,
      "theoretical_loss": 3.5467690465248563,
      "tokens_seen": 1360457728
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002968906720160481,
      "loss": 2.6479,
      "theoretical_loss": 3.546753699209777,
      "tokens_seen": 1360523264
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029688064192577736,
      "loss": 2.5995,
      "theoretical_loss": 3.5467383528409417,
      "tokens_seen": 1360588800
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002968706118355065,
      "loss": 2.6558,
      "theoretical_loss": 3.5467230074182456,
      "tokens_seen": 1360654336
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002968605817452357,
      "loss": 2.7648,
      "theoretical_loss": 3.5467076629415857,
      "tokens_seen": 1360719872
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002968505516549649,
      "loss": 2.6445,
      "theoretical_loss": 3.5466923194108575,
      "tokens_seen": 1360785408
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002968405215646941,
      "loss": 2.3767,
      "theoretical_loss": 3.546676976825957,
      "tokens_seen": 1360850944
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029683049147442327,
      "loss": 2.5866,
      "theoretical_loss": 3.5466616351867812,
      "tokens_seen": 1360916480
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029682046138415245,
      "loss": 2.742,
      "theoretical_loss": 3.5466462944932253,
      "tokens_seen": 1360982016
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029681043129388163,
      "loss": 2.731,
      "theoretical_loss": 3.5466309547451864,
      "tokens_seen": 1361047552
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029680040120361087,
      "loss": 2.913,
      "theoretical_loss": 3.54661561594256,
      "tokens_seen": 1361113088
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029679037111334,
      "loss": 2.7811,
      "theoretical_loss": 3.5466002780852426,
      "tokens_seen": 1361178624
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029678034102306923,
      "loss": 2.7598,
      "theoretical_loss": 3.5465849411731307,
      "tokens_seen": 1361244160
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002967703109327984,
      "loss": 2.6426,
      "theoretical_loss": 3.5465696052061197,
      "tokens_seen": 1361309696
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002967602808425276,
      "loss": 2.5915,
      "theoretical_loss": 3.5465542701841066,
      "tokens_seen": 1361375232
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029675025075225677,
      "loss": 2.855,
      "theoretical_loss": 3.5465389361069874,
      "tokens_seen": 1361440768
    },
    {
      "epoch": 4.05,
      "objective/train/docs_used": 1519309,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.060181140899658,
      "objective/train/theoretical_loss": 3.5465236029746583,
      "objective/train/tokens_used": 1381966304,
      "theoretical_loss": 3.5465236029746583,
      "tokens_seen": 1361506304
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029674022066198595,
      "loss": 2.7874,
      "theoretical_loss": 3.5465236029746583,
      "tokens_seen": 1361506304
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029673019057171513,
      "loss": 2.7999,
      "theoretical_loss": 3.546508270787016,
      "tokens_seen": 1361571840
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029672016048144437,
      "loss": 2.5814,
      "theoretical_loss": 3.546492939543957,
      "tokens_seen": 1361637376
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002967101303911735,
      "loss": 2.6692,
      "theoretical_loss": 3.5464776092453767,
      "tokens_seen": 1361702912
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029670010030090273,
      "loss": 2.9344,
      "theoretical_loss": 3.5464622798911725,
      "tokens_seen": 1361768448
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029669007021063186,
      "loss": 2.7264,
      "theoretical_loss": 3.5464469514812396,
      "tokens_seen": 1361833984
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002966800401203611,
      "loss": 2.8269,
      "theoretical_loss": 3.546431624015476,
      "tokens_seen": 1361899520
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002966700100300903,
      "loss": 2.7462,
      "theoretical_loss": 3.5464162974937765,
      "tokens_seen": 1361965056
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029665997993981946,
      "loss": 2.7305,
      "theoretical_loss": 3.5464009719160385,
      "tokens_seen": 1362030592
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029664994984954864,
      "loss": 2.5155,
      "theoretical_loss": 3.546385647282158,
      "tokens_seen": 1362096128
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002966399197592778,
      "loss": 2.6736,
      "theoretical_loss": 3.546370323592032,
      "tokens_seen": 1362161664
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.000296629889669007,
      "loss": 2.7749,
      "theoretical_loss": 3.5463550008455567,
      "tokens_seen": 1362227200
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029661985957873623,
      "loss": 2.802,
      "theoretical_loss": 3.546339679042628,
      "tokens_seen": 1362292736
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029660982948846536,
      "loss": 2.684,
      "theoretical_loss": 3.546324358183143,
      "tokens_seen": 1362358272
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002965997993981946,
      "loss": 2.5993,
      "theoretical_loss": 3.5463090382669984,
      "tokens_seen": 1362423808
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002965897693079238,
      "loss": 2.6696,
      "theoretical_loss": 3.5462937192940904,
      "tokens_seen": 1362489344
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029657973921765296,
      "loss": 2.8625,
      "theoretical_loss": 3.546278401264316,
      "tokens_seen": 1362554880
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029656970912738214,
      "loss": 2.7273,
      "theoretical_loss": 3.5462630841775713,
      "tokens_seen": 1362620416
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002965596790371113,
      "loss": 2.6861,
      "theoretical_loss": 3.546247768033753,
      "tokens_seen": 1362685952
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002965496489468405,
      "loss": 2.6665,
      "theoretical_loss": 3.546232452832758,
      "tokens_seen": 1362751488
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029653961885656974,
      "loss": 2.6598,
      "theoretical_loss": 3.546217138574482,
      "tokens_seen": 1362817024
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002965295887662989,
      "loss": 2.8421,
      "theoretical_loss": 3.546201825258823,
      "tokens_seen": 1362882560
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002965195586760281,
      "loss": 2.7694,
      "theoretical_loss": 3.546186512885676,
      "tokens_seen": 1362948096
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002965095285857573,
      "loss": 2.7711,
      "theoretical_loss": 3.5461712014549396,
      "tokens_seen": 1363013632
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029649949849548646,
      "loss": 2.6412,
      "theoretical_loss": 3.546155890966509,
      "tokens_seen": 1363079168
    },
    {
      "epoch": 4.05,
      "objective/train/docs_used": 1519912,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8462095260620117,
      "objective/train/theoretical_loss": 3.546140581420282,
      "objective/train/tokens_used": 1383604704,
      "theoretical_loss": 3.546140581420282,
      "tokens_seen": 1363144704
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002964894684052157,
      "loss": 2.7918,
      "theoretical_loss": 3.546140581420282,
      "tokens_seen": 1363144704
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002964794383149448,
      "loss": 2.6971,
      "theoretical_loss": 3.5461252728161545,
      "tokens_seen": 1363210240
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029646940822467406,
      "loss": 2.6667,
      "theoretical_loss": 3.5461099651540233,
      "tokens_seen": 1363275776
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002964593781344032,
      "loss": 2.8574,
      "theoretical_loss": 3.546094658433786,
      "tokens_seen": 1363341312
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002964493480441324,
      "loss": 2.8651,
      "theoretical_loss": 3.5460793526553376,
      "tokens_seen": 1363406848
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002964393179538616,
      "loss": 2.6643,
      "theoretical_loss": 3.546064047818577,
      "tokens_seen": 1363472384
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002964292878635908,
      "loss": 2.831,
      "theoretical_loss": 3.5460487439233996,
      "tokens_seen": 1363537920
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029641925777331996,
      "loss": 2.587,
      "theoretical_loss": 3.546033440969703,
      "tokens_seen": 1363603456
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002964092276830492,
      "loss": 2.6461,
      "theoretical_loss": 3.546018138957383,
      "tokens_seen": 1363668992
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002963991975927783,
      "loss": 2.7813,
      "theoretical_loss": 3.5460028378863377,
      "tokens_seen": 1363734528
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029638916750250756,
      "loss": 2.617,
      "theoretical_loss": 3.5459875377564636,
      "tokens_seen": 1363800064
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002963791374122367,
      "loss": 2.8451,
      "theoretical_loss": 3.5459722385676566,
      "tokens_seen": 1363865600
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002963691073219659,
      "loss": 2.7216,
      "theoretical_loss": 3.5459569403198152,
      "tokens_seen": 1363931136
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002963590772316951,
      "loss": 2.6646,
      "theoretical_loss": 3.545941643012835,
      "tokens_seen": 1363996672
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002963490471414243,
      "loss": 2.5081,
      "theoretical_loss": 3.5459263466466133,
      "tokens_seen": 1364062208
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029633901705115347,
      "loss": 2.8276,
      "theoretical_loss": 3.5459110512210477,
      "tokens_seen": 1364127744
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029632898696088265,
      "loss": 2.5364,
      "theoretical_loss": 3.5458957567360345,
      "tokens_seen": 1364193280
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029631895687061183,
      "loss": 2.8552,
      "theoretical_loss": 3.5458804631914704,
      "tokens_seen": 1364258816
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029630892678034107,
      "loss": 2.7965,
      "theoretical_loss": 3.5458651705872537,
      "tokens_seen": 1364324352
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002962988966900702,
      "loss": 2.8659,
      "theoretical_loss": 3.54584987892328,
      "tokens_seen": 1364389888
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029628886659979943,
      "loss": 2.7637,
      "theoretical_loss": 3.5458345881994466,
      "tokens_seen": 1364455424
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002962788365095286,
      "loss": 2.7808,
      "theoretical_loss": 3.5458192984156516,
      "tokens_seen": 1364520960
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002962688064192578,
      "loss": 2.6831,
      "theoretical_loss": 3.545804009571791,
      "tokens_seen": 1364586496
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029625877632898697,
      "loss": 2.7481,
      "theoretical_loss": 3.545788721667762,
      "tokens_seen": 1364652032
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029624874623871615,
      "loss": 2.6898,
      "theoretical_loss": 3.545773434703462,
      "tokens_seen": 1364717568
    },
    {
      "epoch": 4.05,
      "objective/train/docs_used": 1521305,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.574596881866455,
      "objective/train/theoretical_loss": 3.5457581486787877,
      "objective/train/tokens_used": 1385243104,
      "theoretical_loss": 3.5457581486787877,
      "tokens_seen": 1364783104
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029623871614844533,
      "loss": 2.6539,
      "theoretical_loss": 3.5457581486787877,
      "tokens_seen": 1364783104
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029622868605817457,
      "loss": 2.8171,
      "theoretical_loss": 3.545742863593637,
      "tokens_seen": 1364848640
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002962186559679037,
      "loss": 2.6755,
      "theoretical_loss": 3.5457275794479064,
      "tokens_seen": 1364914176
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029620862587763293,
      "loss": 2.5866,
      "theoretical_loss": 3.5457122962414935,
      "tokens_seen": 1364979712
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029619859578736206,
      "loss": 2.6154,
      "theoretical_loss": 3.5456970139742947,
      "tokens_seen": 1365045248
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002961885656970913,
      "loss": 2.6771,
      "theoretical_loss": 3.545681732646208,
      "tokens_seen": 1365110784
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002961785356068205,
      "loss": 2.8346,
      "theoretical_loss": 3.54566645225713,
      "tokens_seen": 1365176320
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029616850551654966,
      "loss": 2.6756,
      "theoretical_loss": 3.5456511728069584,
      "tokens_seen": 1365241856
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029615847542627884,
      "loss": 2.7741,
      "theoretical_loss": 3.5456358942955903,
      "tokens_seen": 1365307392
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.000296148445336008,
      "loss": 2.8655,
      "theoretical_loss": 3.545620616722923,
      "tokens_seen": 1365372928
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002961384152457372,
      "loss": 2.7293,
      "theoretical_loss": 3.5456053400888536,
      "tokens_seen": 1365438464
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029612838515546643,
      "loss": 2.7781,
      "theoretical_loss": 3.54559006439328,
      "tokens_seen": 1365504000
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029611835506519556,
      "loss": 2.8338,
      "theoretical_loss": 3.545574789636098,
      "tokens_seen": 1365569536
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002961083249749248,
      "loss": 2.719,
      "theoretical_loss": 3.545559515817207,
      "tokens_seen": 1365635072
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.000296098294884654,
      "loss": 2.5393,
      "theoretical_loss": 3.5455442429365025,
      "tokens_seen": 1365700608
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029608826479438316,
      "loss": 2.82,
      "theoretical_loss": 3.545528970993883,
      "tokens_seen": 1365766144
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029607823470411234,
      "loss": 2.5854,
      "theoretical_loss": 3.545513699989246,
      "tokens_seen": 1365831680
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002960682046138415,
      "loss": 2.6055,
      "theoretical_loss": 3.545498429922487,
      "tokens_seen": 1365897216
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002960581745235707,
      "loss": 2.7797,
      "theoretical_loss": 3.545483160793506,
      "tokens_seen": 1365962752
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029604814443329994,
      "loss": 2.7801,
      "theoretical_loss": 3.5454678926021987,
      "tokens_seen": 1366028288
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029603811434302906,
      "loss": 2.6195,
      "theoretical_loss": 3.5454526253484633,
      "tokens_seen": 1366093824
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002960280842527583,
      "loss": 2.8992,
      "theoretical_loss": 3.5454373590321966,
      "tokens_seen": 1366159360
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002960180541624874,
      "loss": 2.5448,
      "theoretical_loss": 3.545422093653297,
      "tokens_seen": 1366224896
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029600802407221666,
      "loss": 3.0915,
      "theoretical_loss": 3.545406829211661,
      "tokens_seen": 1366290432
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029599799398194584,
      "loss": 2.6512,
      "theoretical_loss": 3.545391565707187,
      "tokens_seen": 1366355968
    },
    {
      "epoch": 4.05,
      "objective/train/docs_used": 1521877,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7869224548339844,
      "objective/train/theoretical_loss": 3.5453763031397716,
      "objective/train/tokens_used": 1386881504,
      "theoretical_loss": 3.5453763031397716,
      "tokens_seen": 1366421504
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.000295987963891675,
      "loss": 2.62,
      "theoretical_loss": 3.5453763031397716,
      "tokens_seen": 1366421504
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002959779338014042,
      "loss": 2.8634,
      "theoretical_loss": 3.545361041509313,
      "tokens_seen": 1366487040
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002959679037111334,
      "loss": 2.8177,
      "theoretical_loss": 3.5453457808157087,
      "tokens_seen": 1366552576
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029595787362086257,
      "loss": 2.8319,
      "theoretical_loss": 3.545330521058856,
      "tokens_seen": 1366618112
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002959478435305918,
      "loss": 2.8518,
      "theoretical_loss": 3.5453152622386526,
      "tokens_seen": 1366683648
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029593781344032093,
      "loss": 2.6757,
      "theoretical_loss": 3.545300004354996,
      "tokens_seen": 1366749184
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029592778335005016,
      "loss": 2.5993,
      "theoretical_loss": 3.545284747407784,
      "tokens_seen": 1366814720
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029591775325977935,
      "loss": 2.7115,
      "theoretical_loss": 3.5452694913969145,
      "tokens_seen": 1366880256
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029590772316950853,
      "loss": 2.9373,
      "theoretical_loss": 3.5452542363222843,
      "tokens_seen": 1366945792
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002958976930792377,
      "loss": 2.7642,
      "theoretical_loss": 3.5452389821837915,
      "tokens_seen": 1367011328
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002958876629889669,
      "loss": 2.6147,
      "theoretical_loss": 3.545223728981334,
      "tokens_seen": 1367076864
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029587763289869607,
      "loss": 2.5695,
      "theoretical_loss": 3.5452084767148095,
      "tokens_seen": 1367142400
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002958676028084253,
      "loss": 2.5337,
      "theoretical_loss": 3.545193225384115,
      "tokens_seen": 1367207936
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029585757271815443,
      "loss": 2.5786,
      "theoretical_loss": 3.545177974989149,
      "tokens_seen": 1367273472
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029584754262788367,
      "loss": 2.6016,
      "theoretical_loss": 3.545162725529809,
      "tokens_seen": 1367339008
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002958375125376128,
      "loss": 2.6825,
      "theoretical_loss": 3.545147477005993,
      "tokens_seen": 1367404544
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029582748244734203,
      "loss": 2.7878,
      "theoretical_loss": 3.545132229417599,
      "tokens_seen": 1367470080
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002958174523570712,
      "loss": 2.369,
      "theoretical_loss": 3.5451169827645233,
      "tokens_seen": 1367535616
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002958074222668004,
      "loss": 2.4724,
      "theoretical_loss": 3.5451017370466653,
      "tokens_seen": 1367601152
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002957973921765296,
      "loss": 2.6906,
      "theoretical_loss": 3.5450864922639225,
      "tokens_seen": 1367666688
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002957873620862588,
      "loss": 2.7822,
      "theoretical_loss": 3.5450712484161917,
      "tokens_seen": 1367732224
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.000295777331995988,
      "loss": 2.6861,
      "theoretical_loss": 3.5450560055033717,
      "tokens_seen": 1367797760
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029576730190571717,
      "loss": 2.868,
      "theoretical_loss": 3.5450407635253605,
      "tokens_seen": 1367863296
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029575727181544635,
      "loss": 2.5099,
      "theoretical_loss": 3.5450255224820557,
      "tokens_seen": 1367928832
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029574724172517553,
      "loss": 2.6857,
      "theoretical_loss": 3.5450102823733554,
      "tokens_seen": 1367994368
    },
    {
      "epoch": 4.05,
      "objective/train/docs_used": 1523240,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.509470224380493,
      "objective/train/theoretical_loss": 3.5449950431991573,
      "objective/train/tokens_used": 1388519904,
      "theoretical_loss": 3.5449950431991573,
      "tokens_seen": 1368059904
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029573721163490477,
      "loss": 2.6579,
      "theoretical_loss": 3.5449950431991573,
      "tokens_seen": 1368059904
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002957271815446339,
      "loss": 2.673,
      "theoretical_loss": 3.544979804959359,
      "tokens_seen": 1368125440
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029571715145436313,
      "loss": 2.5067,
      "theoretical_loss": 3.544964567653859,
      "tokens_seen": 1368190976
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029570712136409226,
      "loss": 2.6669,
      "theoretical_loss": 3.544949331282555,
      "tokens_seen": 1368256512
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002956970912738215,
      "loss": 2.496,
      "theoretical_loss": 3.544934095845345,
      "tokens_seen": 1368322048
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.0002956870611835507,
      "loss": 2.662,
      "theoretical_loss": 3.5449188613421274,
      "tokens_seen": 1368387584
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029567703109327986,
      "loss": 2.8064,
      "theoretical_loss": 3.5449036277727997,
      "tokens_seen": 1368453120
    },
    {
      "epoch": 4.05,
      "learning_rate": 0.00029566700100300904,
      "loss": 2.6266,
      "theoretical_loss": 3.54488839513726,
      "tokens_seen": 1368518656
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002956569709127382,
      "loss": 2.9046,
      "theoretical_loss": 3.5448731634354065,
      "tokens_seen": 1368584192
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002956469408224674,
      "loss": 2.6641,
      "theoretical_loss": 3.5448579326671372,
      "tokens_seen": 1368649728
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029563691073219663,
      "loss": 2.8689,
      "theoretical_loss": 3.5448427028323506,
      "tokens_seen": 1368715264
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029562688064192576,
      "loss": 2.6824,
      "theoretical_loss": 3.544827473930944,
      "tokens_seen": 1368780800
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.000295616850551655,
      "loss": 2.6393,
      "theoretical_loss": 3.5448122459628157,
      "tokens_seen": 1368846336
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002956068204613842,
      "loss": 2.5224,
      "theoretical_loss": 3.544797018927864,
      "tokens_seen": 1368911872
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029559679037111336,
      "loss": 2.7649,
      "theoretical_loss": 3.5447817928259875,
      "tokens_seen": 1368977408
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029558676028084254,
      "loss": 2.516,
      "theoretical_loss": 3.544766567657084,
      "tokens_seen": 1369042944
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002955767301905717,
      "loss": 2.4305,
      "theoretical_loss": 3.5447513434210514,
      "tokens_seen": 1369108480
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002955667001003009,
      "loss": 2.9265,
      "theoretical_loss": 3.544736120117788,
      "tokens_seen": 1369174016
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029555667001003014,
      "loss": 2.8476,
      "theoretical_loss": 3.544720897747192,
      "tokens_seen": 1369239552
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029554663991975926,
      "loss": 2.7277,
      "theoretical_loss": 3.544705676309162,
      "tokens_seen": 1369305088
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002955366098294885,
      "loss": 2.6965,
      "theoretical_loss": 3.544690455803596,
      "tokens_seen": 1369370624
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002955265797392176,
      "loss": 2.6146,
      "theoretical_loss": 3.5446752362303924,
      "tokens_seen": 1369436160
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029551654964894686,
      "loss": 2.7972,
      "theoretical_loss": 3.5446600175894485,
      "tokens_seen": 1369501696
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029550651955867604,
      "loss": 2.6935,
      "theoretical_loss": 3.544644799880664,
      "tokens_seen": 1369567232
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002954964894684052,
      "loss": 2.7009,
      "theoretical_loss": 3.5446295831039363,
      "tokens_seen": 1369632768
    },
    {
      "epoch": 4.06,
      "objective/train/docs_used": 1523836,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.621335744857788,
      "objective/train/theoretical_loss": 3.5446143672591646,
      "objective/train/tokens_used": 1390158304,
      "theoretical_loss": 3.5446143672591646,
      "tokens_seen": 1369698304
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002954864593781344,
      "loss": 2.811,
      "theoretical_loss": 3.5446143672591646,
      "tokens_seen": 1369698304
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002954764292878636,
      "loss": 2.7063,
      "theoretical_loss": 3.5445991523462457,
      "tokens_seen": 1369763840
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029546639919759277,
      "loss": 2.5469,
      "theoretical_loss": 3.5445839383650797,
      "tokens_seen": 1369829376
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.000295456369107322,
      "loss": 2.6845,
      "theoretical_loss": 3.544568725315564,
      "tokens_seen": 1369894912
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029544633901705113,
      "loss": 2.4527,
      "theoretical_loss": 3.544553513197597,
      "tokens_seen": 1369960448
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029543630892678037,
      "loss": 2.7605,
      "theoretical_loss": 3.5445383020110772,
      "tokens_seen": 1370025984
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029542627883650955,
      "loss": 2.7853,
      "theoretical_loss": 3.544523091755903,
      "tokens_seen": 1370091520
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029541624874623873,
      "loss": 2.4475,
      "theoretical_loss": 3.544507882431973,
      "tokens_seen": 1370157056
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002954062186559679,
      "loss": 2.8883,
      "theoretical_loss": 3.5444926740391853,
      "tokens_seen": 1370222592
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002953961885656971,
      "loss": 2.7648,
      "theoretical_loss": 3.5444774665774395,
      "tokens_seen": 1370288128
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029538615847542627,
      "loss": 2.8952,
      "theoretical_loss": 3.5444622600466325,
      "tokens_seen": 1370353664
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002953761283851555,
      "loss": 2.6118,
      "theoretical_loss": 3.544447054446663,
      "tokens_seen": 1370419200
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029536609829488463,
      "loss": 2.7484,
      "theoretical_loss": 3.5444318497774305,
      "tokens_seen": 1370484736
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029535606820461387,
      "loss": 2.6515,
      "theoretical_loss": 3.544416646038833,
      "tokens_seen": 1370550272
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.000295346038114343,
      "loss": 2.5422,
      "theoretical_loss": 3.544401443230769,
      "tokens_seen": 1370615808
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029533600802407223,
      "loss": 2.587,
      "theoretical_loss": 3.544386241353137,
      "tokens_seen": 1370681344
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002953259779338014,
      "loss": 2.7654,
      "theoretical_loss": 3.544371040405836,
      "tokens_seen": 1370746880
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002953159478435306,
      "loss": 2.4942,
      "theoretical_loss": 3.5443558403887643,
      "tokens_seen": 1370812416
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002953059177532598,
      "loss": 2.6671,
      "theoretical_loss": 3.5443406413018197,
      "tokens_seen": 1370877952
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.000295295887662989,
      "loss": 2.5489,
      "theoretical_loss": 3.544325443144902,
      "tokens_seen": 1370943488
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029528585757271814,
      "loss": 2.5203,
      "theoretical_loss": 3.5443102459179094,
      "tokens_seen": 1371009024
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029527582748244737,
      "loss": 2.9006,
      "theoretical_loss": 3.5442950496207413,
      "tokens_seen": 1371074560
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002952657973921765,
      "loss": 2.9251,
      "theoretical_loss": 3.5442798542532947,
      "tokens_seen": 1371140096
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029525576730190573,
      "loss": 2.6135,
      "theoretical_loss": 3.54426465981547,
      "tokens_seen": 1371205632
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002952457372116349,
      "loss": 2.7752,
      "theoretical_loss": 3.544249466307164,
      "tokens_seen": 1371271168
    },
    {
      "epoch": 4.06,
      "objective/train/docs_used": 1525089,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.015775203704834,
      "objective/train/theoretical_loss": 3.5442342737282777,
      "objective/train/tokens_used": 1391796704,
      "theoretical_loss": 3.5442342737282777,
      "tokens_seen": 1371336704
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002952357071213641,
      "loss": 2.9369,
      "theoretical_loss": 3.5442342737282777,
      "tokens_seen": 1371336704
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002952256770310933,
      "loss": 2.5958,
      "theoretical_loss": 3.5442190820787083,
      "tokens_seen": 1371402240
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029521564694082246,
      "loss": 2.8745,
      "theoretical_loss": 3.544203891358355,
      "tokens_seen": 1371467776
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029520561685055164,
      "loss": 2.5521,
      "theoretical_loss": 3.5441887015671165,
      "tokens_seen": 1371533312
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002951955867602809,
      "loss": 2.7852,
      "theoretical_loss": 3.5441735127048917,
      "tokens_seen": 1371598848
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029518555667001,
      "loss": 2.6479,
      "theoretical_loss": 3.544158324771579,
      "tokens_seen": 1371664384
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029517552657973924,
      "loss": 2.8104,
      "theoretical_loss": 3.5441431377670773,
      "tokens_seen": 1371729920
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029516549648946836,
      "loss": 2.7465,
      "theoretical_loss": 3.5441279516912862,
      "tokens_seen": 1371795456
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002951554663991976,
      "loss": 2.5285,
      "theoretical_loss": 3.5441127665441035,
      "tokens_seen": 1371860992
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002951454363089268,
      "loss": 2.6205,
      "theoretical_loss": 3.5440975823254286,
      "tokens_seen": 1371926528
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029513540621865596,
      "loss": 2.5794,
      "theoretical_loss": 3.5440823990351604,
      "tokens_seen": 1371992064
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029512537612838514,
      "loss": 2.6045,
      "theoretical_loss": 3.5440672166731977,
      "tokens_seen": 1372057600
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002951153460381144,
      "loss": 2.8003,
      "theoretical_loss": 3.5440520352394396,
      "tokens_seen": 1372123136
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002951053159478435,
      "loss": 2.9421,
      "theoretical_loss": 3.5440368547337844,
      "tokens_seen": 1372188672
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029509528585757274,
      "loss": 2.5751,
      "theoretical_loss": 3.5440216751561318,
      "tokens_seen": 1372254208
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029508525576730187,
      "loss": 2.8673,
      "theoretical_loss": 3.5440064965063804,
      "tokens_seen": 1372319744
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002950752256770311,
      "loss": 2.8663,
      "theoretical_loss": 3.543991318784429,
      "tokens_seen": 1372385280
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002950651955867603,
      "loss": 2.6738,
      "theoretical_loss": 3.5439761419901767,
      "tokens_seen": 1372450816
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029505516549648946,
      "loss": 2.6099,
      "theoretical_loss": 3.5439609661235227,
      "tokens_seen": 1372516352
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029504513540621865,
      "loss": 2.6654,
      "theoretical_loss": 3.543945791184366,
      "tokens_seen": 1372581888
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002950351053159478,
      "loss": 2.9022,
      "theoretical_loss": 3.5439306171726055,
      "tokens_seen": 1372647424
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029502507522567706,
      "loss": 2.6579,
      "theoretical_loss": 3.54391544408814,
      "tokens_seen": 1372712960
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029501504513540624,
      "loss": 2.8065,
      "theoretical_loss": 3.5439002719308696,
      "tokens_seen": 1372778496
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002950050150451354,
      "loss": 2.8156,
      "theoretical_loss": 3.5438851007006917,
      "tokens_seen": 1372844032
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002949949849548646,
      "loss": 2.6603,
      "theoretical_loss": 3.543869930397507,
      "tokens_seen": 1372909568
    },
    {
      "epoch": 4.06,
      "objective/train/docs_used": 1525961,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.1202309131622314,
      "objective/train/theoretical_loss": 3.5438547610212137,
      "objective/train/tokens_used": 1393435104,
      "theoretical_loss": 3.5438547610212137,
      "tokens_seen": 1372975104
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002949849548645938,
      "loss": 2.8852,
      "theoretical_loss": 3.5438547610212137,
      "tokens_seen": 1372975104
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029497492477432297,
      "loss": 2.841,
      "theoretical_loss": 3.543839592571711,
      "tokens_seen": 1373040640
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002949648946840522,
      "loss": 2.7847,
      "theoretical_loss": 3.543824425048898,
      "tokens_seen": 1373106176
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029495486459378133,
      "loss": 2.6657,
      "theoretical_loss": 3.543809258452675,
      "tokens_seen": 1373171712
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029494483450351057,
      "loss": 2.8445,
      "theoretical_loss": 3.5437940927829388,
      "tokens_seen": 1373237248
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029493480441323975,
      "loss": 2.8022,
      "theoretical_loss": 3.5437789280395915,
      "tokens_seen": 1373302784
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029492477432296893,
      "loss": 2.6647,
      "theoretical_loss": 3.54376376422253,
      "tokens_seen": 1373368320
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002949147442326981,
      "loss": 2.7687,
      "theoretical_loss": 3.5437486013316546,
      "tokens_seen": 1373433856
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002949047141424273,
      "loss": 2.8542,
      "theoretical_loss": 3.5437334393668642,
      "tokens_seen": 1373499392
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029489468405215647,
      "loss": 2.4105,
      "theoretical_loss": 3.543718278328058,
      "tokens_seen": 1373564928
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002948846539618857,
      "loss": 2.8407,
      "theoretical_loss": 3.5437031182151357,
      "tokens_seen": 1373630464
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029487462387161483,
      "loss": 2.7031,
      "theoretical_loss": 3.543687959027996,
      "tokens_seen": 1373696000
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029486459378134407,
      "loss": 2.8304,
      "theoretical_loss": 3.5436728007665392,
      "tokens_seen": 1373761536
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002948545636910732,
      "loss": 2.7599,
      "theoretical_loss": 3.543657643430663,
      "tokens_seen": 1373827072
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029484453360080243,
      "loss": 2.7486,
      "theoretical_loss": 3.5436424870202683,
      "tokens_seen": 1373892608
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002948345035105316,
      "loss": 2.7511,
      "theoretical_loss": 3.5436273315352538,
      "tokens_seen": 1373958144
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002948244734202608,
      "loss": 2.7855,
      "theoretical_loss": 3.543612176975519,
      "tokens_seen": 1374023680
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029481444332999,
      "loss": 2.836,
      "theoretical_loss": 3.543597023340963,
      "tokens_seen": 1374089216
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002948044132397192,
      "loss": 2.4758,
      "theoretical_loss": 3.5435818706314848,
      "tokens_seen": 1374154752
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029479438314944834,
      "loss": 2.6613,
      "theoretical_loss": 3.5435667188469853,
      "tokens_seen": 1374220288
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029478435305917757,
      "loss": 2.6733,
      "theoretical_loss": 3.5435515679873624,
      "tokens_seen": 1374285824
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002947743229689067,
      "loss": 2.6872,
      "theoretical_loss": 3.5435364180525166,
      "tokens_seen": 1374351360
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029476429287863593,
      "loss": 3.1353,
      "theoretical_loss": 3.5435212690423463,
      "tokens_seen": 1374416896
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002947542627883651,
      "loss": 2.7879,
      "theoretical_loss": 3.543506120956752,
      "tokens_seen": 1374482432
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002947442326980943,
      "loss": 2.5917,
      "theoretical_loss": 3.543490973795633,
      "tokens_seen": 1374547968
    },
    {
      "epoch": 4.06,
      "objective/train/docs_used": 1531216,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.648439407348633,
      "objective/train/theoretical_loss": 3.543475827558888,
      "objective/train/tokens_used": 1395073504,
      "theoretical_loss": 3.543475827558888,
      "tokens_seen": 1374613504
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002947342026078235,
      "loss": 2.9129,
      "theoretical_loss": 3.543475827558888,
      "tokens_seen": 1374613504
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029472417251755266,
      "loss": 2.5568,
      "theoretical_loss": 3.543460682246417,
      "tokens_seen": 1374679040
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029471414242728184,
      "loss": 2.6088,
      "theoretical_loss": 3.54344553785812,
      "tokens_seen": 1374744576
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002947041123370111,
      "loss": 2.8646,
      "theoretical_loss": 3.5434303943938965,
      "tokens_seen": 1374810112
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002946940822467402,
      "loss": 2.8851,
      "theoretical_loss": 3.543415251853645,
      "tokens_seen": 1374875648
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029468405215646944,
      "loss": 2.6553,
      "theoretical_loss": 3.5434001102372665,
      "tokens_seen": 1374941184
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029467402206619856,
      "loss": 2.9477,
      "theoretical_loss": 3.54338496954466,
      "tokens_seen": 1375006720
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002946639919759278,
      "loss": 2.605,
      "theoretical_loss": 3.543369829775725,
      "tokens_seen": 1375072256
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.000294653961885657,
      "loss": 2.9116,
      "theoretical_loss": 3.543354690930361,
      "tokens_seen": 1375137792
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029464393179538616,
      "loss": 2.7982,
      "theoretical_loss": 3.543339553008468,
      "tokens_seen": 1375203328
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029463390170511534,
      "loss": 2.6475,
      "theoretical_loss": 3.543324416009945,
      "tokens_seen": 1375268864
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002946238716148446,
      "loss": 2.7576,
      "theoretical_loss": 3.5433092799346926,
      "tokens_seen": 1375334400
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002946138415245737,
      "loss": 2.9175,
      "theoretical_loss": 3.5432941447826103,
      "tokens_seen": 1375399936
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029460381143430294,
      "loss": 2.9014,
      "theoretical_loss": 3.543279010553597,
      "tokens_seen": 1375465472
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029459378134403207,
      "loss": 2.6803,
      "theoretical_loss": 3.5432638772475533,
      "tokens_seen": 1375531008
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002945837512537613,
      "loss": 2.816,
      "theoretical_loss": 3.543248744864379,
      "tokens_seen": 1375596544
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002945737211634905,
      "loss": 2.6571,
      "theoretical_loss": 3.5432336134039732,
      "tokens_seen": 1375662080
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029456369107321966,
      "loss": 2.6177,
      "theoretical_loss": 3.5432184828662363,
      "tokens_seen": 1375727616
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029455366098294885,
      "loss": 2.7331,
      "theoretical_loss": 3.543203353251067,
      "tokens_seen": 1375793152
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.000294543630892678,
      "loss": 2.8117,
      "theoretical_loss": 3.5431882245583664,
      "tokens_seen": 1375858688
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002945336008024072,
      "loss": 2.6867,
      "theoretical_loss": 3.543173096788034,
      "tokens_seen": 1375924224
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029452357071213644,
      "loss": 2.5584,
      "theoretical_loss": 3.543157969939969,
      "tokens_seen": 1375989760
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029451354062186557,
      "loss": 2.6233,
      "theoretical_loss": 3.5431428440140724,
      "tokens_seen": 1376055296
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002945035105315948,
      "loss": 2.7276,
      "theoretical_loss": 3.5431277190102426,
      "tokens_seen": 1376120832
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029449348044132393,
      "loss": 2.9551,
      "theoretical_loss": 3.543112594928381,
      "tokens_seen": 1376186368
    },
    {
      "epoch": 4.06,
      "objective/train/docs_used": 1536193,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.979797601699829,
      "objective/train/theoretical_loss": 3.543097471768386,
      "objective/train/tokens_used": 1396711904,
      "theoretical_loss": 3.543097471768386,
      "tokens_seen": 1376251904
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029448345035105317,
      "loss": 2.8658,
      "theoretical_loss": 3.543097471768386,
      "tokens_seen": 1376251904
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029447342026078235,
      "loss": 2.7535,
      "theoretical_loss": 3.5430823495301587,
      "tokens_seen": 1376317440
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029446339017051153,
      "loss": 2.8096,
      "theoretical_loss": 3.5430672282135984,
      "tokens_seen": 1376382976
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002944533600802407,
      "loss": 2.5962,
      "theoretical_loss": 3.5430521078186055,
      "tokens_seen": 1376448512
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029444332998996995,
      "loss": 2.9369,
      "theoretical_loss": 3.5430369883450794,
      "tokens_seen": 1376514048
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002944332998996991,
      "loss": 2.8077,
      "theoretical_loss": 3.5430218697929208,
      "tokens_seen": 1376579584
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002944232698094283,
      "loss": 2.5626,
      "theoretical_loss": 3.5430067521620288,
      "tokens_seen": 1376645120
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029441323971915744,
      "loss": 2.7621,
      "theoretical_loss": 3.542991635452304,
      "tokens_seen": 1376710656
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029440320962888667,
      "loss": 2.5922,
      "theoretical_loss": 3.5429765196636462,
      "tokens_seen": 1376776192
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029439317953861585,
      "loss": 2.659,
      "theoretical_loss": 3.542961404795956,
      "tokens_seen": 1376841728
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029438314944834503,
      "loss": 2.6147,
      "theoretical_loss": 3.542946290849132,
      "tokens_seen": 1376907264
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002943731193580742,
      "loss": 2.6262,
      "theoretical_loss": 3.5429311778230765,
      "tokens_seen": 1376972800
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002943630892678034,
      "loss": 2.7149,
      "theoretical_loss": 3.5429160657176877,
      "tokens_seen": 1377038336
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002943530591775326,
      "loss": 2.8174,
      "theoretical_loss": 3.5429009545328665,
      "tokens_seen": 1377103872
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002943430290872618,
      "loss": 2.9529,
      "theoretical_loss": 3.542885844268513,
      "tokens_seen": 1377169408
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029433299899699094,
      "loss": 2.9298,
      "theoretical_loss": 3.5428707349245263,
      "tokens_seen": 1377234944
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002943229689067202,
      "loss": 2.846,
      "theoretical_loss": 3.5428556265008084,
      "tokens_seen": 1377300480
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002943129388164493,
      "loss": 2.5321,
      "theoretical_loss": 3.5428405189972585,
      "tokens_seen": 1377366016
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029430290872617854,
      "loss": 2.5836,
      "theoretical_loss": 3.5428254124137766,
      "tokens_seen": 1377431552
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029429287863590777,
      "loss": 2.6707,
      "theoretical_loss": 3.5428103067502628,
      "tokens_seen": 1377497088
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002942828485456369,
      "loss": 2.6525,
      "theoretical_loss": 3.5427952020066176,
      "tokens_seen": 1377562624
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029427281845536613,
      "loss": 2.6044,
      "theoretical_loss": 3.5427800981827415,
      "tokens_seen": 1377628160
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002942627883650953,
      "loss": 2.6824,
      "theoretical_loss": 3.5427649952785343,
      "tokens_seen": 1377693696
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002942527582748245,
      "loss": 2.5064,
      "theoretical_loss": 3.5427498932938972,
      "tokens_seen": 1377759232
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002942427281845537,
      "loss": 2.8199,
      "theoretical_loss": 3.5427347922287287,
      "tokens_seen": 1377824768
    },
    {
      "epoch": 4.06,
      "objective/train/docs_used": 1541285,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9072976112365723,
      "objective/train/theoretical_loss": 3.5427196920829305,
      "objective/train/tokens_used": 1398350304,
      "theoretical_loss": 3.5427196920829305,
      "tokens_seen": 1377890304
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029423269809428286,
      "loss": 2.781,
      "theoretical_loss": 3.5427196920829305,
      "tokens_seen": 1377890304
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029422266800401204,
      "loss": 2.6368,
      "theoretical_loss": 3.5427045928564027,
      "tokens_seen": 1377955840
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002942126379137413,
      "loss": 2.9298,
      "theoretical_loss": 3.5426894945490446,
      "tokens_seen": 1378021376
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002942026078234704,
      "loss": 2.6657,
      "theoretical_loss": 3.5426743971607584,
      "tokens_seen": 1378086912
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029419257773319964,
      "loss": 2.7524,
      "theoretical_loss": 3.542659300691443,
      "tokens_seen": 1378152448
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029418254764292876,
      "loss": 2.7688,
      "theoretical_loss": 3.5426442051409985,
      "tokens_seen": 1378217984
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.000294172517552658,
      "loss": 2.8809,
      "theoretical_loss": 3.542629110509327,
      "tokens_seen": 1378283520
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002941624874623872,
      "loss": 2.8175,
      "theoretical_loss": 3.542614016796327,
      "tokens_seen": 1378349056
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029415245737211636,
      "loss": 2.5861,
      "theoretical_loss": 3.5425989240019,
      "tokens_seen": 1378414592
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029414242728184554,
      "loss": 2.547,
      "theoretical_loss": 3.5425838321259464,
      "tokens_seen": 1378480128
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002941323971915748,
      "loss": 2.653,
      "theoretical_loss": 3.542568741168366,
      "tokens_seen": 1378545664
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002941223671013039,
      "loss": 2.7472,
      "theoretical_loss": 3.54255365112906,
      "tokens_seen": 1378611200
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029411233701103314,
      "loss": 2.5158,
      "theoretical_loss": 3.5425385620079286,
      "tokens_seen": 1378676736
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029410230692076227,
      "loss": 2.7334,
      "theoretical_loss": 3.5425234738048723,
      "tokens_seen": 1378742272
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002940922768304915,
      "loss": 2.7095,
      "theoretical_loss": 3.5425083865197915,
      "tokens_seen": 1378807808
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002940822467402207,
      "loss": 2.6454,
      "theoretical_loss": 3.5424933001525867,
      "tokens_seen": 1378873344
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029407221664994986,
      "loss": 2.6793,
      "theoretical_loss": 3.5424782147031584,
      "tokens_seen": 1378938880
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029406218655967905,
      "loss": 2.7721,
      "theoretical_loss": 3.5424631301714076,
      "tokens_seen": 1379004416
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002940521564694082,
      "loss": 2.8642,
      "theoretical_loss": 3.5424480465572343,
      "tokens_seen": 1379069952
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002940421263791374,
      "loss": 2.719,
      "theoretical_loss": 3.5424329638605396,
      "tokens_seen": 1379135488
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029403209628886664,
      "loss": 2.6414,
      "theoretical_loss": 3.5424178820812235,
      "tokens_seen": 1379201024
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029402206619859577,
      "loss": 2.6062,
      "theoretical_loss": 3.542402801219187,
      "tokens_seen": 1379266560
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.000294012036108325,
      "loss": 2.721,
      "theoretical_loss": 3.5423877212743307,
      "tokens_seen": 1379332096
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029400200601805413,
      "loss": 2.6475,
      "theoretical_loss": 3.542372642246555,
      "tokens_seen": 1379397632
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029399197592778337,
      "loss": 2.6133,
      "theoretical_loss": 3.542357564135761,
      "tokens_seen": 1379463168
    },
    {
      "epoch": 4.06,
      "objective/train/docs_used": 1546362,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.725088119506836,
      "objective/train/theoretical_loss": 3.542342486941849,
      "objective/train/tokens_used": 1399988704,
      "theoretical_loss": 3.542342486941849,
      "tokens_seen": 1379528704
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029398194583751255,
      "loss": 2.7855,
      "theoretical_loss": 3.542342486941849,
      "tokens_seen": 1379528704
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029397191574724173,
      "loss": 2.7274,
      "theoretical_loss": 3.5423274106647202,
      "tokens_seen": 1379594240
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002939618856569709,
      "loss": 2.7593,
      "theoretical_loss": 3.542312335304275,
      "tokens_seen": 1379659776
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029395185556670015,
      "loss": 2.5319,
      "theoretical_loss": 3.5422972608604133,
      "tokens_seen": 1379725312
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002939418254764293,
      "loss": 2.5069,
      "theoretical_loss": 3.5422821873330372,
      "tokens_seen": 1379790848
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002939317953861585,
      "loss": 2.8197,
      "theoretical_loss": 3.542267114722047,
      "tokens_seen": 1379856384
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029392176529588764,
      "loss": 2.6525,
      "theoretical_loss": 3.542252043027343,
      "tokens_seen": 1379921920
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029391173520561687,
      "loss": 2.783,
      "theoretical_loss": 3.5422369722488263,
      "tokens_seen": 1379987456
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029390170511534605,
      "loss": 2.641,
      "theoretical_loss": 3.5422219023863977,
      "tokens_seen": 1380052992
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029389167502507523,
      "loss": 2.7446,
      "theoretical_loss": 3.542206833439958,
      "tokens_seen": 1380118528
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002938816449348044,
      "loss": 2.8127,
      "theoretical_loss": 3.5421917654094086,
      "tokens_seen": 1380184064
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002938716148445336,
      "loss": 2.7673,
      "theoretical_loss": 3.542176698294649,
      "tokens_seen": 1380249600
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002938615847542628,
      "loss": 2.7564,
      "theoretical_loss": 3.5421616320955813,
      "tokens_seen": 1380315136
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.000293851554663992,
      "loss": 2.6335,
      "theoretical_loss": 3.5421465668121064,
      "tokens_seen": 1380380672
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029384152457372114,
      "loss": 2.8381,
      "theoretical_loss": 3.542131502444124,
      "tokens_seen": 1380446208
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002938314944834504,
      "loss": 2.9114,
      "theoretical_loss": 3.542116438991536,
      "tokens_seen": 1380511744
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002938214643931795,
      "loss": 2.6381,
      "theoretical_loss": 3.542101376454243,
      "tokens_seen": 1380577280
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029381143430290874,
      "loss": 2.6397,
      "theoretical_loss": 3.5420863148321464,
      "tokens_seen": 1380642816
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002938014042126379,
      "loss": 2.5404,
      "theoretical_loss": 3.5420712541251462,
      "tokens_seen": 1380708352
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002937913741223671,
      "loss": 2.5423,
      "theoretical_loss": 3.5420561943331443,
      "tokens_seen": 1380773888
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002937813440320963,
      "loss": 2.795,
      "theoretical_loss": 3.5420411354560413,
      "tokens_seen": 1380839424
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002937713139418255,
      "loss": 2.565,
      "theoretical_loss": 3.542026077493738,
      "tokens_seen": 1380904960
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029376128385155464,
      "loss": 2.636,
      "theoretical_loss": 3.542011020446136,
      "tokens_seen": 1380970496
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002937512537612839,
      "loss": 2.8553,
      "theoretical_loss": 3.541995964313136,
      "tokens_seen": 1381036032
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.000293741223671013,
      "loss": 2.5084,
      "theoretical_loss": 3.5419809090946384,
      "tokens_seen": 1381101568
    },
    {
      "epoch": 4.06,
      "objective/train/docs_used": 1551454,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5593550205230713,
      "objective/train/theoretical_loss": 3.5419658547905453,
      "objective/train/tokens_used": 1401627104,
      "theoretical_loss": 3.5419658547905453,
      "tokens_seen": 1381167104
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029373119358074224,
      "loss": 2.5186,
      "theoretical_loss": 3.5419658547905453,
      "tokens_seen": 1381167104
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002937211634904714,
      "loss": 2.4587,
      "theoretical_loss": 3.5419508014007572,
      "tokens_seen": 1381232640
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002937111334002006,
      "loss": 2.6895,
      "theoretical_loss": 3.5419357489251757,
      "tokens_seen": 1381298176
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002937011033099298,
      "loss": 2.7143,
      "theoretical_loss": 3.541920697363701,
      "tokens_seen": 1381363712
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029369107321965896,
      "loss": 2.7308,
      "theoretical_loss": 3.5419056467162355,
      "tokens_seen": 1381429248
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029368104312938815,
      "loss": 2.8324,
      "theoretical_loss": 3.5418905969826793,
      "tokens_seen": 1381494784
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002936710130391174,
      "loss": 2.6625,
      "theoretical_loss": 3.541875548162934,
      "tokens_seen": 1381560320
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002936609829488465,
      "loss": 2.7212,
      "theoretical_loss": 3.5418605002569006,
      "tokens_seen": 1381625856
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029365095285857574,
      "loss": 2.706,
      "theoretical_loss": 3.5418454532644796,
      "tokens_seen": 1381691392
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029364092276830487,
      "loss": 2.4649,
      "theoretical_loss": 3.5418304071855737,
      "tokens_seen": 1381756928
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002936308926780341,
      "loss": 2.7022,
      "theoretical_loss": 3.5418153620200834,
      "tokens_seen": 1381822464
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002936208625877633,
      "loss": 2.6936,
      "theoretical_loss": 3.5418003177679096,
      "tokens_seen": 1381888000
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029361083249749247,
      "loss": 2.9936,
      "theoretical_loss": 3.541785274428954,
      "tokens_seen": 1381953536
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029360080240722165,
      "loss": 2.7206,
      "theoretical_loss": 3.5417702320031177,
      "tokens_seen": 1382019072
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002935907723169509,
      "loss": 2.5303,
      "theoretical_loss": 3.541755190490302,
      "tokens_seen": 1382084608
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029358074222668,
      "loss": 2.9346,
      "theoretical_loss": 3.541740149890408,
      "tokens_seen": 1382150144
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029357071213640925,
      "loss": 2.3615,
      "theoretical_loss": 3.541725110203337,
      "tokens_seen": 1382215680
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002935606820461384,
      "loss": 2.7336,
      "theoretical_loss": 3.541710071428991,
      "tokens_seen": 1382281216
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002935506519558676,
      "loss": 2.5758,
      "theoretical_loss": 3.5416950335672706,
      "tokens_seen": 1382346752
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029354062186559684,
      "loss": 2.6175,
      "theoretical_loss": 3.541679996618077,
      "tokens_seen": 1382412288
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029353059177532597,
      "loss": 2.7783,
      "theoretical_loss": 3.541664960581312,
      "tokens_seen": 1382477824
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002935205616850552,
      "loss": 2.6292,
      "theoretical_loss": 3.541649925456878,
      "tokens_seen": 1382543360
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029351053159478433,
      "loss": 2.7404,
      "theoretical_loss": 3.5416348912446742,
      "tokens_seen": 1382608896
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029350050150451357,
      "loss": 2.5828,
      "theoretical_loss": 3.541619857944604,
      "tokens_seen": 1382674432
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029349047141424275,
      "loss": 2.6829,
      "theoretical_loss": 3.5416048255565675,
      "tokens_seen": 1382739968
    },
    {
      "epoch": 4.06,
      "objective/train/docs_used": 1552054,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.0920627117156982,
      "objective/train/theoretical_loss": 3.5415897940804664,
      "objective/train/tokens_used": 1403265504,
      "theoretical_loss": 3.5415897940804664,
      "tokens_seen": 1382805504
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029348044132397193,
      "loss": 2.8501,
      "theoretical_loss": 3.5415897940804664,
      "tokens_seen": 1382805504
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002934704112337011,
      "loss": 2.7506,
      "theoretical_loss": 3.541574763516203,
      "tokens_seen": 1382871040
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029346038114343035,
      "loss": 2.7521,
      "theoretical_loss": 3.5415597338636777,
      "tokens_seen": 1382936576
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002934503510531595,
      "loss": 2.869,
      "theoretical_loss": 3.5415447051227926,
      "tokens_seen": 1383002112
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002934403209628887,
      "loss": 2.6461,
      "theoretical_loss": 3.5415296772934495,
      "tokens_seen": 1383067648
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029343029087261784,
      "loss": 2.7489,
      "theoretical_loss": 3.541514650375549,
      "tokens_seen": 1383133184
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029342026078234707,
      "loss": 2.9143,
      "theoretical_loss": 3.541499624368993,
      "tokens_seen": 1383198720
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029341023069207625,
      "loss": 2.7198,
      "theoretical_loss": 3.5414845992736836,
      "tokens_seen": 1383264256
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029340020060180543,
      "loss": 2.4439,
      "theoretical_loss": 3.541469575089522,
      "tokens_seen": 1383329792
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002933901705115346,
      "loss": 2.6594,
      "theoretical_loss": 3.5414545518164093,
      "tokens_seen": 1383395328
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002933801404212638,
      "loss": 2.6616,
      "theoretical_loss": 3.541439529454248,
      "tokens_seen": 1383460864
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.000293370110330993,
      "loss": 2.879,
      "theoretical_loss": 3.541424508002939,
      "tokens_seen": 1383526400
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002933600802407222,
      "loss": 2.6416,
      "theoretical_loss": 3.541409487462384,
      "tokens_seen": 1383591936
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029335005015045134,
      "loss": 2.6155,
      "theoretical_loss": 3.541394467832485,
      "tokens_seen": 1383657472
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002933400200601806,
      "loss": 2.6164,
      "theoretical_loss": 3.541379449113144,
      "tokens_seen": 1383723008
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002933299899699097,
      "loss": 2.8243,
      "theoretical_loss": 3.5413644313042614,
      "tokens_seen": 1383788544
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029331995987963894,
      "loss": 2.6085,
      "theoretical_loss": 3.5413494144057402,
      "tokens_seen": 1383854080
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002933099297893681,
      "loss": 2.6678,
      "theoretical_loss": 3.541334398417481,
      "tokens_seen": 1383919616
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002932998996990973,
      "loss": 2.6854,
      "theoretical_loss": 3.541319383339387,
      "tokens_seen": 1383985152
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002932898696088265,
      "loss": 2.4524,
      "theoretical_loss": 3.5413043691713586,
      "tokens_seen": 1384050688
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002932798395185557,
      "loss": 2.6441,
      "theoretical_loss": 3.541289355913298,
      "tokens_seen": 1384116224
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029326980942828484,
      "loss": 2.7086,
      "theoretical_loss": 3.5412743435651066,
      "tokens_seen": 1384181760
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002932597793380141,
      "loss": 2.6357,
      "theoretical_loss": 3.541259332126687,
      "tokens_seen": 1384247296
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002932497492477432,
      "loss": 2.5468,
      "theoretical_loss": 3.54124432159794,
      "tokens_seen": 1384312832
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029323971915747244,
      "loss": 2.7114,
      "theoretical_loss": 3.5412293119787686,
      "tokens_seen": 1384378368
    },
    {
      "epoch": 4.06,
      "objective/train/docs_used": 1553388,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.496983051300049,
      "objective/train/theoretical_loss": 3.5412143032690735,
      "objective/train/tokens_used": 1404903904,
      "theoretical_loss": 3.5412143032690735,
      "tokens_seen": 1384443904
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002932296890672016,
      "loss": 2.5614,
      "theoretical_loss": 3.5412143032690735,
      "tokens_seen": 1384443904
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002932196589769308,
      "loss": 2.6653,
      "theoretical_loss": 3.5411992954687572,
      "tokens_seen": 1384509440
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029320962888666,
      "loss": 2.656,
      "theoretical_loss": 3.541184288577721,
      "tokens_seen": 1384574976
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029319959879638916,
      "loss": 2.7702,
      "theoretical_loss": 3.5411692825958676,
      "tokens_seen": 1384640512
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029318956870611835,
      "loss": 2.6418,
      "theoretical_loss": 3.5411542775230984,
      "tokens_seen": 1384706048
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002931795386158476,
      "loss": 2.6088,
      "theoretical_loss": 3.541139273359315,
      "tokens_seen": 1384771584
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002931695085255767,
      "loss": 2.9696,
      "theoretical_loss": 3.54112427010442,
      "tokens_seen": 1384837120
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029315947843530594,
      "loss": 2.6929,
      "theoretical_loss": 3.541109267758315,
      "tokens_seen": 1384902656
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029314944834503507,
      "loss": 2.9294,
      "theoretical_loss": 3.541094266320902,
      "tokens_seen": 1384968192
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002931394182547643,
      "loss": 2.5998,
      "theoretical_loss": 3.541079265792083,
      "tokens_seen": 1385033728
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002931293881644935,
      "loss": 2.5983,
      "theoretical_loss": 3.5410642661717597,
      "tokens_seen": 1385099264
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029311935807422267,
      "loss": 2.7155,
      "theoretical_loss": 3.541049267459834,
      "tokens_seen": 1385164800
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029310932798395185,
      "loss": 2.8903,
      "theoretical_loss": 3.5410342696562083,
      "tokens_seen": 1385230336
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002930992978936811,
      "loss": 2.6472,
      "theoretical_loss": 3.541019272760785,
      "tokens_seen": 1385295872
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002930892678034102,
      "loss": 2.8402,
      "theoretical_loss": 3.5410042767734655,
      "tokens_seen": 1385361408
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029307923771313945,
      "loss": 2.6837,
      "theoretical_loss": 3.5409892816941513,
      "tokens_seen": 1385426944
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002930692076228686,
      "loss": 2.6073,
      "theoretical_loss": 3.5409742875227463,
      "tokens_seen": 1385492480
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002930591775325978,
      "loss": 2.6603,
      "theoretical_loss": 3.540959294259151,
      "tokens_seen": 1385558016
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.000293049147442327,
      "loss": 2.7129,
      "theoretical_loss": 3.540944301903268,
      "tokens_seen": 1385623552
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029303911735205617,
      "loss": 2.4107,
      "theoretical_loss": 3.540929310454999,
      "tokens_seen": 1385689088
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029302908726178535,
      "loss": 2.8114,
      "theoretical_loss": 3.540914319914247,
      "tokens_seen": 1385754624
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029301905717151453,
      "loss": 2.68,
      "theoretical_loss": 3.5408993302809133,
      "tokens_seen": 1385820160
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002930090270812437,
      "loss": 2.6964,
      "theoretical_loss": 3.540884341554901,
      "tokens_seen": 1385885696
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029299899699097295,
      "loss": 2.6005,
      "theoretical_loss": 3.5408693537361113,
      "tokens_seen": 1385951232
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002929889669007021,
      "loss": 2.7721,
      "theoretical_loss": 3.540854366824447,
      "tokens_seen": 1386016768
    },
    {
      "epoch": 4.06,
      "objective/train/docs_used": 1554184,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.679152250289917,
      "objective/train/theoretical_loss": 3.5408393808198104,
      "objective/train/tokens_used": 1406542304,
      "theoretical_loss": 3.5408393808198104,
      "tokens_seen": 1386082304
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002929789368104313,
      "loss": 2.7258,
      "theoretical_loss": 3.5408393808198104,
      "tokens_seen": 1386082304
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002929689067201605,
      "loss": 2.9701,
      "theoretical_loss": 3.540824395722103,
      "tokens_seen": 1386147840
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002929588766298897,
      "loss": 2.6006,
      "theoretical_loss": 3.540809411531228,
      "tokens_seen": 1386213376
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029294884653961885,
      "loss": 2.8074,
      "theoretical_loss": 3.5407944282470867,
      "tokens_seen": 1386278912
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029293881644934804,
      "loss": 2.5889,
      "theoretical_loss": 3.540779445869582,
      "tokens_seen": 1386344448
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002929287863590772,
      "loss": 2.7228,
      "theoretical_loss": 3.540764464398616,
      "tokens_seen": 1386409984
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029291875626880645,
      "loss": 2.6788,
      "theoretical_loss": 3.540749483834091,
      "tokens_seen": 1386475520
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002929087261785356,
      "loss": 2.8537,
      "theoretical_loss": 3.540734504175909,
      "tokens_seen": 1386541056
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002928986960882648,
      "loss": 2.8571,
      "theoretical_loss": 3.540719525423973,
      "tokens_seen": 1386606592
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029288866599799394,
      "loss": 2.8249,
      "theoretical_loss": 3.540704547578185,
      "tokens_seen": 1386672128
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002928786359077232,
      "loss": 2.819,
      "theoretical_loss": 3.540689570638447,
      "tokens_seen": 1386737664
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029286860581745236,
      "loss": 2.8209,
      "theoretical_loss": 3.5406745946046616,
      "tokens_seen": 1386803200
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029285857572718154,
      "loss": 2.8291,
      "theoretical_loss": 3.540659619476732,
      "tokens_seen": 1386868736
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002928485456369107,
      "loss": 2.6679,
      "theoretical_loss": 3.5406446452545595,
      "tokens_seen": 1386934272
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002928385155466399,
      "loss": 2.8749,
      "theoretical_loss": 3.5406296719380475,
      "tokens_seen": 1386999808
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002928284854563691,
      "loss": 2.7219,
      "theoretical_loss": 3.5406146995270973,
      "tokens_seen": 1387065344
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002928184553660983,
      "loss": 2.7416,
      "theoretical_loss": 3.540599728021612,
      "tokens_seen": 1387130880
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029280842527582744,
      "loss": 2.7386,
      "theoretical_loss": 3.540584757421494,
      "tokens_seen": 1387196416
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002927983951855567,
      "loss": 2.4624,
      "theoretical_loss": 3.5405697877266458,
      "tokens_seen": 1387261952
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002927883650952859,
      "loss": 2.6384,
      "theoretical_loss": 3.54055481893697,
      "tokens_seen": 1387327488
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029277833500501504,
      "loss": 2.6868,
      "theoretical_loss": 3.540539851052369,
      "tokens_seen": 1387393024
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002927683049147443,
      "loss": 2.5707,
      "theoretical_loss": 3.5405248840727452,
      "tokens_seen": 1387458560
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002927582748244734,
      "loss": 2.7434,
      "theoretical_loss": 3.5405099179980013,
      "tokens_seen": 1387524096
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029274824473420264,
      "loss": 2.6813,
      "theoretical_loss": 3.5404949528280394,
      "tokens_seen": 1387589632
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002927382146439318,
      "loss": 2.73,
      "theoretical_loss": 3.5404799885627627,
      "tokens_seen": 1387655168
    },
    {
      "epoch": 4.06,
      "objective/train/docs_used": 1555725,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6399734020233154,
      "objective/train/theoretical_loss": 3.540465025202074,
      "objective/train/tokens_used": 1408180704,
      "theoretical_loss": 3.540465025202074,
      "tokens_seen": 1387720704
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.000292728184553661,
      "loss": 2.8601,
      "theoretical_loss": 3.540465025202074,
      "tokens_seen": 1387720704
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002927181544633902,
      "loss": 2.7116,
      "theoretical_loss": 3.540450062745875,
      "tokens_seen": 1387786240
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029270812437311936,
      "loss": 2.6407,
      "theoretical_loss": 3.540435101194069,
      "tokens_seen": 1387851776
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029269809428284855,
      "loss": 2.7734,
      "theoretical_loss": 3.5404201405465585,
      "tokens_seen": 1387917312
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002926880641925778,
      "loss": 2.8065,
      "theoretical_loss": 3.540405180803246,
      "tokens_seen": 1387982848
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002926780341023069,
      "loss": 2.4892,
      "theoretical_loss": 3.540390221964034,
      "tokens_seen": 1388048384
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029266800401203614,
      "loss": 2.615,
      "theoretical_loss": 3.5403752640288255,
      "tokens_seen": 1388113920
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029265797392176527,
      "loss": 2.6055,
      "theoretical_loss": 3.540360306997523,
      "tokens_seen": 1388179456
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002926479438314945,
      "loss": 2.7447,
      "theoretical_loss": 3.5403453508700298,
      "tokens_seen": 1388244992
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002926379137412237,
      "loss": 2.6932,
      "theoretical_loss": 3.5403303956462477,
      "tokens_seen": 1388310528
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029262788365095287,
      "loss": 2.5505,
      "theoretical_loss": 3.5403154413260802,
      "tokens_seen": 1388376064
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029261785356068205,
      "loss": 2.877,
      "theoretical_loss": 3.5403004879094295,
      "tokens_seen": 1388441600
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002926078234704113,
      "loss": 2.5885,
      "theoretical_loss": 3.5402855353961984,
      "tokens_seen": 1388507136
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002925977933801404,
      "loss": 2.6713,
      "theoretical_loss": 3.54027058378629,
      "tokens_seen": 1388572672
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029258776328986965,
      "loss": 2.496,
      "theoretical_loss": 3.540255633079607,
      "tokens_seen": 1388638208
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002925777331995988,
      "loss": 2.5759,
      "theoretical_loss": 3.5402406832760525,
      "tokens_seen": 1388703744
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.000292567703109328,
      "loss": 2.5836,
      "theoretical_loss": 3.540225734375529,
      "tokens_seen": 1388769280
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002925576730190572,
      "loss": 2.5895,
      "theoretical_loss": 3.540210786377939,
      "tokens_seen": 1388834816
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029254764292878637,
      "loss": 2.6178,
      "theoretical_loss": 3.5401958392831854,
      "tokens_seen": 1388900352
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029253761283851555,
      "loss": 2.4121,
      "theoretical_loss": 3.540180893091172,
      "tokens_seen": 1388965888
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029252758274824473,
      "loss": 2.6844,
      "theoretical_loss": 3.5401659478018006,
      "tokens_seen": 1389031424
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002925175526579739,
      "loss": 2.6171,
      "theoretical_loss": 3.540151003414975,
      "tokens_seen": 1389096960
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029250752256770315,
      "loss": 2.6658,
      "theoretical_loss": 3.5401360599305973,
      "tokens_seen": 1389162496
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002924974924774323,
      "loss": 2.774,
      "theoretical_loss": 3.540121117348571,
      "tokens_seen": 1389228032
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002924874623871615,
      "loss": 2.7155,
      "theoretical_loss": 3.540106175668799,
      "tokens_seen": 1389293568
    },
    {
      "epoch": 4.06,
      "objective/train/docs_used": 1556470,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5523812770843506,
      "objective/train/theoretical_loss": 3.540091234891184,
      "objective/train/tokens_used": 1409819104,
      "theoretical_loss": 3.540091234891184,
      "tokens_seen": 1389359104
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002924774322968907,
      "loss": 2.7275,
      "theoretical_loss": 3.540091234891184,
      "tokens_seen": 1389359104
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002924674022066199,
      "loss": 2.6922,
      "theoretical_loss": 3.540076295015629,
      "tokens_seen": 1389424640
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029245737211634906,
      "loss": 2.7791,
      "theoretical_loss": 3.5400613560420373,
      "tokens_seen": 1389490176
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029244734202607824,
      "loss": 2.8218,
      "theoretical_loss": 3.5400464179703115,
      "tokens_seen": 1389555712
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002924373119358074,
      "loss": 2.4633,
      "theoretical_loss": 3.540031480800355,
      "tokens_seen": 1389621248
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029242728184553665,
      "loss": 2.8368,
      "theoretical_loss": 3.5400165445320706,
      "tokens_seen": 1389686784
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002924172517552658,
      "loss": 2.6233,
      "theoretical_loss": 3.540001609165362,
      "tokens_seen": 1389752320
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.000292407221664995,
      "loss": 2.7833,
      "theoretical_loss": 3.539986674700131,
      "tokens_seen": 1389817856
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029239719157472414,
      "loss": 2.6574,
      "theoretical_loss": 3.5399717411362817,
      "tokens_seen": 1389883392
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002923871614844534,
      "loss": 2.7321,
      "theoretical_loss": 3.5399568084737165,
      "tokens_seen": 1389948928
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029237713139418256,
      "loss": 2.7022,
      "theoretical_loss": 3.539941876712339,
      "tokens_seen": 1390014464
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029236710130391174,
      "loss": 2.8063,
      "theoretical_loss": 3.5399269458520526,
      "tokens_seen": 1390080000
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002923570712136409,
      "loss": 2.786,
      "theoretical_loss": 3.53991201589276,
      "tokens_seen": 1390145536
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002923470411233701,
      "loss": 2.6066,
      "theoretical_loss": 3.539897086834364,
      "tokens_seen": 1390211072
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002923370110330993,
      "loss": 2.6857,
      "theoretical_loss": 3.539882158676768,
      "tokens_seen": 1390276608
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002923269809428285,
      "loss": 2.6082,
      "theoretical_loss": 3.5398672314198762,
      "tokens_seen": 1390342144
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029231695085255765,
      "loss": 2.6024,
      "theoretical_loss": 3.539852305063591,
      "tokens_seen": 1390407680
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002923069207622869,
      "loss": 2.6098,
      "theoretical_loss": 3.539837379607815,
      "tokens_seen": 1390473216
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029229689067201606,
      "loss": 2.5219,
      "theoretical_loss": 3.539822455052452,
      "tokens_seen": 1390538752
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029228686058174524,
      "loss": 2.6146,
      "theoretical_loss": 3.539807531397406,
      "tokens_seen": 1390604288
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002922768304914744,
      "loss": 2.6021,
      "theoretical_loss": 3.539792608642579,
      "tokens_seen": 1390669824
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002922668004012036,
      "loss": 2.8935,
      "theoretical_loss": 3.539777686787875,
      "tokens_seen": 1390735360
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002922567703109328,
      "loss": 2.566,
      "theoretical_loss": 3.539762765833197,
      "tokens_seen": 1390800896
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.000292246740220662,
      "loss": 2.7801,
      "theoretical_loss": 3.5397478457784484,
      "tokens_seen": 1390866432
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029223671013039115,
      "loss": 2.6456,
      "theoretical_loss": 3.5397329266235324,
      "tokens_seen": 1390931968
    },
    {
      "epoch": 4.06,
      "objective/train/docs_used": 1557630,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9702374935150146,
      "objective/train/theoretical_loss": 3.539718008368353,
      "objective/train/tokens_used": 1411457504,
      "theoretical_loss": 3.539718008368353,
      "tokens_seen": 1390997504
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002922266800401204,
      "loss": 2.6979,
      "theoretical_loss": 3.539718008368353,
      "tokens_seen": 1390997504
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002922166499498495,
      "loss": 2.7617,
      "theoretical_loss": 3.5397030910128127,
      "tokens_seen": 1391063040
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029220661985957875,
      "loss": 2.7569,
      "theoretical_loss": 3.539688174556815,
      "tokens_seen": 1391128576
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002921965897693079,
      "loss": 2.6994,
      "theoretical_loss": 3.5396732590002635,
      "tokens_seen": 1391194112
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002921865596790371,
      "loss": 2.7891,
      "theoretical_loss": 3.5396583443430623,
      "tokens_seen": 1391259648
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002921765295887663,
      "loss": 2.6115,
      "theoretical_loss": 3.539643430585113,
      "tokens_seen": 1391325184
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029216649949849547,
      "loss": 2.9884,
      "theoretical_loss": 3.5396285177263214,
      "tokens_seen": 1391390720
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029215646940822465,
      "loss": 2.7357,
      "theoretical_loss": 3.539613605766589,
      "tokens_seen": 1391456256
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002921464393179539,
      "loss": 2.7335,
      "theoretical_loss": 3.53959869470582,
      "tokens_seen": 1391521792
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.000292136409227683,
      "loss": 2.7324,
      "theoretical_loss": 3.5395837845439173,
      "tokens_seen": 1391587328
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029212637913741225,
      "loss": 2.8466,
      "theoretical_loss": 3.5395688752807857,
      "tokens_seen": 1391652864
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029211634904714143,
      "loss": 2.8641,
      "theoretical_loss": 3.539553966916327,
      "tokens_seen": 1391718400
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002921063189568706,
      "loss": 2.6256,
      "theoretical_loss": 3.539539059450447,
      "tokens_seen": 1391783936
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002920962888665998,
      "loss": 2.7322,
      "theoretical_loss": 3.539524152883047,
      "tokens_seen": 1391849472
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.000292086258776329,
      "loss": 2.6218,
      "theoretical_loss": 3.539509247214031,
      "tokens_seen": 1391915008
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029207622868605815,
      "loss": 2.6119,
      "theoretical_loss": 3.539494342443303,
      "tokens_seen": 1391980544
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002920661985957874,
      "loss": 2.8052,
      "theoretical_loss": 3.539479438570767,
      "tokens_seen": 1392046080
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002920561685055165,
      "loss": 2.4703,
      "theoretical_loss": 3.5394645355963257,
      "tokens_seen": 1392111616
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029204613841524575,
      "loss": 2.838,
      "theoretical_loss": 3.539449633519883,
      "tokens_seen": 1392177152
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029203610832497493,
      "loss": 2.8107,
      "theoretical_loss": 3.539434732341343,
      "tokens_seen": 1392242688
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002920260782347041,
      "loss": 2.6331,
      "theoretical_loss": 3.5394198320606094,
      "tokens_seen": 1392308224
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029201604814443335,
      "loss": 2.962,
      "theoretical_loss": 3.5394049326775847,
      "tokens_seen": 1392373760
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002920060180541625,
      "loss": 2.6808,
      "theoretical_loss": 3.5393900341921736,
      "tokens_seen": 1392439296
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002919959879638917,
      "loss": 2.746,
      "theoretical_loss": 3.5393751366042796,
      "tokens_seen": 1392504832
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002919859578736209,
      "loss": 2.7237,
      "theoretical_loss": 3.5393602399138056,
      "tokens_seen": 1392570368
    },
    {
      "epoch": 4.06,
      "objective/train/docs_used": 1558283,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4765679836273193,
      "objective/train/theoretical_loss": 3.5393453441206564,
      "objective/train/tokens_used": 1413095904,
      "theoretical_loss": 3.5393453441206564,
      "tokens_seen": 1392635904
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002919759277833501,
      "loss": 2.6099,
      "theoretical_loss": 3.5393453441206564,
      "tokens_seen": 1392635904
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029196589769307926,
      "loss": 2.64,
      "theoretical_loss": 3.5393304492247353,
      "tokens_seen": 1392701440
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029195586760280844,
      "loss": 2.7599,
      "theoretical_loss": 3.5393155552259463,
      "tokens_seen": 1392766976
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002919458375125376,
      "loss": 2.5263,
      "theoretical_loss": 3.5393006621241927,
      "tokens_seen": 1392832512
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029193580742226685,
      "loss": 2.4846,
      "theoretical_loss": 3.5392857699193785,
      "tokens_seen": 1392898048
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.000291925777331996,
      "loss": 2.7272,
      "theoretical_loss": 3.539270878611407,
      "tokens_seen": 1392963584
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002919157472417252,
      "loss": 2.8948,
      "theoretical_loss": 3.539255988200183,
      "tokens_seen": 1393029120
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029190571715145434,
      "loss": 2.6318,
      "theoretical_loss": 3.5392410986856095,
      "tokens_seen": 1393094656
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002918956870611836,
      "loss": 2.6671,
      "theoretical_loss": 3.5392262100675906,
      "tokens_seen": 1393160192
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029188565697091276,
      "loss": 2.6477,
      "theoretical_loss": 3.53921132234603,
      "tokens_seen": 1393225728
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029187562688064194,
      "loss": 2.6761,
      "theoretical_loss": 3.5391964355208323,
      "tokens_seen": 1393291264
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002918655967903711,
      "loss": 2.6871,
      "theoretical_loss": 3.5391815495919,
      "tokens_seen": 1393356800
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002918555667001003,
      "loss": 2.7094,
      "theoretical_loss": 3.539166664559138,
      "tokens_seen": 1393422336
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002918455366098295,
      "loss": 2.7393,
      "theoretical_loss": 3.5391517804224506,
      "tokens_seen": 1393487872
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002918355065195587,
      "loss": 2.6149,
      "theoretical_loss": 3.5391368971817405,
      "tokens_seen": 1393553408
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029182547642928785,
      "loss": 2.6424,
      "theoretical_loss": 3.539122014836912,
      "tokens_seen": 1393618944
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002918154463390171,
      "loss": 2.7817,
      "theoretical_loss": 3.539107133387869,
      "tokens_seen": 1393684480
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029180541624874626,
      "loss": 2.7797,
      "theoretical_loss": 3.539092252834517,
      "tokens_seen": 1393750016
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029179538615847544,
      "loss": 2.4839,
      "theoretical_loss": 3.5390773731767577,
      "tokens_seen": 1393815552
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002917853560682046,
      "loss": 2.8268,
      "theoretical_loss": 3.5390624944144964,
      "tokens_seen": 1393881088
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002917753259779338,
      "loss": 2.6579,
      "theoretical_loss": 3.5390476165476366,
      "tokens_seen": 1393946624
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.000291765295887663,
      "loss": 2.5451,
      "theoretical_loss": 3.5390327395760828,
      "tokens_seen": 1394012160
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002917552657973922,
      "loss": 2.6735,
      "theoretical_loss": 3.5390178634997382,
      "tokens_seen": 1394077696
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029174523570712135,
      "loss": 2.6657,
      "theoretical_loss": 3.539002988318508,
      "tokens_seen": 1394143232
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002917352056168506,
      "loss": 2.6527,
      "theoretical_loss": 3.5389881140322954,
      "tokens_seen": 1394208768
    },
    {
      "epoch": 4.06,
      "objective/train/docs_used": 1559602,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4316999912261963,
      "objective/train/theoretical_loss": 3.5389732406410044,
      "objective/train/tokens_used": 1414734304,
      "theoretical_loss": 3.5389732406410044,
      "tokens_seen": 1394274304
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002917251755265797,
      "loss": 2.7982,
      "theoretical_loss": 3.5389732406410044,
      "tokens_seen": 1394274304
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029171514543630895,
      "loss": 2.7828,
      "theoretical_loss": 3.53895836814454,
      "tokens_seen": 1394339840
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002917051153460381,
      "loss": 2.6216,
      "theoretical_loss": 3.5389434965428057,
      "tokens_seen": 1394405376
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002916950852557673,
      "loss": 2.6401,
      "theoretical_loss": 3.5389286258357053,
      "tokens_seen": 1394470912
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002916850551654965,
      "loss": 2.5921,
      "theoretical_loss": 3.5389137560231436,
      "tokens_seen": 1394536448
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029167502507522567,
      "loss": 2.5946,
      "theoretical_loss": 3.5388988871050246,
      "tokens_seen": 1394601984
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029166499498495485,
      "loss": 2.7326,
      "theoretical_loss": 3.538884019081252,
      "tokens_seen": 1394667520
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002916549648946841,
      "loss": 2.6635,
      "theoretical_loss": 3.5388691519517304,
      "tokens_seen": 1394733056
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002916449348044132,
      "loss": 2.6518,
      "theoretical_loss": 3.538854285716364,
      "tokens_seen": 1394798592
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029163490471414245,
      "loss": 2.7532,
      "theoretical_loss": 3.5388394203750573,
      "tokens_seen": 1394864128
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029162487462387163,
      "loss": 2.6184,
      "theoretical_loss": 3.5388245559277136,
      "tokens_seen": 1394929664
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002916148445336008,
      "loss": 2.7091,
      "theoretical_loss": 3.5388096923742385,
      "tokens_seen": 1394995200
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029160481444333,
      "loss": 2.4223,
      "theoretical_loss": 3.5387948297145346,
      "tokens_seen": 1395060736
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002915947843530592,
      "loss": 2.5134,
      "theoretical_loss": 3.5387799679485075,
      "tokens_seen": 1395126272
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029158475426278835,
      "loss": 2.5809,
      "theoretical_loss": 3.5387651070760606,
      "tokens_seen": 1395191808
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002915747241725176,
      "loss": 2.7848,
      "theoretical_loss": 3.538750247097099,
      "tokens_seen": 1395257344
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002915646940822467,
      "loss": 2.6923,
      "theoretical_loss": 3.538735388011527,
      "tokens_seen": 1395322880
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029155466399197595,
      "loss": 2.5315,
      "theoretical_loss": 3.5387205298192477,
      "tokens_seen": 1395388416
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002915446339017051,
      "loss": 2.7831,
      "theoretical_loss": 3.538705672520167,
      "tokens_seen": 1395453952
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002915346038114343,
      "loss": 2.5866,
      "theoretical_loss": 3.5386908161141886,
      "tokens_seen": 1395519488
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002915245737211635,
      "loss": 2.5707,
      "theoretical_loss": 3.538675960601217,
      "tokens_seen": 1395585024
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002915145436308927,
      "loss": 2.5683,
      "theoretical_loss": 3.5386611059811557,
      "tokens_seen": 1395650560
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029150451354062186,
      "loss": 2.8669,
      "theoretical_loss": 3.53864625225391,
      "tokens_seen": 1395716096
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002914944834503511,
      "loss": 2.4638,
      "theoretical_loss": 3.5386313994193843,
      "tokens_seen": 1395781632
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002914844533600802,
      "loss": 2.4921,
      "theoretical_loss": 3.538616547477483,
      "tokens_seen": 1395847168
    },
    {
      "epoch": 4.06,
      "objective/train/docs_used": 1560364,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.397831916809082,
      "objective/train/theoretical_loss": 3.5386016964281106,
      "objective/train/tokens_used": 1416372704,
      "theoretical_loss": 3.5386016964281106,
      "tokens_seen": 1395912704
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029147442326980946,
      "loss": 2.7208,
      "theoretical_loss": 3.5386016964281106,
      "tokens_seen": 1395912704
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002914643931795386,
      "loss": 2.6569,
      "theoretical_loss": 3.538586846271172,
      "tokens_seen": 1395978240
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002914543630892678,
      "loss": 2.8526,
      "theoretical_loss": 3.5385719970065703,
      "tokens_seen": 1396043776
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.000291444332998997,
      "loss": 2.7612,
      "theoretical_loss": 3.5385571486342107,
      "tokens_seen": 1396109312
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002914343029087262,
      "loss": 2.7004,
      "theoretical_loss": 3.5385423011539983,
      "tokens_seen": 1396174848
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029142427281845536,
      "loss": 2.6992,
      "theoretical_loss": 3.5385274545658367,
      "tokens_seen": 1396240384
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029141424272818454,
      "loss": 2.6948,
      "theoretical_loss": 3.5385126088696315,
      "tokens_seen": 1396305920
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002914042126379137,
      "loss": 2.4636,
      "theoretical_loss": 3.538497764065286,
      "tokens_seen": 1396371456
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029139418254764296,
      "loss": 2.7728,
      "theoretical_loss": 3.5384829201527057,
      "tokens_seen": 1396436992
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002913841524573721,
      "loss": 2.6798,
      "theoretical_loss": 3.538468077131795,
      "tokens_seen": 1396502528
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002913741223671013,
      "loss": 2.813,
      "theoretical_loss": 3.538453235002458,
      "tokens_seen": 1396568064
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029136409227683045,
      "loss": 2.5104,
      "theoretical_loss": 3.5384383937646,
      "tokens_seen": 1396633600
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002913540621865597,
      "loss": 2.4528,
      "theoretical_loss": 3.538423553418125,
      "tokens_seen": 1396699136
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029134403209628886,
      "loss": 2.546,
      "theoretical_loss": 3.538408713962938,
      "tokens_seen": 1396764672
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029133400200601805,
      "loss": 2.4749,
      "theoretical_loss": 3.5383938753989437,
      "tokens_seen": 1396830208
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002913239719157472,
      "loss": 2.5486,
      "theoretical_loss": 3.538379037726047,
      "tokens_seen": 1396895744
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029131394182547646,
      "loss": 2.7272,
      "theoretical_loss": 3.5383642009441516,
      "tokens_seen": 1396961280
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002913039117352056,
      "loss": 2.775,
      "theoretical_loss": 3.5383493650531634,
      "tokens_seen": 1397026816
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002912938816449348,
      "loss": 2.7766,
      "theoretical_loss": 3.5383345300529863,
      "tokens_seen": 1397092352
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.000291283851554664,
      "loss": 2.5114,
      "theoretical_loss": 3.5383196959435255,
      "tokens_seen": 1397157888
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002912738214643932,
      "loss": 2.717,
      "theoretical_loss": 3.5383048627246856,
      "tokens_seen": 1397223424
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002912637913741224,
      "loss": 2.6179,
      "theoretical_loss": 3.5382900303963707,
      "tokens_seen": 1397288960
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029125376128385155,
      "loss": 2.6235,
      "theoretical_loss": 3.5382751989584866,
      "tokens_seen": 1397354496
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002912437311935808,
      "loss": 2.628,
      "theoretical_loss": 3.538260368410938,
      "tokens_seen": 1397420032
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002912337011033099,
      "loss": 2.5759,
      "theoretical_loss": 3.538245538753629,
      "tokens_seen": 1397485568
    },
    {
      "epoch": 4.06,
      "objective/train/docs_used": 1561743,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9683220386505127,
      "objective/train/theoretical_loss": 3.5382307099864647,
      "objective/train/tokens_used": 1418011104,
      "theoretical_loss": 3.5382307099864647,
      "tokens_seen": 1397551104
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029122367101303915,
      "loss": 2.7218,
      "theoretical_loss": 3.5382307099864647,
      "tokens_seen": 1397551104
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029121364092276833,
      "loss": 2.5486,
      "theoretical_loss": 3.53821588210935,
      "tokens_seen": 1397616640
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002912036108324975,
      "loss": 2.5438,
      "theoretical_loss": 3.5382010551221903,
      "tokens_seen": 1397682176
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002911935807422267,
      "loss": 2.6942,
      "theoretical_loss": 3.5381862290248893,
      "tokens_seen": 1397747712
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029118355065195587,
      "loss": 2.6858,
      "theoretical_loss": 3.5381714038173526,
      "tokens_seen": 1397813248
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029117352056168505,
      "loss": 2.7571,
      "theoretical_loss": 3.538156579499485,
      "tokens_seen": 1397878784
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002911634904714143,
      "loss": 2.7165,
      "theoretical_loss": 3.538141756071192,
      "tokens_seen": 1397944320
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002911534603811434,
      "loss": 2.6838,
      "theoretical_loss": 3.538126933532377,
      "tokens_seen": 1398009856
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029114343029087265,
      "loss": 2.4973,
      "theoretical_loss": 3.5381121118829464,
      "tokens_seen": 1398075392
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029113340020060183,
      "loss": 2.541,
      "theoretical_loss": 3.538097291122804,
      "tokens_seen": 1398140928
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.000291123370110331,
      "loss": 2.6344,
      "theoretical_loss": 3.538082471251856,
      "tokens_seen": 1398206464
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002911133400200602,
      "loss": 2.5838,
      "theoretical_loss": 3.5380676522700067,
      "tokens_seen": 1398272000
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002911033099297894,
      "loss": 2.4584,
      "theoretical_loss": 3.5380528341771607,
      "tokens_seen": 1398337536
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029109327983951855,
      "loss": 2.5209,
      "theoretical_loss": 3.5380380169732235,
      "tokens_seen": 1398403072
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002910832497492478,
      "loss": 2.7778,
      "theoretical_loss": 3.5380232006581003,
      "tokens_seen": 1398468608
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002910732196589769,
      "loss": 2.719,
      "theoretical_loss": 3.538008385231696,
      "tokens_seen": 1398534144
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029106318956870615,
      "loss": 2.6171,
      "theoretical_loss": 3.537993570693915,
      "tokens_seen": 1398599680
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002910531594784353,
      "loss": 2.5908,
      "theoretical_loss": 3.5379787570446632,
      "tokens_seen": 1398665216
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002910431293881645,
      "loss": 2.6341,
      "theoretical_loss": 3.5379639442838453,
      "tokens_seen": 1398730752
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002910330992978937,
      "loss": 2.7537,
      "theoretical_loss": 3.5379491324113665,
      "tokens_seen": 1398796288
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002910230692076229,
      "loss": 2.8437,
      "theoretical_loss": 3.537934321427132,
      "tokens_seen": 1398861824
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029101303911735206,
      "loss": 2.6106,
      "theoretical_loss": 3.537919511331047,
      "tokens_seen": 1398927360
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002910030090270813,
      "loss": 2.6991,
      "theoretical_loss": 3.537904702123016,
      "tokens_seen": 1398992896
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002909929789368104,
      "loss": 2.5194,
      "theoretical_loss": 3.537889893802945,
      "tokens_seen": 1399058432
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029098294884653966,
      "loss": 2.6954,
      "theoretical_loss": 3.537875086370738,
      "tokens_seen": 1399123968
    },
    {
      "epoch": 4.06,
      "objective/train/docs_used": 1562396,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.784968137741089,
      "objective/train/theoretical_loss": 3.537860279826302,
      "objective/train/tokens_used": 1419649504,
      "theoretical_loss": 3.537860279826302,
      "tokens_seen": 1399189504
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002909729187562688,
      "loss": 2.6959,
      "theoretical_loss": 3.537860279826302,
      "tokens_seen": 1399189504
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.000290962888665998,
      "loss": 2.7707,
      "theoretical_loss": 3.53784547416954,
      "tokens_seen": 1399255040
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002909528585757272,
      "loss": 2.6713,
      "theoretical_loss": 3.5378306694003596,
      "tokens_seen": 1399320576
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002909428284854564,
      "loss": 2.5541,
      "theoretical_loss": 3.537815865518664,
      "tokens_seen": 1399386112
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029093279839518556,
      "loss": 2.7361,
      "theoretical_loss": 3.537801062524359,
      "tokens_seen": 1399451648
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029092276830491474,
      "loss": 2.5931,
      "theoretical_loss": 3.5377862604173504,
      "tokens_seen": 1399517184
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002909127382146439,
      "loss": 2.6053,
      "theoretical_loss": 3.5377714591975433,
      "tokens_seen": 1399582720
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029090270812437316,
      "loss": 2.6946,
      "theoretical_loss": 3.5377566588648426,
      "tokens_seen": 1399648256
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002908926780341023,
      "loss": 2.6109,
      "theoretical_loss": 3.537741859419154,
      "tokens_seen": 1399713792
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002908826479438315,
      "loss": 2.5046,
      "theoretical_loss": 3.537727060860383,
      "tokens_seen": 1399779328
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029087261785356065,
      "loss": 2.7201,
      "theoretical_loss": 3.5377122631884337,
      "tokens_seen": 1399844864
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002908625877632899,
      "loss": 2.5966,
      "theoretical_loss": 3.5376974664032126,
      "tokens_seen": 1399910400
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029085255767301906,
      "loss": 2.5142,
      "theoretical_loss": 3.5376826705046254,
      "tokens_seen": 1399975936
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029084252758274825,
      "loss": 2.3679,
      "theoretical_loss": 3.537667875492576,
      "tokens_seen": 1400041472
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002908324974924774,
      "loss": 2.6371,
      "theoretical_loss": 3.537653081366971,
      "tokens_seen": 1400107008
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029082246740220666,
      "loss": 2.6135,
      "theoretical_loss": 3.5376382881277153,
      "tokens_seen": 1400172544
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002908124373119358,
      "loss": 2.5081,
      "theoretical_loss": 3.5376234957747146,
      "tokens_seen": 1400238080
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.000290802407221665,
      "loss": 2.6574,
      "theoretical_loss": 3.537608704307874,
      "tokens_seen": 1400303616
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029079237713139415,
      "loss": 2.4363,
      "theoretical_loss": 3.537593913727099,
      "tokens_seen": 1400369152
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002907823470411234,
      "loss": 2.6265,
      "theoretical_loss": 3.5375791240322956,
      "tokens_seen": 1400434688
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029077231695085257,
      "loss": 2.8114,
      "theoretical_loss": 3.5375643352233683,
      "tokens_seen": 1400500224
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029076228686058175,
      "loss": 2.444,
      "theoretical_loss": 3.537549547300223,
      "tokens_seen": 1400565760
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029075225677031093,
      "loss": 2.6148,
      "theoretical_loss": 3.537534760262766,
      "tokens_seen": 1400631296
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002907422266800401,
      "loss": 2.5593,
      "theoretical_loss": 3.537519974110902,
      "tokens_seen": 1400696832
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002907321965897693,
      "loss": 2.5544,
      "theoretical_loss": 3.537505188844536,
      "tokens_seen": 1400762368
    },
    {
      "epoch": 4.06,
      "objective/train/docs_used": 1563023,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.582470655441284,
      "objective/train/theoretical_loss": 3.5374904044635747,
      "objective/train/tokens_used": 1421287904,
      "theoretical_loss": 3.5374904044635747,
      "tokens_seen": 1400827904
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029072216649949853,
      "loss": 2.6132,
      "theoretical_loss": 3.5374904044635747,
      "tokens_seen": 1400827904
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029071213640922765,
      "loss": 2.778,
      "theoretical_loss": 3.537475620967923,
      "tokens_seen": 1400893440
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002907021063189569,
      "loss": 2.6992,
      "theoretical_loss": 3.5374608383574873,
      "tokens_seen": 1400958976
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.000290692076228686,
      "loss": 2.7268,
      "theoretical_loss": 3.537446056632172,
      "tokens_seen": 1401024512
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029068204613841525,
      "loss": 2.3954,
      "theoretical_loss": 3.5374312757918833,
      "tokens_seen": 1401090048
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029067201604814443,
      "loss": 2.7221,
      "theoretical_loss": 3.537416495836527,
      "tokens_seen": 1401155584
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002906619859578736,
      "loss": 2.4826,
      "theoretical_loss": 3.5374017167660075,
      "tokens_seen": 1401221120
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002906519558676028,
      "loss": 2.5533,
      "theoretical_loss": 3.537386938580233,
      "tokens_seen": 1401286656
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029064192577733203,
      "loss": 2.7717,
      "theoretical_loss": 3.537372161279106,
      "tokens_seen": 1401352192
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.00029063189568706116,
      "loss": 2.566,
      "theoretical_loss": 3.537357384862535,
      "tokens_seen": 1401417728
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002906218655967904,
      "loss": 2.6829,
      "theoretical_loss": 3.5373426093304237,
      "tokens_seen": 1401483264
    },
    {
      "epoch": 4.06,
      "learning_rate": 0.0002906118355065195,
      "loss": 2.7533,
      "theoretical_loss": 3.537327834682679,
      "tokens_seen": 1401548800
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00029060180541624876,
      "loss": 2.3787,
      "theoretical_loss": 3.537313060919206,
      "tokens_seen": 1401614336
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00029059177532597794,
      "loss": 2.3038,
      "theoretical_loss": 3.5372982880399104,
      "tokens_seen": 1401679872
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002905817452357071,
      "loss": 2.4445,
      "theoretical_loss": 3.537283516044698,
      "tokens_seen": 1401745408
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002905717151454363,
      "loss": 2.7568,
      "theoretical_loss": 3.5372687449334754,
      "tokens_seen": 1401810944
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002905616850551655,
      "loss": 2.6059,
      "theoretical_loss": 3.5372539747061476,
      "tokens_seen": 1401876480
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00029055165496489466,
      "loss": 2.5998,
      "theoretical_loss": 3.53723920536262,
      "tokens_seen": 1401942016
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002905416248746239,
      "loss": 2.6002,
      "theoretical_loss": 3.5372244369027994,
      "tokens_seen": 1402007552
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002905315947843531,
      "loss": 2.7421,
      "theoretical_loss": 3.537209669326591,
      "tokens_seen": 1402073088
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00029052156469408226,
      "loss": 2.6018,
      "theoretical_loss": 3.5371949026339005,
      "tokens_seen": 1402138624
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002905115346038115,
      "loss": 2.7309,
      "theoretical_loss": 3.5371801368246345,
      "tokens_seen": 1402204160
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002905015045135406,
      "loss": 2.7116,
      "theoretical_loss": 3.5371653718986975,
      "tokens_seen": 1402269696
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00029049147442326986,
      "loss": 2.7406,
      "theoretical_loss": 3.537150607855997,
      "tokens_seen": 1402335232
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.000290481444332999,
      "loss": 2.6909,
      "theoretical_loss": 3.537135844696438,
      "tokens_seen": 1402400768
    },
    {
      "epoch": 4.07,
      "objective/train/docs_used": 1564219,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.3950443267822266,
      "objective/train/theoretical_loss": 3.5371210824199264,
      "objective/train/tokens_used": 1422926304,
      "theoretical_loss": 3.5371210824199264,
      "tokens_seen": 1402466304
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002904714142427282,
      "loss": 2.5619,
      "theoretical_loss": 3.5371210824199264,
      "tokens_seen": 1402466304
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002904613841524574,
      "loss": 2.7336,
      "theoretical_loss": 3.5371063210263682,
      "tokens_seen": 1402531840
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002904513540621866,
      "loss": 2.7499,
      "theoretical_loss": 3.5370915605156696,
      "tokens_seen": 1402597376
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00029044132397191576,
      "loss": 2.5718,
      "theoretical_loss": 3.537076800887736,
      "tokens_seen": 1402662912
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00029043129388164494,
      "loss": 2.5744,
      "theoretical_loss": 3.537062042142474,
      "tokens_seen": 1402728448
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002904212637913741,
      "loss": 2.651,
      "theoretical_loss": 3.5370472842797893,
      "tokens_seen": 1402793984
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00029041123370110336,
      "loss": 2.6257,
      "theoretical_loss": 3.537032527299588,
      "tokens_seen": 1402859520
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002904012036108325,
      "loss": 2.9672,
      "theoretical_loss": 3.537017771201776,
      "tokens_seen": 1402925056
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002903911735205617,
      "loss": 2.7034,
      "theoretical_loss": 3.537003015986259,
      "tokens_seen": 1402990592
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00029038114343029085,
      "loss": 2.8006,
      "theoretical_loss": 3.536988261652943,
      "tokens_seen": 1403056128
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002903711133400201,
      "loss": 2.7806,
      "theoretical_loss": 3.536973508201735,
      "tokens_seen": 1403121664
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00029036108324974926,
      "loss": 2.7626,
      "theoretical_loss": 3.5369587556325404,
      "tokens_seen": 1403187200
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00029035105315947845,
      "loss": 2.884,
      "theoretical_loss": 3.536944003945265,
      "tokens_seen": 1403252736
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002903410230692076,
      "loss": 2.6588,
      "theoretical_loss": 3.5369292531398155,
      "tokens_seen": 1403318272
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00029033099297893686,
      "loss": 2.5644,
      "theoretical_loss": 3.5369145032160976,
      "tokens_seen": 1403383808
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.000290320962888666,
      "loss": 2.5827,
      "theoretical_loss": 3.536899754174018,
      "tokens_seen": 1403449344
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002903109327983952,
      "loss": 2.5899,
      "theoretical_loss": 3.536885006013482,
      "tokens_seen": 1403514880
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00029030090270812435,
      "loss": 2.7309,
      "theoretical_loss": 3.536870258734396,
      "tokens_seen": 1403580416
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002902908726178536,
      "loss": 2.5328,
      "theoretical_loss": 3.5368555123366665,
      "tokens_seen": 1403645952
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00029028084252758277,
      "loss": 2.8273,
      "theoretical_loss": 3.5368407668202,
      "tokens_seen": 1403711488
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00029027081243731195,
      "loss": 2.8104,
      "theoretical_loss": 3.5368260221849015,
      "tokens_seen": 1403777024
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00029026078234704113,
      "loss": 2.7349,
      "theoretical_loss": 3.536811278430678,
      "tokens_seen": 1403842560
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002902507522567703,
      "loss": 2.7155,
      "theoretical_loss": 3.536796535557436,
      "tokens_seen": 1403908096
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002902407221664995,
      "loss": 2.5073,
      "theoretical_loss": 3.536781793565081,
      "tokens_seen": 1403973632
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00029023069207622873,
      "loss": 2.5849,
      "theoretical_loss": 3.53676705245352,
      "tokens_seen": 1404039168
    },
    {
      "epoch": 4.07,
      "objective/train/docs_used": 1565469,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9763519763946533,
      "objective/train/theoretical_loss": 3.5367523122226583,
      "objective/train/tokens_used": 1424564704,
      "theoretical_loss": 3.5367523122226583,
      "tokens_seen": 1404104704
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00029022066198595785,
      "loss": 2.7733,
      "theoretical_loss": 3.5367523122226583,
      "tokens_seen": 1404104704
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002902106318956871,
      "loss": 2.7111,
      "theoretical_loss": 3.5367375728724033,
      "tokens_seen": 1404170240
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002902006018054162,
      "loss": 2.6443,
      "theoretical_loss": 3.5367228344026604,
      "tokens_seen": 1404235776
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00029019057171514545,
      "loss": 2.7371,
      "theoretical_loss": 3.5367080968133364,
      "tokens_seen": 1404301312
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00029018054162487463,
      "loss": 2.7286,
      "theoretical_loss": 3.5366933601043375,
      "tokens_seen": 1404366848
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002901705115346038,
      "loss": 2.522,
      "theoretical_loss": 3.53667862427557,
      "tokens_seen": 1404432384
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.000290160481444333,
      "loss": 2.7689,
      "theoretical_loss": 3.5366638893269404,
      "tokens_seen": 1404497920
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00029015045135406223,
      "loss": 2.6427,
      "theoretical_loss": 3.5366491552583548,
      "tokens_seen": 1404563456
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00029014042126379136,
      "loss": 2.4538,
      "theoretical_loss": 3.53663442206972,
      "tokens_seen": 1404628992
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002901303911735206,
      "loss": 2.6588,
      "theoretical_loss": 3.5366196897609417,
      "tokens_seen": 1404694528
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002901203610832497,
      "loss": 2.7638,
      "theoretical_loss": 3.536604958331927,
      "tokens_seen": 1404760064
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00029011033099297896,
      "loss": 2.5228,
      "theoretical_loss": 3.5365902277825816,
      "tokens_seen": 1404825600
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00029010030090270814,
      "loss": 2.5804,
      "theoretical_loss": 3.5365754981128132,
      "tokens_seen": 1404891136
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002900902708124373,
      "loss": 2.6405,
      "theoretical_loss": 3.5365607693225267,
      "tokens_seen": 1404956672
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002900802407221665,
      "loss": 2.4659,
      "theoretical_loss": 3.5365460414116296,
      "tokens_seen": 1405022208
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002900702106318957,
      "loss": 2.5059,
      "theoretical_loss": 3.536531314380028,
      "tokens_seen": 1405087744
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00029006018054162486,
      "loss": 2.4888,
      "theoretical_loss": 3.5365165882276286,
      "tokens_seen": 1405153280
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002900501504513541,
      "loss": 2.6572,
      "theoretical_loss": 3.5365018629543377,
      "tokens_seen": 1405218816
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002900401203610832,
      "loss": 2.7108,
      "theoretical_loss": 3.5364871385600622,
      "tokens_seen": 1405284352
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00029003009027081246,
      "loss": 2.6045,
      "theoretical_loss": 3.5364724150447078,
      "tokens_seen": 1405349888
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002900200601805416,
      "loss": 2.7128,
      "theoretical_loss": 3.5364576924081823,
      "tokens_seen": 1405415424
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002900100300902708,
      "loss": 2.6778,
      "theoretical_loss": 3.5364429706503913,
      "tokens_seen": 1405480960
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00029,
      "loss": 2.6373,
      "theoretical_loss": 3.5364282497712414,
      "tokens_seen": 1405546496
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002899899699097292,
      "loss": 2.5146,
      "theoretical_loss": 3.5364135297706394,
      "tokens_seen": 1405612032
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028997993981945836,
      "loss": 2.2929,
      "theoretical_loss": 3.5363988106484925,
      "tokens_seen": 1405677568
    },
    {
      "epoch": 4.07,
      "objective/train/docs_used": 1565832,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.80586838722229,
      "objective/train/theoretical_loss": 3.5363840924047065,
      "objective/train/tokens_used": 1426203104,
      "theoretical_loss": 3.5363840924047065,
      "tokens_seen": 1405743104
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002899699097291876,
      "loss": 2.7417,
      "theoretical_loss": 3.5363840924047065,
      "tokens_seen": 1405743104
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002899598796389167,
      "loss": 2.5818,
      "theoretical_loss": 3.5363693750391882,
      "tokens_seen": 1405808640
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028994984954864596,
      "loss": 2.6419,
      "theoretical_loss": 3.5363546585518444,
      "tokens_seen": 1405874176
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002899398194583751,
      "loss": 2.5231,
      "theoretical_loss": 3.5363399429425817,
      "tokens_seen": 1405939712
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002899297893681043,
      "loss": 2.6878,
      "theoretical_loss": 3.5363252282113073,
      "tokens_seen": 1406005248
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002899197592778335,
      "loss": 2.837,
      "theoretical_loss": 3.536310514357927,
      "tokens_seen": 1406070784
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002899097291875627,
      "loss": 2.7435,
      "theoretical_loss": 3.536295801382348,
      "tokens_seen": 1406136320
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028989969909729187,
      "loss": 2.6983,
      "theoretical_loss": 3.5362810892844774,
      "tokens_seen": 1406201856
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028988966900702105,
      "loss": 2.5337,
      "theoretical_loss": 3.5362663780642207,
      "tokens_seen": 1406267392
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028987963891675023,
      "loss": 2.5224,
      "theoretical_loss": 3.536251667721486,
      "tokens_seen": 1406332928
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028986960882647946,
      "loss": 2.5878,
      "theoretical_loss": 3.536236958256179,
      "tokens_seen": 1406398464
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002898595787362086,
      "loss": 2.6069,
      "theoretical_loss": 3.5362222496682074,
      "tokens_seen": 1406464000
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002898495486459378,
      "loss": 2.5348,
      "theoretical_loss": 3.536207541957477,
      "tokens_seen": 1406529536
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.000289839518555667,
      "loss": 2.7274,
      "theoretical_loss": 3.536192835123896,
      "tokens_seen": 1406595072
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002898294884653962,
      "loss": 2.6032,
      "theoretical_loss": 3.53617812916737,
      "tokens_seen": 1406660608
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028981945837512537,
      "loss": 2.5445,
      "theoretical_loss": 3.536163424087807,
      "tokens_seen": 1406726144
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028980942828485455,
      "loss": 2.7627,
      "theoretical_loss": 3.5361487198851123,
      "tokens_seen": 1406791680
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028979939819458373,
      "loss": 2.6414,
      "theoretical_loss": 3.5361340165591937,
      "tokens_seen": 1406857216
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028978936810431297,
      "loss": 2.8794,
      "theoretical_loss": 3.536119314109958,
      "tokens_seen": 1406922752
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028977933801404215,
      "loss": 2.6667,
      "theoretical_loss": 3.5361046125373123,
      "tokens_seen": 1406988288
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028976930792377133,
      "loss": 2.5148,
      "theoretical_loss": 3.536089911841163,
      "tokens_seen": 1407053824
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002897592778335005,
      "loss": 2.489,
      "theoretical_loss": 3.5360752120214167,
      "tokens_seen": 1407119360
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002897492477432297,
      "loss": 2.8873,
      "theoretical_loss": 3.536060513077982,
      "tokens_seen": 1407184896
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028973921765295893,
      "loss": 2.7641,
      "theoretical_loss": 3.5360458150107643,
      "tokens_seen": 1407250432
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028972918756268805,
      "loss": 2.5708,
      "theoretical_loss": 3.536031117819671,
      "tokens_seen": 1407315968
    },
    {
      "epoch": 4.07,
      "objective/train/docs_used": 1567098,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.710190773010254,
      "objective/train/theoretical_loss": 3.536016421504609,
      "objective/train/tokens_used": 1427841504,
      "theoretical_loss": 3.536016421504609,
      "tokens_seen": 1407381504
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002897191574724173,
      "loss": 2.6155,
      "theoretical_loss": 3.536016421504609,
      "tokens_seen": 1407381504
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002897091273821464,
      "loss": 2.5796,
      "theoretical_loss": 3.5360017260654857,
      "tokens_seen": 1407447040
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028969909729187565,
      "loss": 2.5956,
      "theoretical_loss": 3.5359870315022075,
      "tokens_seen": 1407512576
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028968906720160483,
      "loss": 2.8037,
      "theoretical_loss": 3.535972337814682,
      "tokens_seen": 1407578112
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.000289679037111334,
      "loss": 2.4828,
      "theoretical_loss": 3.5359576450028163,
      "tokens_seen": 1407643648
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002896690070210632,
      "loss": 2.8295,
      "theoretical_loss": 3.5359429530665167,
      "tokens_seen": 1407709184
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028965897693079243,
      "loss": 2.5858,
      "theoretical_loss": 3.5359282620056907,
      "tokens_seen": 1407774720
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028964894684052156,
      "loss": 2.9195,
      "theoretical_loss": 3.5359135718202452,
      "tokens_seen": 1407840256
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002896389167502508,
      "loss": 2.532,
      "theoretical_loss": 3.5358988825100877,
      "tokens_seen": 1407905792
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002896288866599799,
      "loss": 2.8376,
      "theoretical_loss": 3.535884194075125,
      "tokens_seen": 1407971328
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028961885656970916,
      "loss": 2.7427,
      "theoretical_loss": 3.5358695065152643,
      "tokens_seen": 1408036864
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028960882647943834,
      "loss": 2.674,
      "theoretical_loss": 3.535854819830413,
      "tokens_seen": 1408102400
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002895987963891675,
      "loss": 2.6753,
      "theoretical_loss": 3.535840134020478,
      "tokens_seen": 1408167936
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002895887662988967,
      "loss": 2.6783,
      "theoretical_loss": 3.5358254490853662,
      "tokens_seen": 1408233472
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002895787362086259,
      "loss": 2.6066,
      "theoretical_loss": 3.5358107650249853,
      "tokens_seen": 1408299008
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028956870611835506,
      "loss": 2.7882,
      "theoretical_loss": 3.535796081839242,
      "tokens_seen": 1408364544
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002895586760280843,
      "loss": 2.7076,
      "theoretical_loss": 3.5357813995280436,
      "tokens_seen": 1408430080
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002895486459378134,
      "loss": 2.5654,
      "theoretical_loss": 3.5357667180912973,
      "tokens_seen": 1408495616
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028953861584754266,
      "loss": 2.5214,
      "theoretical_loss": 3.535752037528911,
      "tokens_seen": 1408561152
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002895285857572718,
      "loss": 2.8717,
      "theoretical_loss": 3.535737357840791,
      "tokens_seen": 1408626688
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.000289518555667001,
      "loss": 2.5479,
      "theoretical_loss": 3.5357226790268452,
      "tokens_seen": 1408692224
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002895085255767302,
      "loss": 2.418,
      "theoretical_loss": 3.535708001086981,
      "tokens_seen": 1408757760
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002894984954864594,
      "loss": 2.6083,
      "theoretical_loss": 3.5356933240211053,
      "tokens_seen": 1408823296
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028948846539618856,
      "loss": 2.6263,
      "theoretical_loss": 3.5356786478291253,
      "tokens_seen": 1408888832
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002894784353059178,
      "loss": 2.653,
      "theoretical_loss": 3.535663972510948,
      "tokens_seen": 1408954368
    },
    {
      "epoch": 4.07,
      "objective/train/docs_used": 1567854,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5623223781585693,
      "objective/train/theoretical_loss": 3.535649298066482,
      "objective/train/tokens_used": 1429479904,
      "theoretical_loss": 3.535649298066482,
      "tokens_seen": 1409019904
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002894684052156469,
      "loss": 2.599,
      "theoretical_loss": 3.535649298066482,
      "tokens_seen": 1409019904
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028945837512537616,
      "loss": 2.7912,
      "theoretical_loss": 3.5356346244956334,
      "tokens_seen": 1409085440
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002894483450351053,
      "loss": 2.6228,
      "theoretical_loss": 3.53561995179831,
      "tokens_seen": 1409150976
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002894383149448345,
      "loss": 2.5917,
      "theoretical_loss": 3.53560527997442,
      "tokens_seen": 1409216512
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002894282848545637,
      "loss": 2.544,
      "theoretical_loss": 3.535590609023869,
      "tokens_seen": 1409282048
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002894182547642929,
      "loss": 2.4938,
      "theoretical_loss": 3.535575938946566,
      "tokens_seen": 1409347584
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028940822467402207,
      "loss": 2.6501,
      "theoretical_loss": 3.5355612697424177,
      "tokens_seen": 1409413120
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028939819458375125,
      "loss": 2.5533,
      "theoretical_loss": 3.5355466014113315,
      "tokens_seen": 1409478656
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028938816449348043,
      "loss": 2.7157,
      "theoretical_loss": 3.535531933953215,
      "tokens_seen": 1409544192
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028937813440320966,
      "loss": 2.5321,
      "theoretical_loss": 3.5355172673679762,
      "tokens_seen": 1409609728
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002893681043129388,
      "loss": 2.5582,
      "theoretical_loss": 3.5355026016555215,
      "tokens_seen": 1409675264
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.000289358074222668,
      "loss": 2.6674,
      "theoretical_loss": 3.535487936815759,
      "tokens_seen": 1409740800
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002893480441323972,
      "loss": 2.6364,
      "theoretical_loss": 3.5354732728485967,
      "tokens_seen": 1409806336
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002893380140421264,
      "loss": 2.3849,
      "theoretical_loss": 3.535458609753941,
      "tokens_seen": 1409871872
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028932798395185557,
      "loss": 2.6955,
      "theoretical_loss": 3.5354439475317,
      "tokens_seen": 1409937408
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028931795386158475,
      "loss": 2.6259,
      "theoretical_loss": 3.5354292861817815,
      "tokens_seen": 1410002944
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028930792377131393,
      "loss": 2.6021,
      "theoretical_loss": 3.5354146257040924,
      "tokens_seen": 1410068480
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028929789368104317,
      "loss": 2.7806,
      "theoretical_loss": 3.535399966098541,
      "tokens_seen": 1410134016
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002892878635907723,
      "loss": 2.5702,
      "theoretical_loss": 3.5353853073650345,
      "tokens_seen": 1410199552
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028927783350050153,
      "loss": 2.6023,
      "theoretical_loss": 3.5353706495034807,
      "tokens_seen": 1410265088
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028926780341023066,
      "loss": 2.4183,
      "theoretical_loss": 3.5353559925137867,
      "tokens_seen": 1410330624
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002892577733199599,
      "loss": 2.4087,
      "theoretical_loss": 3.5353413363958612,
      "tokens_seen": 1410396160
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002892477432296891,
      "loss": 2.7805,
      "theoretical_loss": 3.535326681149611,
      "tokens_seen": 1410461696
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028923771313941825,
      "loss": 2.7112,
      "theoretical_loss": 3.5353120267749434,
      "tokens_seen": 1410527232
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028922768304914744,
      "loss": 2.6916,
      "theoretical_loss": 3.5352973732717667,
      "tokens_seen": 1410592768
    },
    {
      "epoch": 4.07,
      "objective/train/docs_used": 1569260,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.391869068145752,
      "objective/train/theoretical_loss": 3.5352827206399886,
      "objective/train/tokens_used": 1431118304,
      "theoretical_loss": 3.5352827206399886,
      "tokens_seen": 1410658304
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002892176529588766,
      "loss": 2.3331,
      "theoretical_loss": 3.5352827206399886,
      "tokens_seen": 1410658304
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002892076228686058,
      "loss": 2.6157,
      "theoretical_loss": 3.5352680688795166,
      "tokens_seen": 1410723840
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028919759277833503,
      "loss": 2.5459,
      "theoretical_loss": 3.535253417990259,
      "tokens_seen": 1410789376
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028918756268806416,
      "loss": 2.6511,
      "theoretical_loss": 3.5352387679721224,
      "tokens_seen": 1410854912
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002891775325977934,
      "loss": 2.5784,
      "theoretical_loss": 3.535224118825015,
      "tokens_seen": 1410920448
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002891675025075226,
      "loss": 2.603,
      "theoretical_loss": 3.5352094705488453,
      "tokens_seen": 1410985984
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028915747241725176,
      "loss": 2.7896,
      "theoretical_loss": 3.53519482314352,
      "tokens_seen": 1411051520
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028914744232698094,
      "loss": 2.6002,
      "theoretical_loss": 3.5351801766089483,
      "tokens_seen": 1411117056
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002891374122367101,
      "loss": 2.6282,
      "theoretical_loss": 3.5351655309450365,
      "tokens_seen": 1411182592
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002891273821464393,
      "loss": 2.3283,
      "theoretical_loss": 3.535150886151693,
      "tokens_seen": 1411248128
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028911735205616854,
      "loss": 2.6072,
      "theoretical_loss": 3.5351362422288255,
      "tokens_seen": 1411313664
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028910732196589766,
      "loss": 2.5544,
      "theoretical_loss": 3.5351215991763416,
      "tokens_seen": 1411379200
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002890972918756269,
      "loss": 2.6091,
      "theoretical_loss": 3.5351069569941505,
      "tokens_seen": 1411444736
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.000289087261785356,
      "loss": 2.6773,
      "theoretical_loss": 3.5350923156821583,
      "tokens_seen": 1411510272
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028907723169508526,
      "loss": 2.5854,
      "theoretical_loss": 3.5350776752402737,
      "tokens_seen": 1411575808
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028906720160481444,
      "loss": 2.7125,
      "theoretical_loss": 3.535063035668405,
      "tokens_seen": 1411641344
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002890571715145436,
      "loss": 2.5343,
      "theoretical_loss": 3.5350483969664594,
      "tokens_seen": 1411706880
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002890471414242728,
      "loss": 2.5468,
      "theoretical_loss": 3.5350337591343455,
      "tokens_seen": 1411772416
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.000289037111334002,
      "loss": 2.6072,
      "theoretical_loss": 3.5350191221719705,
      "tokens_seen": 1411837952
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002890270812437312,
      "loss": 2.7006,
      "theoretical_loss": 3.5350044860792424,
      "tokens_seen": 1411903488
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002890170511534604,
      "loss": 2.5986,
      "theoretical_loss": 3.53498985085607,
      "tokens_seen": 1411969024
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002890070210631896,
      "loss": 2.6427,
      "theoretical_loss": 3.5349752165023602,
      "tokens_seen": 1412034560
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028899699097291876,
      "loss": 2.6691,
      "theoretical_loss": 3.534960583018022,
      "tokens_seen": 1412100096
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.000288986960882648,
      "loss": 2.7271,
      "theoretical_loss": 3.534945950402963,
      "tokens_seen": 1412165632
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002889769307923771,
      "loss": 2.7876,
      "theoretical_loss": 3.534931318657091,
      "tokens_seen": 1412231168
    },
    {
      "epoch": 4.07,
      "objective/train/docs_used": 1569919,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.615917921066284,
      "objective/train/theoretical_loss": 3.5349166877803144,
      "objective/train/tokens_used": 1432756704,
      "theoretical_loss": 3.5349166877803144,
      "tokens_seen": 1412296704
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028896690070210636,
      "loss": 2.5592,
      "theoretical_loss": 3.5349166877803144,
      "tokens_seen": 1412296704
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002889568706118355,
      "loss": 2.5736,
      "theoretical_loss": 3.534902057772541,
      "tokens_seen": 1412362240
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002889468405215647,
      "loss": 2.8109,
      "theoretical_loss": 3.5348874286336787,
      "tokens_seen": 1412427776
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002889368104312939,
      "loss": 2.6942,
      "theoretical_loss": 3.534872800363636,
      "tokens_seen": 1412493312
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002889267803410231,
      "loss": 2.7169,
      "theoretical_loss": 3.5348581729623207,
      "tokens_seen": 1412558848
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028891675025075227,
      "loss": 2.6482,
      "theoretical_loss": 3.534843546429641,
      "tokens_seen": 1412624384
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028890672016048145,
      "loss": 2.917,
      "theoretical_loss": 3.534828920765505,
      "tokens_seen": 1412689920
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028889669007021063,
      "loss": 2.4384,
      "theoretical_loss": 3.534814295969821,
      "tokens_seen": 1412755456
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028888665997993986,
      "loss": 2.8092,
      "theoretical_loss": 3.5347996720424972,
      "tokens_seen": 1412820992
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.000288876629889669,
      "loss": 2.6139,
      "theoretical_loss": 3.5347850489834416,
      "tokens_seen": 1412886528
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028886659979939823,
      "loss": 2.747,
      "theoretical_loss": 3.534770426792562,
      "tokens_seen": 1412952064
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002888565697091274,
      "loss": 2.7182,
      "theoretical_loss": 3.534755805469767,
      "tokens_seen": 1413017600
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002888465396188566,
      "loss": 2.7166,
      "theoretical_loss": 3.534741185014965,
      "tokens_seen": 1413083136
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028883650952858577,
      "loss": 2.515,
      "theoretical_loss": 3.5347265654280635,
      "tokens_seen": 1413148672
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028882647943831495,
      "loss": 2.7623,
      "theoretical_loss": 3.534711946708972,
      "tokens_seen": 1413214208
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028881644934804413,
      "loss": 2.6792,
      "theoretical_loss": 3.5346973288575976,
      "tokens_seen": 1413279744
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028880641925777337,
      "loss": 2.7728,
      "theoretical_loss": 3.534682711873849,
      "tokens_seen": 1413345280
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002887963891675025,
      "loss": 2.6048,
      "theoretical_loss": 3.5346680957576337,
      "tokens_seen": 1413410816
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028878635907723173,
      "loss": 2.7749,
      "theoretical_loss": 3.5346534805088616,
      "tokens_seen": 1413476352
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028877632898696086,
      "loss": 2.5928,
      "theoretical_loss": 3.5346388661274393,
      "tokens_seen": 1413541888
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002887662988966901,
      "loss": 2.4303,
      "theoretical_loss": 3.5346242526132765,
      "tokens_seen": 1413607424
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002887562688064193,
      "loss": 2.727,
      "theoretical_loss": 3.534609639966281,
      "tokens_seen": 1413672960
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028874623871614845,
      "loss": 2.5918,
      "theoretical_loss": 3.5345950281863607,
      "tokens_seen": 1413738496
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028873620862587764,
      "loss": 2.6209,
      "theoretical_loss": 3.5345804172734248,
      "tokens_seen": 1413804032
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002887261785356068,
      "loss": 2.5483,
      "theoretical_loss": 3.5345658072273807,
      "tokens_seen": 1413869568
    },
    {
      "epoch": 4.07,
      "objective/train/docs_used": 1570961,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.574770927429199,
      "objective/train/theoretical_loss": 3.5345511980481374,
      "objective/train/tokens_used": 1434395104,
      "theoretical_loss": 3.5345511980481374,
      "tokens_seen": 1413935104
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.000288716148445336,
      "loss": 2.646,
      "theoretical_loss": 3.5345511980481374,
      "tokens_seen": 1413935104
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028870611835506523,
      "loss": 2.4373,
      "theoretical_loss": 3.5345365897356036,
      "tokens_seen": 1414000640
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028869608826479436,
      "loss": 2.6903,
      "theoretical_loss": 3.5345219822896867,
      "tokens_seen": 1414066176
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002886860581745236,
      "loss": 2.2652,
      "theoretical_loss": 3.5345073757102963,
      "tokens_seen": 1414131712
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002886760280842528,
      "loss": 2.5962,
      "theoretical_loss": 3.53449276999734,
      "tokens_seen": 1414197248
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028866599799398196,
      "loss": 2.4221,
      "theoretical_loss": 3.5344781651507264,
      "tokens_seen": 1414262784
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028865596790371114,
      "loss": 2.4368,
      "theoretical_loss": 3.5344635611703645,
      "tokens_seen": 1414328320
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002886459378134403,
      "loss": 2.5616,
      "theoretical_loss": 3.5344489580561627,
      "tokens_seen": 1414393856
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002886359077231695,
      "loss": 2.5009,
      "theoretical_loss": 3.5344343558080284,
      "tokens_seen": 1414459392
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028862587763289874,
      "loss": 2.7095,
      "theoretical_loss": 3.534419754425871,
      "tokens_seen": 1414524928
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028861584754262786,
      "loss": 2.5479,
      "theoretical_loss": 3.5344051539095993,
      "tokens_seen": 1414590464
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002886058174523571,
      "loss": 2.8498,
      "theoretical_loss": 3.5343905542591214,
      "tokens_seen": 1414656000
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002885957873620862,
      "loss": 2.5168,
      "theoretical_loss": 3.5343759554743457,
      "tokens_seen": 1414721536
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028858575727181546,
      "loss": 2.5852,
      "theoretical_loss": 3.5343613575551815,
      "tokens_seen": 1414787072
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028857572718154464,
      "loss": 2.692,
      "theoretical_loss": 3.5343467605015366,
      "tokens_seen": 1414852608
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002885656970912738,
      "loss": 2.4516,
      "theoretical_loss": 3.53433216431332,
      "tokens_seen": 1414918144
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.000288555667001003,
      "loss": 2.6683,
      "theoretical_loss": 3.53431756899044,
      "tokens_seen": 1414983680
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002885456369107322,
      "loss": 2.6535,
      "theoretical_loss": 3.534302974532805,
      "tokens_seen": 1415049216
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028853560682046137,
      "loss": 2.8792,
      "theoretical_loss": 3.5342883809403247,
      "tokens_seen": 1415114752
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002885255767301906,
      "loss": 2.798,
      "theoretical_loss": 3.534273788212907,
      "tokens_seen": 1415180288
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028851554663991973,
      "loss": 2.4001,
      "theoretical_loss": 3.5342591963504604,
      "tokens_seen": 1415245824
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028850551654964896,
      "loss": 2.5954,
      "theoretical_loss": 3.534244605352894,
      "tokens_seen": 1415311360
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028849548645937815,
      "loss": 2.586,
      "theoretical_loss": 3.5342300152201167,
      "tokens_seen": 1415376896
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002884854563691073,
      "loss": 2.7573,
      "theoretical_loss": 3.534215425952036,
      "tokens_seen": 1415442432
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002884754262788365,
      "loss": 2.7199,
      "theoretical_loss": 3.5342008375485623,
      "tokens_seen": 1415507968
    },
    {
      "epoch": 4.07,
      "objective/train/docs_used": 1571590,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.518846035003662,
      "objective/train/theoretical_loss": 3.534186250009603,
      "objective/train/tokens_used": 1436033504,
      "theoretical_loss": 3.534186250009603,
      "tokens_seen": 1415573504
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002884653961885657,
      "loss": 2.686,
      "theoretical_loss": 3.534186250009603,
      "tokens_seen": 1415573504
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028845536609829487,
      "loss": 2.5762,
      "theoretical_loss": 3.5341716633350675,
      "tokens_seen": 1415639040
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002884453360080241,
      "loss": 2.6943,
      "theoretical_loss": 3.5341570775248643,
      "tokens_seen": 1415704576
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028843530591775323,
      "loss": 2.5015,
      "theoretical_loss": 3.5341424925789022,
      "tokens_seen": 1415770112
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028842527582748247,
      "loss": 2.6678,
      "theoretical_loss": 3.5341279084970902,
      "tokens_seen": 1415835648
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002884152457372116,
      "loss": 2.7635,
      "theoretical_loss": 3.5341133252793373,
      "tokens_seen": 1415901184
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028840521564694083,
      "loss": 2.7631,
      "theoretical_loss": 3.5340987429255515,
      "tokens_seen": 1415966720
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028839518555667,
      "loss": 2.6395,
      "theoretical_loss": 3.534084161435642,
      "tokens_seen": 1416032256
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002883851554663992,
      "loss": 2.5964,
      "theoretical_loss": 3.5340695808095184,
      "tokens_seen": 1416097792
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002883751253761284,
      "loss": 2.6531,
      "theoretical_loss": 3.5340550010470886,
      "tokens_seen": 1416163328
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002883650952858576,
      "loss": 2.6817,
      "theoretical_loss": 3.5340404221482618,
      "tokens_seen": 1416228864
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028835506519558674,
      "loss": 2.4982,
      "theoretical_loss": 3.534025844112947,
      "tokens_seen": 1416294400
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028834503510531597,
      "loss": 2.6608,
      "theoretical_loss": 3.534011266941053,
      "tokens_seen": 1416359936
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002883350050150451,
      "loss": 2.5268,
      "theoretical_loss": 3.5339966906324882,
      "tokens_seen": 1416425472
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028832497492477433,
      "loss": 2.6619,
      "theoretical_loss": 3.533982115187162,
      "tokens_seen": 1416491008
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002883149448345035,
      "loss": 2.6603,
      "theoretical_loss": 3.5339675406049844,
      "tokens_seen": 1416556544
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002883049147442327,
      "loss": 2.6605,
      "theoretical_loss": 3.5339529668858622,
      "tokens_seen": 1416622080
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002882948846539619,
      "loss": 2.7202,
      "theoretical_loss": 3.533938394029706,
      "tokens_seen": 1416687616
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028828485456369106,
      "loss": 2.4932,
      "theoretical_loss": 3.5339238220364244,
      "tokens_seen": 1416753152
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002882748244734203,
      "loss": 2.659,
      "theoretical_loss": 3.533909250905926,
      "tokens_seen": 1416818688
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002882647943831495,
      "loss": 2.6021,
      "theoretical_loss": 3.53389468063812,
      "tokens_seen": 1416884224
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028825476429287866,
      "loss": 2.7558,
      "theoretical_loss": 3.533880111232916,
      "tokens_seen": 1416949760
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028824473420260784,
      "loss": 2.7942,
      "theoretical_loss": 3.533865542690222,
      "tokens_seen": 1417015296
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.000288234704112337,
      "loss": 2.7587,
      "theoretical_loss": 3.5338509750099485,
      "tokens_seen": 1417080832
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002882246740220662,
      "loss": 2.6476,
      "theoretical_loss": 3.533836408192003,
      "tokens_seen": 1417146368
    },
    {
      "epoch": 4.07,
      "objective/train/docs_used": 1572254,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7383265495300293,
      "objective/train/theoretical_loss": 3.533821842236295,
      "objective/train/tokens_used": 1437671904,
      "theoretical_loss": 3.533821842236295,
      "tokens_seen": 1417211904
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028821464393179543,
      "loss": 2.904,
      "theoretical_loss": 3.533821842236295,
      "tokens_seen": 1417211904
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028820461384152456,
      "loss": 2.7954,
      "theoretical_loss": 3.533807277142734,
      "tokens_seen": 1417277440
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002881945837512538,
      "loss": 2.466,
      "theoretical_loss": 3.533792712911229,
      "tokens_seen": 1417342976
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.000288184553660983,
      "loss": 2.723,
      "theoretical_loss": 3.533778149541689,
      "tokens_seen": 1417408512
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028817452357071216,
      "loss": 2.7506,
      "theoretical_loss": 3.5337635870340236,
      "tokens_seen": 1417474048
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028816449348044134,
      "loss": 2.5223,
      "theoretical_loss": 3.5337490253881416,
      "tokens_seen": 1417539584
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002881544633901705,
      "loss": 2.5958,
      "theoretical_loss": 3.5337344646039517,
      "tokens_seen": 1417605120
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002881444332998997,
      "loss": 2.6868,
      "theoretical_loss": 3.533719904681363,
      "tokens_seen": 1417670656
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028813440320962894,
      "loss": 2.5505,
      "theoretical_loss": 3.5337053456202865,
      "tokens_seen": 1417736192
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028812437311935806,
      "loss": 2.5364,
      "theoretical_loss": 3.533690787420629,
      "tokens_seen": 1417801728
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002881143430290873,
      "loss": 2.5286,
      "theoretical_loss": 3.5336762300823015,
      "tokens_seen": 1417867264
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002881043129388164,
      "loss": 2.84,
      "theoretical_loss": 3.533661673605212,
      "tokens_seen": 1417932800
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028809428284854566,
      "loss": 2.6731,
      "theoretical_loss": 3.5336471179892706,
      "tokens_seen": 1417998336
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028808425275827484,
      "loss": 2.6108,
      "theoretical_loss": 3.533632563234386,
      "tokens_seen": 1418063872
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.000288074222668004,
      "loss": 2.8358,
      "theoretical_loss": 3.5336180093404677,
      "tokens_seen": 1418129408
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002880641925777332,
      "loss": 2.8047,
      "theoretical_loss": 3.533603456307425,
      "tokens_seen": 1418194944
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002880541624874624,
      "loss": 2.6638,
      "theoretical_loss": 3.5335889041351676,
      "tokens_seen": 1418260480
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028804413239719157,
      "loss": 2.7179,
      "theoretical_loss": 3.533574352823604,
      "tokens_seen": 1418326016
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002880341023069208,
      "loss": 2.6641,
      "theoretical_loss": 3.533559802372644,
      "tokens_seen": 1418391552
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028802407221664993,
      "loss": 2.6614,
      "theoretical_loss": 3.533545252782197,
      "tokens_seen": 1418457088
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028801404212637916,
      "loss": 2.8433,
      "theoretical_loss": 3.5335307040521724,
      "tokens_seen": 1418522624
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028800401203610835,
      "loss": 2.282,
      "theoretical_loss": 3.533516156182479,
      "tokens_seen": 1418588160
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002879939819458375,
      "loss": 2.5477,
      "theoretical_loss": 3.533501609173027,
      "tokens_seen": 1418653696
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002879839518555667,
      "loss": 2.6269,
      "theoretical_loss": 3.533487063023725,
      "tokens_seen": 1418719232
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002879739217652959,
      "loss": 2.5457,
      "theoretical_loss": 3.5334725177344835,
      "tokens_seen": 1418784768
    },
    {
      "epoch": 4.07,
      "objective/train/docs_used": 1573335,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.203737497329712,
      "objective/train/theoretical_loss": 3.53345797330521,
      "objective/train/tokens_used": 1439310304,
      "theoretical_loss": 3.53345797330521,
      "tokens_seen": 1418850304
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028796389167502507,
      "loss": 2.6313,
      "theoretical_loss": 3.53345797330521,
      "tokens_seen": 1418850304
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002879538615847543,
      "loss": 2.4921,
      "theoretical_loss": 3.533443429735816,
      "tokens_seen": 1418915840
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028794383149448343,
      "loss": 2.6774,
      "theoretical_loss": 3.53342888702621,
      "tokens_seen": 1418981376
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028793380140421267,
      "loss": 2.7072,
      "theoretical_loss": 3.5334143451763014,
      "tokens_seen": 1419046912
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002879237713139418,
      "loss": 2.6758,
      "theoretical_loss": 3.533399804186,
      "tokens_seen": 1419112448
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028791374122367103,
      "loss": 2.6,
      "theoretical_loss": 3.5333852640552155,
      "tokens_seen": 1419177984
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002879037111334002,
      "loss": 2.538,
      "theoretical_loss": 3.5333707247838566,
      "tokens_seen": 1419243520
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002878936810431294,
      "loss": 2.5949,
      "theoretical_loss": 3.533356186371833,
      "tokens_seen": 1419309056
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002878836509528586,
      "loss": 2.7255,
      "theoretical_loss": 3.5333416488190554,
      "tokens_seen": 1419374592
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002878736208625878,
      "loss": 2.6536,
      "theoretical_loss": 3.5333271121254315,
      "tokens_seen": 1419440128
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028786359077231694,
      "loss": 2.7465,
      "theoretical_loss": 3.5333125762908724,
      "tokens_seen": 1419505664
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028785356068204617,
      "loss": 2.581,
      "theoretical_loss": 3.5332980413152866,
      "tokens_seen": 1419571200
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002878435305917753,
      "loss": 2.615,
      "theoretical_loss": 3.5332835071985844,
      "tokens_seen": 1419636736
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028783350050150453,
      "loss": 2.7341,
      "theoretical_loss": 3.5332689739406753,
      "tokens_seen": 1419702272
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002878234704112337,
      "loss": 2.5138,
      "theoretical_loss": 3.5332544415414686,
      "tokens_seen": 1419767808
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002878134403209629,
      "loss": 2.6365,
      "theoretical_loss": 3.5332399100008747,
      "tokens_seen": 1419833344
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002878034102306921,
      "loss": 2.8449,
      "theoretical_loss": 3.533225379318802,
      "tokens_seen": 1419898880
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028779338014042126,
      "loss": 2.6908,
      "theoretical_loss": 3.533210849495161,
      "tokens_seen": 1419964416
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028778335005015044,
      "loss": 2.3565,
      "theoretical_loss": 3.5331963205298615,
      "tokens_seen": 1420029952
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002877733199598797,
      "loss": 2.7702,
      "theoretical_loss": 3.5331817924228126,
      "tokens_seen": 1420095488
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002877632898696088,
      "loss": 2.5993,
      "theoretical_loss": 3.5331672651739243,
      "tokens_seen": 1420161024
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028775325977933804,
      "loss": 2.6276,
      "theoretical_loss": 3.533152738783106,
      "tokens_seen": 1420226560
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028774322968906716,
      "loss": 2.9462,
      "theoretical_loss": 3.5331382132502682,
      "tokens_seen": 1420292096
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002877331995987964,
      "loss": 2.6313,
      "theoretical_loss": 3.5331236885753197,
      "tokens_seen": 1420357632
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002877231695085256,
      "loss": 2.7853,
      "theoretical_loss": 3.533109164758171,
      "tokens_seen": 1420423168
    },
    {
      "epoch": 4.07,
      "objective/train/docs_used": 1574634,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9888744354248047,
      "objective/train/theoretical_loss": 3.5330946417987312,
      "objective/train/tokens_used": 1440948704,
      "theoretical_loss": 3.5330946417987312,
      "tokens_seen": 1420488704
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028771313941825476,
      "loss": 2.6085,
      "theoretical_loss": 3.5330946417987312,
      "tokens_seen": 1420488704
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028770310932798394,
      "loss": 2.5191,
      "theoretical_loss": 3.533080119696911,
      "tokens_seen": 1420554240
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002876930792377132,
      "loss": 2.7856,
      "theoretical_loss": 3.5330655984526196,
      "tokens_seen": 1420619776
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002876830491474423,
      "loss": 2.586,
      "theoretical_loss": 3.5330510780657662,
      "tokens_seen": 1420685312
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028767301905717154,
      "loss": 2.7563,
      "theoretical_loss": 3.5330365585362618,
      "tokens_seen": 1420750848
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028766298896690067,
      "loss": 2.8421,
      "theoretical_loss": 3.5330220398640155,
      "tokens_seen": 1420816384
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002876529588766299,
      "loss": 2.6803,
      "theoretical_loss": 3.5330075220489374,
      "tokens_seen": 1420881920
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002876429287863591,
      "loss": 2.5614,
      "theoretical_loss": 3.5329930050909373,
      "tokens_seen": 1420947456
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028763289869608826,
      "loss": 2.654,
      "theoretical_loss": 3.532978488989925,
      "tokens_seen": 1421012992
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028762286860581745,
      "loss": 2.5684,
      "theoretical_loss": 3.5329639737458107,
      "tokens_seen": 1421078528
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002876128385155466,
      "loss": 2.6544,
      "theoretical_loss": 3.5329494593585036,
      "tokens_seen": 1421144064
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002876028084252758,
      "loss": 2.7579,
      "theoretical_loss": 3.5329349458279147,
      "tokens_seen": 1421209600
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028759277833500504,
      "loss": 2.7276,
      "theoretical_loss": 3.5329204331539534,
      "tokens_seen": 1421275136
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028758274824473417,
      "loss": 2.6584,
      "theoretical_loss": 3.5329059213365293,
      "tokens_seen": 1421340672
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002875727181544634,
      "loss": 2.7329,
      "theoretical_loss": 3.5328914103755524,
      "tokens_seen": 1421406208
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028756268806419253,
      "loss": 2.8148,
      "theoretical_loss": 3.5328769002709333,
      "tokens_seen": 1421471744
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028755265797392177,
      "loss": 2.7957,
      "theoretical_loss": 3.532862391022581,
      "tokens_seen": 1421537280
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028754262788365095,
      "loss": 2.9081,
      "theoretical_loss": 3.5328478826304064,
      "tokens_seen": 1421602816
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028753259779338013,
      "loss": 2.7681,
      "theoretical_loss": 3.5328333750943193,
      "tokens_seen": 1421668352
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028752256770310936,
      "loss": 2.7856,
      "theoretical_loss": 3.532818868414229,
      "tokens_seen": 1421733888
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028751253761283855,
      "loss": 2.744,
      "theoretical_loss": 3.5328043625900474,
      "tokens_seen": 1421799424
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002875025075225677,
      "loss": 2.7367,
      "theoretical_loss": 3.5327898576216823,
      "tokens_seen": 1421864960
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002874924774322969,
      "loss": 2.4577,
      "theoretical_loss": 3.532775353509045,
      "tokens_seen": 1421930496
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002874824473420261,
      "loss": 2.7095,
      "theoretical_loss": 3.5327608502520453,
      "tokens_seen": 1421996032
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028747241725175527,
      "loss": 2.8389,
      "theoretical_loss": 3.5327463478505936,
      "tokens_seen": 1422061568
    },
    {
      "epoch": 4.07,
      "objective/train/docs_used": 1575381,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.637848377227783,
      "objective/train/theoretical_loss": 3.5327318463046,
      "objective/train/tokens_used": 1442587104,
      "theoretical_loss": 3.5327318463046,
      "tokens_seen": 1422127104
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002874623871614845,
      "loss": 2.8563,
      "theoretical_loss": 3.5327318463046,
      "tokens_seen": 1422127104
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028745235707121363,
      "loss": 2.8276,
      "theoretical_loss": 3.5327173456139733,
      "tokens_seen": 1422192640
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028744232698094287,
      "loss": 2.7305,
      "theoretical_loss": 3.5327028457786254,
      "tokens_seen": 1422258176
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.000287432296890672,
      "loss": 2.5748,
      "theoretical_loss": 3.5326883467984658,
      "tokens_seen": 1422323712
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028742226680040123,
      "loss": 2.6212,
      "theoretical_loss": 3.5326738486734044,
      "tokens_seen": 1422389248
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002874122367101304,
      "loss": 2.5636,
      "theoretical_loss": 3.5326593514033515,
      "tokens_seen": 1422454784
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002874022066198596,
      "loss": 2.6832,
      "theoretical_loss": 3.532644854988218,
      "tokens_seen": 1422520320
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002873921765295888,
      "loss": 2.6347,
      "theoretical_loss": 3.532630359427913,
      "tokens_seen": 1422585856
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.000287382146439318,
      "loss": 2.5502,
      "theoretical_loss": 3.532615864722347,
      "tokens_seen": 1422651392
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028737211634904714,
      "loss": 2.4627,
      "theoretical_loss": 3.532601370871431,
      "tokens_seen": 1422716928
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028736208625877637,
      "loss": 2.6478,
      "theoretical_loss": 3.532586877875074,
      "tokens_seen": 1422782464
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002873520561685055,
      "loss": 2.6382,
      "theoretical_loss": 3.532572385733187,
      "tokens_seen": 1422848000
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028734202607823473,
      "loss": 2.736,
      "theoretical_loss": 3.532557894445681,
      "tokens_seen": 1422913536
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002873319959879639,
      "loss": 2.6704,
      "theoretical_loss": 3.532543404012465,
      "tokens_seen": 1422979072
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002873219658976931,
      "loss": 2.7003,
      "theoretical_loss": 3.532528914433449,
      "tokens_seen": 1423044608
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002873119358074223,
      "loss": 2.6309,
      "theoretical_loss": 3.532514425708545,
      "tokens_seen": 1423110144
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028730190571715146,
      "loss": 2.5853,
      "theoretical_loss": 3.5324999378376623,
      "tokens_seen": 1423175680
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028729187562688064,
      "loss": 2.7507,
      "theoretical_loss": 3.532485450820711,
      "tokens_seen": 1423241216
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002872818455366099,
      "loss": 2.6948,
      "theoretical_loss": 3.532470964657602,
      "tokens_seen": 1423306752
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.000287271815446339,
      "loss": 2.5424,
      "theoretical_loss": 3.5324564793482454,
      "tokens_seen": 1423372288
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028726178535606824,
      "loss": 2.5131,
      "theoretical_loss": 3.5324419948925514,
      "tokens_seen": 1423437824
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028725175526579736,
      "loss": 2.636,
      "theoretical_loss": 3.5324275112904306,
      "tokens_seen": 1423503360
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002872417251755266,
      "loss": 2.7088,
      "theoretical_loss": 3.532413028541794,
      "tokens_seen": 1423568896
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002872316950852558,
      "loss": 2.6311,
      "theoretical_loss": 3.532398546646551,
      "tokens_seen": 1423634432
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028722166499498496,
      "loss": 2.9132,
      "theoretical_loss": 3.532384065604613,
      "tokens_seen": 1423699968
    },
    {
      "epoch": 4.07,
      "objective/train/docs_used": 1576738,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.3536036014556885,
      "objective/train/theoretical_loss": 3.532369585415889,
      "objective/train/tokens_used": 1444225504,
      "theoretical_loss": 3.532369585415889,
      "tokens_seen": 1423765504
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028721163490471414,
      "loss": 2.5959,
      "theoretical_loss": 3.532369585415889,
      "tokens_seen": 1423765504
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002872016048144434,
      "loss": 2.4618,
      "theoretical_loss": 3.532355106080291,
      "tokens_seen": 1423831040
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002871915747241725,
      "loss": 2.8228,
      "theoretical_loss": 3.532340627597729,
      "tokens_seen": 1423896576
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028718154463390174,
      "loss": 2.927,
      "theoretical_loss": 3.532326149968113,
      "tokens_seen": 1423962112
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028717151454363087,
      "loss": 2.5715,
      "theoretical_loss": 3.5323116731913538,
      "tokens_seen": 1424027648
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002871614844533601,
      "loss": 2.7448,
      "theoretical_loss": 3.5322971972673622,
      "tokens_seen": 1424093184
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002871514543630893,
      "loss": 2.6791,
      "theoretical_loss": 3.532282722196048,
      "tokens_seen": 1424158720
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028714142427281846,
      "loss": 2.5363,
      "theoretical_loss": 3.5322682479773224,
      "tokens_seen": 1424224256
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028713139418254765,
      "loss": 2.5642,
      "theoretical_loss": 3.532253774611096,
      "tokens_seen": 1424289792
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002871213640922768,
      "loss": 2.6663,
      "theoretical_loss": 3.5322393020972793,
      "tokens_seen": 1424355328
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.000287111334002006,
      "loss": 2.8504,
      "theoretical_loss": 3.5322248304357826,
      "tokens_seen": 1424420864
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028710130391173524,
      "loss": 2.8866,
      "theoretical_loss": 3.532210359626516,
      "tokens_seen": 1424486400
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028709127382146437,
      "loss": 2.8101,
      "theoretical_loss": 3.5321958896693917,
      "tokens_seen": 1424551936
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002870812437311936,
      "loss": 2.6613,
      "theoretical_loss": 3.532181420564319,
      "tokens_seen": 1424617472
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028707121364092273,
      "loss": 2.6923,
      "theoretical_loss": 3.5321669523112087,
      "tokens_seen": 1424683008
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028706118355065197,
      "loss": 2.7143,
      "theoretical_loss": 3.5321524849099717,
      "tokens_seen": 1424748544
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028705115346038115,
      "loss": 2.7368,
      "theoretical_loss": 3.532138018360518,
      "tokens_seen": 1424814080
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028704112337011033,
      "loss": 2.709,
      "theoretical_loss": 3.53212355266276,
      "tokens_seen": 1424879616
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002870310932798395,
      "loss": 2.624,
      "theoretical_loss": 3.5321090878166066,
      "tokens_seen": 1424945152
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028702106318956875,
      "loss": 2.7142,
      "theoretical_loss": 3.532094623821969,
      "tokens_seen": 1425010688
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002870110330992979,
      "loss": 2.9406,
      "theoretical_loss": 3.5320801606787584,
      "tokens_seen": 1425076224
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002870010030090271,
      "loss": 2.578,
      "theoretical_loss": 3.5320656983868854,
      "tokens_seen": 1425141760
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028699097291875624,
      "loss": 2.6162,
      "theoretical_loss": 3.53205123694626,
      "tokens_seen": 1425207296
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028698094282848547,
      "loss": 2.6059,
      "theoretical_loss": 3.5320367763567937,
      "tokens_seen": 1425272832
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028697091273821465,
      "loss": 2.5249,
      "theoretical_loss": 3.532022316618397,
      "tokens_seen": 1425338368
    },
    {
      "epoch": 4.07,
      "objective/train/docs_used": 1577330,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.346597671508789,
      "objective/train/theoretical_loss": 3.5320078577309806,
      "objective/train/tokens_used": 1445863904,
      "theoretical_loss": 3.5320078577309806,
      "tokens_seen": 1425403904
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028696088264794383,
      "loss": 2.605,
      "theoretical_loss": 3.5320078577309806,
      "tokens_seen": 1425403904
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.000286950852557673,
      "loss": 2.644,
      "theoretical_loss": 3.531993399694456,
      "tokens_seen": 1425469440
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002869408224674022,
      "loss": 2.7256,
      "theoretical_loss": 3.531978942508733,
      "tokens_seen": 1425534976
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002869307923771314,
      "loss": 2.6636,
      "theoretical_loss": 3.531964486173723,
      "tokens_seen": 1425600512
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002869207622868606,
      "loss": 2.9186,
      "theoretical_loss": 3.5319500306893366,
      "tokens_seen": 1425666048
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028691073219658974,
      "loss": 2.8108,
      "theoretical_loss": 3.531935576055485,
      "tokens_seen": 1425731584
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.000286900702106319,
      "loss": 2.8785,
      "theoretical_loss": 3.5319211222720788,
      "tokens_seen": 1425797120
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002868906720160481,
      "loss": 2.8474,
      "theoretical_loss": 3.531906669339029,
      "tokens_seen": 1425862656
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028688064192577734,
      "loss": 2.6304,
      "theoretical_loss": 3.531892217256246,
      "tokens_seen": 1425928192
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002868706118355065,
      "loss": 2.753,
      "theoretical_loss": 3.531877766023641,
      "tokens_seen": 1425993728
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002868605817452357,
      "loss": 2.8028,
      "theoretical_loss": 3.531863315641125,
      "tokens_seen": 1426059264
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002868505516549649,
      "loss": 2.5905,
      "theoretical_loss": 3.53184886610861,
      "tokens_seen": 1426124800
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002868405215646941,
      "loss": 3.0052,
      "theoretical_loss": 3.531834417426005,
      "tokens_seen": 1426190336
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028683049147442324,
      "loss": 2.3945,
      "theoretical_loss": 3.5318199695932218,
      "tokens_seen": 1426255872
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002868204613841525,
      "loss": 2.5216,
      "theoretical_loss": 3.5318055226101714,
      "tokens_seen": 1426321408
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002868104312938816,
      "loss": 2.8529,
      "theoretical_loss": 3.5317910764767655,
      "tokens_seen": 1426386944
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028680040120361084,
      "loss": 2.6891,
      "theoretical_loss": 3.5317766311929133,
      "tokens_seen": 1426452480
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028679037111334,
      "loss": 2.5532,
      "theoretical_loss": 3.5317621867585274,
      "tokens_seen": 1426518016
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002867803410230692,
      "loss": 2.6447,
      "theoretical_loss": 3.5317477431735185,
      "tokens_seen": 1426583552
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028677031093279844,
      "loss": 2.5323,
      "theoretical_loss": 3.5317333004377973,
      "tokens_seen": 1426649088
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028676028084252756,
      "loss": 2.6454,
      "theoretical_loss": 3.531718858551275,
      "tokens_seen": 1426714624
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002867502507522568,
      "loss": 2.8039,
      "theoretical_loss": 3.531704417513863,
      "tokens_seen": 1426780160
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.000286740220661986,
      "loss": 2.3451,
      "theoretical_loss": 3.531689977325472,
      "tokens_seen": 1426845696
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028673019057171516,
      "loss": 2.9817,
      "theoretical_loss": 3.531675537986013,
      "tokens_seen": 1426911232
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028672016048144434,
      "loss": 2.749,
      "theoretical_loss": 3.5316610994953974,
      "tokens_seen": 1426976768
    },
    {
      "epoch": 4.07,
      "objective/train/docs_used": 1578940,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.657034158706665,
      "objective/train/theoretical_loss": 3.531646661853536,
      "objective/train/tokens_used": 1447502304,
      "theoretical_loss": 3.531646661853536,
      "tokens_seen": 1427042304
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002867101303911736,
      "loss": 2.8357,
      "theoretical_loss": 3.531646661853536,
      "tokens_seen": 1427042304
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002867001003009027,
      "loss": 2.7306,
      "theoretical_loss": 3.5316322250603402,
      "tokens_seen": 1427107840
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028669007021063194,
      "loss": 2.5669,
      "theoretical_loss": 3.531617789115721,
      "tokens_seen": 1427173376
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028668004012036107,
      "loss": 2.8655,
      "theoretical_loss": 3.5316033540195892,
      "tokens_seen": 1427238912
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002866700100300903,
      "loss": 2.6255,
      "theoretical_loss": 3.531588919771857,
      "tokens_seen": 1427304448
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002866599799398195,
      "loss": 2.7666,
      "theoretical_loss": 3.5315744863724348,
      "tokens_seen": 1427369984
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028664994984954866,
      "loss": 2.7119,
      "theoretical_loss": 3.531560053821234,
      "tokens_seen": 1427435520
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028663991975927785,
      "loss": 2.5825,
      "theoretical_loss": 3.531545622118166,
      "tokens_seen": 1427501056
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.000286629889669007,
      "loss": 2.3998,
      "theoretical_loss": 3.5315311912631415,
      "tokens_seen": 1427566592
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002866198595787362,
      "loss": 2.5511,
      "theoretical_loss": 3.531516761256072,
      "tokens_seen": 1427632128
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028660982948846544,
      "loss": 2.8859,
      "theoretical_loss": 3.531502332096869,
      "tokens_seen": 1427697664
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028659979939819457,
      "loss": 2.5624,
      "theoretical_loss": 3.531487903785443,
      "tokens_seen": 1427763200
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002865897693079238,
      "loss": 2.8676,
      "theoretical_loss": 3.5314734763217066,
      "tokens_seen": 1427828736
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028657973921765293,
      "loss": 2.8014,
      "theoretical_loss": 3.5314590497055702,
      "tokens_seen": 1427894272
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028656970912738217,
      "loss": 2.6093,
      "theoretical_loss": 3.531444623936945,
      "tokens_seen": 1427959808
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028655967903711135,
      "loss": 2.5359,
      "theoretical_loss": 3.5314301990157424,
      "tokens_seen": 1428025344
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028654964894684053,
      "loss": 2.6843,
      "theoretical_loss": 3.531415774941874,
      "tokens_seen": 1428090880
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002865396188565697,
      "loss": 2.7487,
      "theoretical_loss": 3.5314013517152514,
      "tokens_seen": 1428156416
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028652958876629895,
      "loss": 2.4328,
      "theoretical_loss": 3.531386929335785,
      "tokens_seen": 1428221952
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002865195586760281,
      "loss": 2.7507,
      "theoretical_loss": 3.531372507803387,
      "tokens_seen": 1428287488
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002865095285857573,
      "loss": 2.7066,
      "theoretical_loss": 3.531358087117969,
      "tokens_seen": 1428353024
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028649949849548644,
      "loss": 2.5695,
      "theoretical_loss": 3.531343667279441,
      "tokens_seen": 1428418560
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028648946840521567,
      "loss": 2.377,
      "theoretical_loss": 3.531329248287716,
      "tokens_seen": 1428484096
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028647943831494485,
      "loss": 2.8074,
      "theoretical_loss": 3.531314830142705,
      "tokens_seen": 1428549632
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028646940822467403,
      "loss": 2.5772,
      "theoretical_loss": 3.5313004128443195,
      "tokens_seen": 1428615168
    },
    {
      "epoch": 4.07,
      "objective/train/docs_used": 1579734,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7796366214752197,
      "objective/train/theoretical_loss": 3.53128599639247,
      "objective/train/tokens_used": 1449140704,
      "theoretical_loss": 3.53128599639247,
      "tokens_seen": 1428680704
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002864593781344032,
      "loss": 2.912,
      "theoretical_loss": 3.53128599639247,
      "tokens_seen": 1428680704
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002864493480441324,
      "loss": 2.6375,
      "theoretical_loss": 3.5312715807870685,
      "tokens_seen": 1428746240
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002864393179538616,
      "loss": 2.8322,
      "theoretical_loss": 3.531257166028027,
      "tokens_seen": 1428811776
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002864292878635908,
      "loss": 2.5146,
      "theoretical_loss": 3.5312427521152565,
      "tokens_seen": 1428877312
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028641925777331994,
      "loss": 2.7135,
      "theoretical_loss": 3.5312283390486687,
      "tokens_seen": 1428942848
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002864092276830492,
      "loss": 2.7155,
      "theoretical_loss": 3.5312139268281753,
      "tokens_seen": 1429008384
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002863991975927783,
      "loss": 2.5024,
      "theoretical_loss": 3.531199515453687,
      "tokens_seen": 1429073920
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028638916750250754,
      "loss": 2.736,
      "theoretical_loss": 3.5311851049251164,
      "tokens_seen": 1429139456
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002863791374122367,
      "loss": 2.7296,
      "theoretical_loss": 3.5311706952423743,
      "tokens_seen": 1429204992
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002863691073219659,
      "loss": 2.68,
      "theoretical_loss": 3.5311562864053725,
      "tokens_seen": 1429270528
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002863590772316951,
      "loss": 2.5365,
      "theoretical_loss": 3.5311418784140227,
      "tokens_seen": 1429336064
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002863490471414243,
      "loss": 2.7918,
      "theoretical_loss": 3.5311274712682366,
      "tokens_seen": 1429401600
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028633901705115344,
      "loss": 2.5645,
      "theoretical_loss": 3.531113064967925,
      "tokens_seen": 1429467136
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002863289869608827,
      "loss": 2.7413,
      "theoretical_loss": 3.5310986595130007,
      "tokens_seen": 1429532672
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002863189568706118,
      "loss": 2.554,
      "theoretical_loss": 3.5310842549033747,
      "tokens_seen": 1429598208
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028630892678034104,
      "loss": 2.643,
      "theoretical_loss": 3.5310698511389593,
      "tokens_seen": 1429663744
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002862988966900702,
      "loss": 2.5413,
      "theoretical_loss": 3.531055448219665,
      "tokens_seen": 1429729280
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002862888665997994,
      "loss": 2.6201,
      "theoretical_loss": 3.5310410461454045,
      "tokens_seen": 1429794816
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002862788365095286,
      "loss": 2.6887,
      "theoretical_loss": 3.5310266449160888,
      "tokens_seen": 1429860352
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028626880641925776,
      "loss": 2.6714,
      "theoretical_loss": 3.53101224453163,
      "tokens_seen": 1429925888
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028625877632898694,
      "loss": 2.5562,
      "theoretical_loss": 3.5309978449919397,
      "tokens_seen": 1429991424
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002862487462387162,
      "loss": 2.7693,
      "theoretical_loss": 3.5309834462969296,
      "tokens_seen": 1430056960
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002862387161484453,
      "loss": 2.8589,
      "theoretical_loss": 3.5309690484465115,
      "tokens_seen": 1430122496
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028622868605817454,
      "loss": 2.706,
      "theoretical_loss": 3.530954651440597,
      "tokens_seen": 1430188032
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028621865596790367,
      "loss": 2.8354,
      "theoretical_loss": 3.5309402552790985,
      "tokens_seen": 1430253568
    },
    {
      "epoch": 4.07,
      "objective/train/docs_used": 1580976,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8712832927703857,
      "objective/train/theoretical_loss": 3.530925859961927,
      "objective/train/tokens_used": 1450779104,
      "theoretical_loss": 3.530925859961927,
      "tokens_seen": 1430319104
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002862086258776329,
      "loss": 2.7676,
      "theoretical_loss": 3.530925859961927,
      "tokens_seen": 1430319104
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002861985957873621,
      "loss": 2.5262,
      "theoretical_loss": 3.5309114654889946,
      "tokens_seen": 1430384640
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028618856569709127,
      "loss": 2.659,
      "theoretical_loss": 3.530897071860213,
      "tokens_seen": 1430450176
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028617853560682045,
      "loss": 2.7417,
      "theoretical_loss": 3.530882679075494,
      "tokens_seen": 1430515712
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002861685055165497,
      "loss": 2.7481,
      "theoretical_loss": 3.53086828713475,
      "tokens_seen": 1430581248
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002861584754262788,
      "loss": 2.532,
      "theoretical_loss": 3.5308538960378923,
      "tokens_seen": 1430646784
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028614844533600805,
      "loss": 2.6953,
      "theoretical_loss": 3.530839505784833,
      "tokens_seen": 1430712320
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028613841524573717,
      "loss": 2.6791,
      "theoretical_loss": 3.5308251163754836,
      "tokens_seen": 1430777856
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002861283851554664,
      "loss": 2.5014,
      "theoretical_loss": 3.5308107278097562,
      "tokens_seen": 1430843392
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002861183550651956,
      "loss": 2.7834,
      "theoretical_loss": 3.530796340087563,
      "tokens_seen": 1430908928
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028610832497492477,
      "loss": 2.782,
      "theoretical_loss": 3.530781953208816,
      "tokens_seen": 1430974464
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028609829488465395,
      "loss": 2.7377,
      "theoretical_loss": 3.5307675671734264,
      "tokens_seen": 1431040000
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028608826479438313,
      "loss": 2.4967,
      "theoretical_loss": 3.5307531819813063,
      "tokens_seen": 1431105536
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002860782347041123,
      "loss": 2.6931,
      "theoretical_loss": 3.5307387976323685,
      "tokens_seen": 1431171072
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028606820461384155,
      "loss": 2.6167,
      "theoretical_loss": 3.5307244141265235,
      "tokens_seen": 1431236608
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002860581745235707,
      "loss": 2.6194,
      "theoretical_loss": 3.530710031463685,
      "tokens_seen": 1431302144
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002860481444332999,
      "loss": 2.6382,
      "theoretical_loss": 3.5306956496437643,
      "tokens_seen": 1431367680
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002860381143430291,
      "loss": 2.6777,
      "theoretical_loss": 3.5306812686666724,
      "tokens_seen": 1431433216
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002860280842527583,
      "loss": 2.7881,
      "theoretical_loss": 3.5306668885323225,
      "tokens_seen": 1431498752
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002860180541624875,
      "loss": 2.5622,
      "theoretical_loss": 3.5306525092406265,
      "tokens_seen": 1431564288
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028600802407221664,
      "loss": 2.7507,
      "theoretical_loss": 3.5306381307914965,
      "tokens_seen": 1431629824
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028599799398194587,
      "loss": 2.7967,
      "theoretical_loss": 3.530623753184844,
      "tokens_seen": 1431695360
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028598796389167505,
      "loss": 2.6421,
      "theoretical_loss": 3.530609376420582,
      "tokens_seen": 1431760896
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028597793380140423,
      "loss": 2.7102,
      "theoretical_loss": 3.530595000498621,
      "tokens_seen": 1431826432
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002859679037111334,
      "loss": 2.5797,
      "theoretical_loss": 3.5305806254188745,
      "tokens_seen": 1431891968
    },
    {
      "epoch": 4.07,
      "objective/train/docs_used": 1581614,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.419712543487549,
      "objective/train/theoretical_loss": 3.5305662511812548,
      "objective/train/tokens_used": 1452417504,
      "theoretical_loss": 3.5305662511812548,
      "tokens_seen": 1431957504
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002859578736208626,
      "loss": 2.4994,
      "theoretical_loss": 3.5305662511812548,
      "tokens_seen": 1431957504
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002859478435305918,
      "loss": 2.7283,
      "theoretical_loss": 3.5305518777856726,
      "tokens_seen": 1432023040
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.000285937813440321,
      "loss": 2.7827,
      "theoretical_loss": 3.530537505232041,
      "tokens_seen": 1432088576
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028592778335005014,
      "loss": 2.7673,
      "theoretical_loss": 3.5305231335202727,
      "tokens_seen": 1432154112
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002859177532597794,
      "loss": 2.849,
      "theoretical_loss": 3.5305087626502787,
      "tokens_seen": 1432219648
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002859077231695085,
      "loss": 2.8395,
      "theoretical_loss": 3.530494392621972,
      "tokens_seen": 1432285184
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028589769307923774,
      "loss": 2.6592,
      "theoretical_loss": 3.5304800234352642,
      "tokens_seen": 1432350720
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002858876629889669,
      "loss": 2.6934,
      "theoretical_loss": 3.530465655090068,
      "tokens_seen": 1432416256
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002858776328986961,
      "loss": 2.6282,
      "theoretical_loss": 3.530451287586295,
      "tokens_seen": 1432481792
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002858676028084253,
      "loss": 2.6504,
      "theoretical_loss": 3.530436920923858,
      "tokens_seen": 1432547328
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002858575727181545,
      "loss": 2.4934,
      "theoretical_loss": 3.5304225551026693,
      "tokens_seen": 1432612864
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028584754262788364,
      "loss": 2.621,
      "theoretical_loss": 3.5304081901226407,
      "tokens_seen": 1432678400
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002858375125376129,
      "loss": 2.5828,
      "theoretical_loss": 3.530393825983685,
      "tokens_seen": 1432743936
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.000285827482447342,
      "loss": 2.852,
      "theoretical_loss": 3.5303794626857146,
      "tokens_seen": 1432809472
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028581745235707124,
      "loss": 2.5021,
      "theoretical_loss": 3.5303651002286407,
      "tokens_seen": 1432875008
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002858074222668004,
      "loss": 2.7758,
      "theoretical_loss": 3.530350738612377,
      "tokens_seen": 1432940544
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002857973921765296,
      "loss": 2.4004,
      "theoretical_loss": 3.5303363778368344,
      "tokens_seen": 1433006080
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002857873620862588,
      "loss": 2.4702,
      "theoretical_loss": 3.530322017901926,
      "tokens_seen": 1433071616
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028577733199598796,
      "loss": 2.6284,
      "theoretical_loss": 3.5303076588075646,
      "tokens_seen": 1433137152
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028576730190571714,
      "loss": 2.9295,
      "theoretical_loss": 3.5302933005536623,
      "tokens_seen": 1433202688
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002857572718154464,
      "loss": 2.5359,
      "theoretical_loss": 3.530278943140131,
      "tokens_seen": 1433268224
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002857472417251755,
      "loss": 2.5897,
      "theoretical_loss": 3.530264586566883,
      "tokens_seen": 1433333760
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028573721163490474,
      "loss": 2.6742,
      "theoretical_loss": 3.5302502308338317,
      "tokens_seen": 1433399296
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028572718154463387,
      "loss": 2.7224,
      "theoretical_loss": 3.5302358759408885,
      "tokens_seen": 1433464832
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002857171514543631,
      "loss": 2.5141,
      "theoretical_loss": 3.5302215218879667,
      "tokens_seen": 1433530368
    },
    {
      "epoch": 4.07,
      "objective/train/docs_used": 1582179,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.372399091720581,
      "objective/train/theoretical_loss": 3.530207168674978,
      "objective/train/tokens_used": 1454055904,
      "theoretical_loss": 3.530207168674978,
      "tokens_seen": 1433595904
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002857071213640923,
      "loss": 2.5721,
      "theoretical_loss": 3.530207168674978,
      "tokens_seen": 1433595904
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028569709127382147,
      "loss": 2.4279,
      "theoretical_loss": 3.5301928163018346,
      "tokens_seen": 1433661440
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028568706118355065,
      "loss": 2.593,
      "theoretical_loss": 3.5301784647684498,
      "tokens_seen": 1433726976
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002856770310932799,
      "loss": 2.7227,
      "theoretical_loss": 3.5301641140747364,
      "tokens_seen": 1433792512
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.000285667001003009,
      "loss": 2.9401,
      "theoretical_loss": 3.5301497642206057,
      "tokens_seen": 1433858048
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028565697091273825,
      "loss": 2.4208,
      "theoretical_loss": 3.530135415205971,
      "tokens_seen": 1433923584
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028564694082246737,
      "loss": 2.3774,
      "theoretical_loss": 3.530121067030745,
      "tokens_seen": 1433989120
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002856369107321966,
      "loss": 2.7481,
      "theoretical_loss": 3.530106719694839,
      "tokens_seen": 1434054656
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002856268806419258,
      "loss": 2.5669,
      "theoretical_loss": 3.5300923731981664,
      "tokens_seen": 1434120192
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028561685055165497,
      "loss": 2.6281,
      "theoretical_loss": 3.5300780275406405,
      "tokens_seen": 1434185728
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028560682046138415,
      "loss": 2.5687,
      "theoretical_loss": 3.5300636827221727,
      "tokens_seen": 1434251264
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028559679037111333,
      "loss": 2.5335,
      "theoretical_loss": 3.530049338742676,
      "tokens_seen": 1434316800
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002855867602808425,
      "loss": 2.568,
      "theoretical_loss": 3.5300349956020636,
      "tokens_seen": 1434382336
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.00028557673019057175,
      "loss": 2.7643,
      "theoretical_loss": 3.530020653300247,
      "tokens_seen": 1434447872
    },
    {
      "epoch": 4.07,
      "learning_rate": 0.0002855667001003009,
      "loss": 2.5505,
      "theoretical_loss": 3.5300063118371394,
      "tokens_seen": 1434513408
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002855566700100301,
      "loss": 2.7626,
      "theoretical_loss": 3.5299919712126533,
      "tokens_seen": 1434578944
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002855466399197593,
      "loss": 2.6743,
      "theoretical_loss": 3.5299776314267017,
      "tokens_seen": 1434644480
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002855366098294885,
      "loss": 2.7768,
      "theoretical_loss": 3.5299632924791973,
      "tokens_seen": 1434710016
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028552657973921765,
      "loss": 2.8286,
      "theoretical_loss": 3.5299489543700524,
      "tokens_seen": 1434775552
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028551654964894684,
      "loss": 2.5388,
      "theoretical_loss": 3.529934617099179,
      "tokens_seen": 1434841088
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.000285506519558676,
      "loss": 2.6986,
      "theoretical_loss": 3.5299202806664915,
      "tokens_seen": 1434906624
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028549648946840525,
      "loss": 2.6987,
      "theoretical_loss": 3.5299059450719015,
      "tokens_seen": 1434972160
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002854864593781344,
      "loss": 2.9928,
      "theoretical_loss": 3.529891610315322,
      "tokens_seen": 1435037696
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002854764292878636,
      "loss": 2.8229,
      "theoretical_loss": 3.5298772763966655,
      "tokens_seen": 1435103232
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028546639919759274,
      "loss": 2.7209,
      "theoretical_loss": 3.5298629433158455,
      "tokens_seen": 1435168768
    },
    {
      "epoch": 4.08,
      "objective/train/docs_used": 1583288,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.874298572540283,
      "objective/train/theoretical_loss": 3.5298486110727736,
      "objective/train/tokens_used": 1455694304,
      "theoretical_loss": 3.5298486110727736,
      "tokens_seen": 1435234304
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.000285456369107322,
      "loss": 2.8018,
      "theoretical_loss": 3.5298486110727736,
      "tokens_seen": 1435234304
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028544633901705116,
      "loss": 2.578,
      "theoretical_loss": 3.529834279667363,
      "tokens_seen": 1435299840
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028543630892678034,
      "loss": 2.7333,
      "theoretical_loss": 3.5298199490995277,
      "tokens_seen": 1435365376
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002854262788365095,
      "loss": 2.7832,
      "theoretical_loss": 3.529805619369179,
      "tokens_seen": 1435430912
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002854162487462387,
      "loss": 2.7618,
      "theoretical_loss": 3.52979129047623,
      "tokens_seen": 1435496448
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002854062186559679,
      "loss": 2.5082,
      "theoretical_loss": 3.5297769624205944,
      "tokens_seen": 1435561984
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002853961885656971,
      "loss": 2.6455,
      "theoretical_loss": 3.529762635202184,
      "tokens_seen": 1435627520
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028538615847542624,
      "loss": 2.5146,
      "theoretical_loss": 3.5297483088209125,
      "tokens_seen": 1435693056
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002853761283851555,
      "loss": 2.7314,
      "theoretical_loss": 3.529733983276692,
      "tokens_seen": 1435758592
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028536609829488466,
      "loss": 2.7992,
      "theoretical_loss": 3.529719658569436,
      "tokens_seen": 1435824128
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028535606820461384,
      "loss": 2.7624,
      "theoretical_loss": 3.529705334699057,
      "tokens_seen": 1435889664
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.000285346038114343,
      "loss": 2.5357,
      "theoretical_loss": 3.5296910116654683,
      "tokens_seen": 1435955200
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002853360080240722,
      "loss": 2.587,
      "theoretical_loss": 3.529676689468582,
      "tokens_seen": 1436020736
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002853259779338014,
      "loss": 2.6797,
      "theoretical_loss": 3.5296623681083124,
      "tokens_seen": 1436086272
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002853159478435306,
      "loss": 2.6747,
      "theoretical_loss": 3.5296480475845713,
      "tokens_seen": 1436151808
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028530591775325975,
      "loss": 2.9066,
      "theoretical_loss": 3.5296337278972723,
      "tokens_seen": 1436217344
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.000285295887662989,
      "loss": 2.9173,
      "theoretical_loss": 3.5296194090463278,
      "tokens_seen": 1436282880
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002852858575727181,
      "loss": 2.7209,
      "theoretical_loss": 3.5296050910316517,
      "tokens_seen": 1436348416
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028527582748244735,
      "loss": 2.6102,
      "theoretical_loss": 3.5295907738531564,
      "tokens_seen": 1436413952
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002852657973921766,
      "loss": 2.6939,
      "theoretical_loss": 3.529576457510754,
      "tokens_seen": 1436479488
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002852557673019057,
      "loss": 2.8535,
      "theoretical_loss": 3.5295621420043597,
      "tokens_seen": 1436545024
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028524573721163494,
      "loss": 2.7347,
      "theoretical_loss": 3.529547827333885,
      "tokens_seen": 1436610560
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028523570712136407,
      "loss": 2.7182,
      "theoretical_loss": 3.5295335134992434,
      "tokens_seen": 1436676096
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002852256770310933,
      "loss": 2.6752,
      "theoretical_loss": 3.5295192005003475,
      "tokens_seen": 1436741632
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002852156469408225,
      "loss": 2.9172,
      "theoretical_loss": 3.529504888337111,
      "tokens_seen": 1436807168
    },
    {
      "epoch": 4.08,
      "objective/train/docs_used": 1583910,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5404577255249023,
      "objective/train/theoretical_loss": 3.529490577009447,
      "objective/train/tokens_used": 1457332704,
      "theoretical_loss": 3.529490577009447,
      "tokens_seen": 1436872704
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028520561685055167,
      "loss": 2.4372,
      "theoretical_loss": 3.529490577009447,
      "tokens_seen": 1436872704
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028519558676028085,
      "loss": 2.4868,
      "theoretical_loss": 3.529476266517268,
      "tokens_seen": 1436938240
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002851855566700101,
      "loss": 2.4531,
      "theoretical_loss": 3.5294619568604877,
      "tokens_seen": 1437003776
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002851755265797392,
      "loss": 2.7788,
      "theoretical_loss": 3.529447648039019,
      "tokens_seen": 1437069312
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028516549648946845,
      "loss": 2.7561,
      "theoretical_loss": 3.5294333400527753,
      "tokens_seen": 1437134848
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028515546639919757,
      "loss": 2.8554,
      "theoretical_loss": 3.529419032901669,
      "tokens_seen": 1437200384
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002851454363089268,
      "loss": 2.7275,
      "theoretical_loss": 3.5294047265856143,
      "tokens_seen": 1437265920
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.000285135406218656,
      "loss": 2.5726,
      "theoretical_loss": 3.529390421104524,
      "tokens_seen": 1437331456
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028512537612838517,
      "loss": 2.8757,
      "theoretical_loss": 3.5293761164583106,
      "tokens_seen": 1437396992
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028511534603811435,
      "loss": 2.6895,
      "theoretical_loss": 3.5293618126468886,
      "tokens_seen": 1437462528
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028510531594784353,
      "loss": 2.6364,
      "theoretical_loss": 3.52934750967017,
      "tokens_seen": 1437528064
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002850952858575727,
      "loss": 2.6436,
      "theoretical_loss": 3.529333207528069,
      "tokens_seen": 1437593600
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028508525576730195,
      "loss": 2.4625,
      "theoretical_loss": 3.5293189062204986,
      "tokens_seen": 1437659136
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002850752256770311,
      "loss": 2.7444,
      "theoretical_loss": 3.529304605747371,
      "tokens_seen": 1437724672
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002850651955867603,
      "loss": 2.6947,
      "theoretical_loss": 3.529290306108601,
      "tokens_seen": 1437790208
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002850551654964895,
      "loss": 2.6257,
      "theoretical_loss": 3.5292760073041016,
      "tokens_seen": 1437855744
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002850451354062187,
      "loss": 2.544,
      "theoretical_loss": 3.5292617093337855,
      "tokens_seen": 1437921280
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028503510531594785,
      "loss": 2.9978,
      "theoretical_loss": 3.5292474121975665,
      "tokens_seen": 1437986816
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028502507522567704,
      "loss": 2.4569,
      "theoretical_loss": 3.5292331158953574,
      "tokens_seen": 1438052352
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002850150451354062,
      "loss": 2.5917,
      "theoretical_loss": 3.5292188204270722,
      "tokens_seen": 1438117888
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028500501504513545,
      "loss": 2.7325,
      "theoretical_loss": 3.5292045257926237,
      "tokens_seen": 1438183424
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002849949849548646,
      "loss": 2.8818,
      "theoretical_loss": 3.5291902319919255,
      "tokens_seen": 1438248960
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002849849548645938,
      "loss": 2.8042,
      "theoretical_loss": 3.529175939024891,
      "tokens_seen": 1438314496
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028497492477432294,
      "loss": 2.7066,
      "theoretical_loss": 3.529161646891434,
      "tokens_seen": 1438380032
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002849648946840522,
      "loss": 2.6888,
      "theoretical_loss": 3.529147355591467,
      "tokens_seen": 1438445568
    },
    {
      "epoch": 4.08,
      "objective/train/docs_used": 1585264,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9627788066864014,
      "objective/train/theoretical_loss": 3.5291330651249044,
      "objective/train/tokens_used": 1458971104,
      "theoretical_loss": 3.5291330651249044,
      "tokens_seen": 1438511104
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028495486459378136,
      "loss": 2.6728,
      "theoretical_loss": 3.5291330651249044,
      "tokens_seen": 1438511104
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028494483450351054,
      "loss": 2.6711,
      "theoretical_loss": 3.529118775491659,
      "tokens_seen": 1438576640
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002849348044132397,
      "loss": 2.7203,
      "theoretical_loss": 3.5291044866916437,
      "tokens_seen": 1438642176
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002849247743229689,
      "loss": 2.835,
      "theoretical_loss": 3.5290901987247736,
      "tokens_seen": 1438707712
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002849147442326981,
      "loss": 2.7568,
      "theoretical_loss": 3.5290759115909607,
      "tokens_seen": 1438773248
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002849047141424273,
      "loss": 2.7452,
      "theoretical_loss": 3.5290616252901192,
      "tokens_seen": 1438838784
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028489468405215644,
      "loss": 2.5437,
      "theoretical_loss": 3.5290473398221627,
      "tokens_seen": 1438904320
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002848846539618857,
      "loss": 2.9406,
      "theoretical_loss": 3.529033055187004,
      "tokens_seen": 1438969856
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028487462387161486,
      "loss": 2.7922,
      "theoretical_loss": 3.529018771384557,
      "tokens_seen": 1439035392
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028486459378134404,
      "loss": 2.6864,
      "theoretical_loss": 3.529004488414736,
      "tokens_seen": 1439100928
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002848545636910732,
      "loss": 2.873,
      "theoretical_loss": 3.528990206277453,
      "tokens_seen": 1439166464
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002848445336008024,
      "loss": 2.725,
      "theoretical_loss": 3.528975924972623,
      "tokens_seen": 1439232000
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002848345035105316,
      "loss": 2.6527,
      "theoretical_loss": 3.528961644500159,
      "tokens_seen": 1439297536
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002848244734202608,
      "loss": 2.5969,
      "theoretical_loss": 3.528947364859974,
      "tokens_seen": 1439363072
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028481444332998995,
      "loss": 2.6177,
      "theoretical_loss": 3.5289330860519823,
      "tokens_seen": 1439428608
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002848044132397192,
      "loss": 2.7198,
      "theoretical_loss": 3.5289188080760976,
      "tokens_seen": 1439494144
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002847943831494483,
      "loss": 2.4022,
      "theoretical_loss": 3.5289045309322336,
      "tokens_seen": 1439559680
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028478435305917755,
      "loss": 2.7359,
      "theoretical_loss": 3.528890254620303,
      "tokens_seen": 1439625216
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002847743229689067,
      "loss": 2.6777,
      "theoretical_loss": 3.5288759791402207,
      "tokens_seen": 1439690752
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002847642928786359,
      "loss": 2.518,
      "theoretical_loss": 3.5288617044918995,
      "tokens_seen": 1439756288
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002847542627883651,
      "loss": 2.7267,
      "theoretical_loss": 3.5288474306752535,
      "tokens_seen": 1439821824
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028474423269809427,
      "loss": 2.708,
      "theoretical_loss": 3.5288331576901966,
      "tokens_seen": 1439887360
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028473420260782345,
      "loss": 2.6584,
      "theoretical_loss": 3.5288188855366416,
      "tokens_seen": 1439952896
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002847241725175527,
      "loss": 2.5356,
      "theoretical_loss": 3.5288046142145033,
      "tokens_seen": 1440018432
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002847141424272818,
      "loss": 2.7895,
      "theoretical_loss": 3.5287903437236947,
      "tokens_seen": 1440083968
    },
    {
      "epoch": 4.08,
      "objective/train/docs_used": 1585991,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4762494564056396,
      "objective/train/theoretical_loss": 3.5287760740641296,
      "objective/train/tokens_used": 1460609504,
      "theoretical_loss": 3.5287760740641296,
      "tokens_seen": 1440149504
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028470411233701105,
      "loss": 2.4325,
      "theoretical_loss": 3.5287760740641296,
      "tokens_seen": 1440149504
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028469408224674023,
      "loss": 2.9182,
      "theoretical_loss": 3.528761805235722,
      "tokens_seen": 1440215040
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002846840521564694,
      "loss": 2.6841,
      "theoretical_loss": 3.5287475372383854,
      "tokens_seen": 1440280576
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002846740220661986,
      "loss": 2.6416,
      "theoretical_loss": 3.528733270072034,
      "tokens_seen": 1440346112
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002846639919759278,
      "loss": 2.7276,
      "theoretical_loss": 3.5287190037365814,
      "tokens_seen": 1440411648
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028465396188565695,
      "loss": 2.7666,
      "theoretical_loss": 3.528704738231941,
      "tokens_seen": 1440477184
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002846439317953862,
      "loss": 2.4176,
      "theoretical_loss": 3.5286904735580276,
      "tokens_seen": 1440542720
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002846339017051153,
      "loss": 2.6206,
      "theoretical_loss": 3.528676209714754,
      "tokens_seen": 1440608256
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028462387161484455,
      "loss": 2.676,
      "theoretical_loss": 3.5286619467020346,
      "tokens_seen": 1440673792
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002846138415245737,
      "loss": 2.9246,
      "theoretical_loss": 3.5286476845197834,
      "tokens_seen": 1440739328
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002846038114343029,
      "loss": 2.565,
      "theoretical_loss": 3.5286334231679133,
      "tokens_seen": 1440804864
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002845937813440321,
      "loss": 2.5963,
      "theoretical_loss": 3.5286191626463395,
      "tokens_seen": 1440870400
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002845837512537613,
      "loss": 2.309,
      "theoretical_loss": 3.5286049029549753,
      "tokens_seen": 1440935936
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028457372116349046,
      "loss": 2.7321,
      "theoretical_loss": 3.5285906440937342,
      "tokens_seen": 1441001472
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002845636910732197,
      "loss": 2.9154,
      "theoretical_loss": 3.528576386062531,
      "tokens_seen": 1441067008
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002845536609829488,
      "loss": 2.791,
      "theoretical_loss": 3.528562128861279,
      "tokens_seen": 1441132544
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028454363089267805,
      "loss": 2.5997,
      "theoretical_loss": 3.5285478724898924,
      "tokens_seen": 1441198080
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028453360080240724,
      "loss": 2.6299,
      "theoretical_loss": 3.528533616948285,
      "tokens_seen": 1441263616
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002845235707121364,
      "loss": 2.5887,
      "theoretical_loss": 3.5285193622363704,
      "tokens_seen": 1441329152
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028451354062186565,
      "loss": 2.6146,
      "theoretical_loss": 3.5285051083540635,
      "tokens_seen": 1441394688
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002845035105315948,
      "loss": 2.6975,
      "theoretical_loss": 3.528490855301278,
      "tokens_seen": 1441460224
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.000284493480441324,
      "loss": 2.5712,
      "theoretical_loss": 3.5284766030779275,
      "tokens_seen": 1441525760
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028448345035105314,
      "loss": 2.7834,
      "theoretical_loss": 3.528462351683926,
      "tokens_seen": 1441591296
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002844734202607824,
      "loss": 2.5477,
      "theoretical_loss": 3.5284481011191886,
      "tokens_seen": 1441656832
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028446339017051156,
      "loss": 2.693,
      "theoretical_loss": 3.528433851383628,
      "tokens_seen": 1441722368
    },
    {
      "epoch": 4.08,
      "objective/train/docs_used": 1587458,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6036527156829834,
      "objective/train/theoretical_loss": 3.5284196024771592,
      "objective/train/tokens_used": 1462247904,
      "theoretical_loss": 3.5284196024771592,
      "tokens_seen": 1441787904
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028445336008024074,
      "loss": 2.7933,
      "theoretical_loss": 3.5284196024771592,
      "tokens_seen": 1441787904
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002844433299899699,
      "loss": 2.87,
      "theoretical_loss": 3.5284053543996956,
      "tokens_seen": 1441853440
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002844332998996991,
      "loss": 2.5991,
      "theoretical_loss": 3.528391107151152,
      "tokens_seen": 1441918976
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002844232698094283,
      "loss": 2.6916,
      "theoretical_loss": 3.5283768607314414,
      "tokens_seen": 1441984512
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002844132397191575,
      "loss": 2.7467,
      "theoretical_loss": 3.5283626151404794,
      "tokens_seen": 1442050048
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028440320962888664,
      "loss": 2.6921,
      "theoretical_loss": 3.528348370378179,
      "tokens_seen": 1442115584
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002843931795386159,
      "loss": 2.8085,
      "theoretical_loss": 3.528334126444455,
      "tokens_seen": 1442181120
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028438314944834506,
      "loss": 2.7245,
      "theoretical_loss": 3.528319883339221,
      "tokens_seen": 1442246656
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028437311935807424,
      "loss": 2.7377,
      "theoretical_loss": 3.528305641062391,
      "tokens_seen": 1442312192
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002843630892678034,
      "loss": 2.6306,
      "theoretical_loss": 3.5282913996138805,
      "tokens_seen": 1442377728
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002843530591775326,
      "loss": 2.6701,
      "theoretical_loss": 3.5282771589936024,
      "tokens_seen": 1442443264
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002843430290872618,
      "loss": 2.6255,
      "theoretical_loss": 3.528262919201471,
      "tokens_seen": 1442508800
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.000284332998996991,
      "loss": 2.5292,
      "theoretical_loss": 3.528248680237401,
      "tokens_seen": 1442574336
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028432296890672015,
      "loss": 2.5836,
      "theoretical_loss": 3.528234442101307,
      "tokens_seen": 1442639872
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002843129388164494,
      "loss": 2.5171,
      "theoretical_loss": 3.5282202047931017,
      "tokens_seen": 1442705408
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002843029087261785,
      "loss": 2.6853,
      "theoretical_loss": 3.5282059683127014,
      "tokens_seen": 1442770944
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028429287863590775,
      "loss": 2.6028,
      "theoretical_loss": 3.528191732660019,
      "tokens_seen": 1442836480
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002842828485456369,
      "loss": 2.6122,
      "theoretical_loss": 3.528177497834969,
      "tokens_seen": 1442902016
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002842728184553661,
      "loss": 2.7187,
      "theoretical_loss": 3.528163263837466,
      "tokens_seen": 1442967552
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002842627883650953,
      "loss": 2.6089,
      "theoretical_loss": 3.5281490306674237,
      "tokens_seen": 1443033088
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028425275827482447,
      "loss": 2.5862,
      "theoretical_loss": 3.528134798324757,
      "tokens_seen": 1443098624
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028424272818455365,
      "loss": 2.5758,
      "theoretical_loss": 3.5281205668093802,
      "tokens_seen": 1443164160
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002842326980942829,
      "loss": 2.8418,
      "theoretical_loss": 3.5281063361212075,
      "tokens_seen": 1443229696
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.000284222668004012,
      "loss": 2.6395,
      "theoretical_loss": 3.5280921062601536,
      "tokens_seen": 1443295232
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028421263791374125,
      "loss": 2.7947,
      "theoretical_loss": 3.5280778772261314,
      "tokens_seen": 1443360768
    },
    {
      "epoch": 4.08,
      "objective/train/docs_used": 1588273,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.190006971359253,
      "objective/train/theoretical_loss": 3.5280636490190576,
      "objective/train/tokens_used": 1463886304,
      "theoretical_loss": 3.5280636490190576,
      "tokens_seen": 1443426304
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028420260782347043,
      "loss": 2.4971,
      "theoretical_loss": 3.5280636490190576,
      "tokens_seen": 1443426304
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002841925777331996,
      "loss": 2.6572,
      "theoretical_loss": 3.528049421638845,
      "tokens_seen": 1443491840
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002841825476429288,
      "loss": 2.7693,
      "theoretical_loss": 3.5280351950854087,
      "tokens_seen": 1443557376
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.000284172517552658,
      "loss": 2.6632,
      "theoretical_loss": 3.5280209693586624,
      "tokens_seen": 1443622912
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028416248746238715,
      "loss": 2.5126,
      "theoretical_loss": 3.528006744458522,
      "tokens_seen": 1443688448
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002841524573721164,
      "loss": 2.466,
      "theoretical_loss": 3.5279925203849,
      "tokens_seen": 1443753984
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002841424272818455,
      "loss": 2.7694,
      "theoretical_loss": 3.527978297137712,
      "tokens_seen": 1443819520
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028413239719157475,
      "loss": 2.7568,
      "theoretical_loss": 3.5279640747168726,
      "tokens_seen": 1443885056
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002841223671013039,
      "loss": 2.8102,
      "theoretical_loss": 3.5279498531222955,
      "tokens_seen": 1443950592
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002841123370110331,
      "loss": 2.8483,
      "theoretical_loss": 3.5279356323538957,
      "tokens_seen": 1444016128
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002841023069207623,
      "loss": 2.6838,
      "theoretical_loss": 3.527921412411588,
      "tokens_seen": 1444081664
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002840922768304915,
      "loss": 2.4802,
      "theoretical_loss": 3.527907193295287,
      "tokens_seen": 1444147200
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028408224674022066,
      "loss": 2.6184,
      "theoretical_loss": 3.5278929750049057,
      "tokens_seen": 1444212736
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002840722166499499,
      "loss": 2.8073,
      "theoretical_loss": 3.527878757540361,
      "tokens_seen": 1444278272
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.000284062186559679,
      "loss": 2.8705,
      "theoretical_loss": 3.5278645409015654,
      "tokens_seen": 1444343808
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028405215646940825,
      "loss": 2.7107,
      "theoretical_loss": 3.527850325088435,
      "tokens_seen": 1444409344
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002840421263791374,
      "loss": 2.6718,
      "theoretical_loss": 3.5278361101008833,
      "tokens_seen": 1444474880
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002840320962888666,
      "loss": 2.7017,
      "theoretical_loss": 3.5278218959388257,
      "tokens_seen": 1444540416
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002840220661985958,
      "loss": 2.7548,
      "theoretical_loss": 3.527807682602176,
      "tokens_seen": 1444605952
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.000284012036108325,
      "loss": 2.4647,
      "theoretical_loss": 3.5277934700908498,
      "tokens_seen": 1444671488
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028400200601805416,
      "loss": 2.52,
      "theoretical_loss": 3.5277792584047605,
      "tokens_seen": 1444737024
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028399197592778334,
      "loss": 2.7247,
      "theoretical_loss": 3.5277650475438236,
      "tokens_seen": 1444802560
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002839819458375125,
      "loss": 2.6323,
      "theoretical_loss": 3.5277508375079543,
      "tokens_seen": 1444868096
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028397191574724176,
      "loss": 2.763,
      "theoretical_loss": 3.527736628297066,
      "tokens_seen": 1444933632
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002839618856569709,
      "loss": 2.557,
      "theoretical_loss": 3.5277224199110746,
      "tokens_seen": 1444999168
    },
    {
      "epoch": 4.08,
      "objective/train/docs_used": 1589516,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.031261682510376,
      "objective/train/theoretical_loss": 3.5277082123498937,
      "objective/train/tokens_used": 1465524704,
      "theoretical_loss": 3.5277082123498937,
      "tokens_seen": 1445064704
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002839518555667001,
      "loss": 2.8223,
      "theoretical_loss": 3.5277082123498937,
      "tokens_seen": 1445064704
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028394182547642925,
      "loss": 2.7678,
      "theoretical_loss": 3.5276940056134385,
      "tokens_seen": 1445130240
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002839317953861585,
      "loss": 2.5075,
      "theoretical_loss": 3.527679799701624,
      "tokens_seen": 1445195776
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028392176529588766,
      "loss": 2.5564,
      "theoretical_loss": 3.5276655946143642,
      "tokens_seen": 1445261312
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028391173520561684,
      "loss": 2.7625,
      "theoretical_loss": 3.5276513903515747,
      "tokens_seen": 1445326848
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.000283901705115346,
      "loss": 2.6866,
      "theoretical_loss": 3.5276371869131697,
      "tokens_seen": 1445392384
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028389167502507526,
      "loss": 2.5275,
      "theoretical_loss": 3.5276229842990645,
      "tokens_seen": 1445457920
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002838816449348044,
      "loss": 2.4997,
      "theoretical_loss": 3.5276087825091738,
      "tokens_seen": 1445523456
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002838716148445336,
      "loss": 2.8864,
      "theoretical_loss": 3.5275945815434113,
      "tokens_seen": 1445588992
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028386158475426275,
      "loss": 2.5567,
      "theoretical_loss": 3.5275803814016937,
      "tokens_seen": 1445654528
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.000283851554663992,
      "loss": 2.5588,
      "theoretical_loss": 3.527566182083934,
      "tokens_seen": 1445720064
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028384152457372117,
      "loss": 2.6456,
      "theoretical_loss": 3.5275519835900484,
      "tokens_seen": 1445785600
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028383149448345035,
      "loss": 2.6197,
      "theoretical_loss": 3.527537785919951,
      "tokens_seen": 1445851136
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028382146439317953,
      "loss": 2.7781,
      "theoretical_loss": 3.527523589073557,
      "tokens_seen": 1445916672
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002838114343029087,
      "loss": 2.4368,
      "theoretical_loss": 3.527509393050781,
      "tokens_seen": 1445982208
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002838014042126379,
      "loss": 2.6069,
      "theoretical_loss": 3.527495197851538,
      "tokens_seen": 1446047744
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002837913741223671,
      "loss": 2.5526,
      "theoretical_loss": 3.527481003475743,
      "tokens_seen": 1446113280
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002837813440320963,
      "loss": 2.7014,
      "theoretical_loss": 3.527466809923311,
      "tokens_seen": 1446178816
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002837713139418255,
      "loss": 2.5824,
      "theoretical_loss": 3.5274526171941574,
      "tokens_seen": 1446244352
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028376128385155467,
      "loss": 2.5125,
      "theoretical_loss": 3.527438425288196,
      "tokens_seen": 1446309888
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028375125376128385,
      "loss": 2.6314,
      "theoretical_loss": 3.527424234205342,
      "tokens_seen": 1446375424
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002837412236710131,
      "loss": 2.3199,
      "theoretical_loss": 3.527410043945511,
      "tokens_seen": 1446440960
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002837311935807422,
      "loss": 2.8419,
      "theoretical_loss": 3.527395854508618,
      "tokens_seen": 1446506496
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028372116349047145,
      "loss": 2.8332,
      "theoretical_loss": 3.527381665894577,
      "tokens_seen": 1446572032
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028371113340020063,
      "loss": 2.5837,
      "theoretical_loss": 3.5273674781033044,
      "tokens_seen": 1446637568
    },
    {
      "epoch": 4.08,
      "objective/train/docs_used": 1590115,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.46427321434021,
      "objective/train/theoretical_loss": 3.5273532911347143,
      "objective/train/tokens_used": 1467163104,
      "theoretical_loss": 3.5273532911347143,
      "tokens_seen": 1446703104
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002837011033099298,
      "loss": 2.628,
      "theoretical_loss": 3.5273532911347143,
      "tokens_seen": 1446703104
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.000283691073219659,
      "loss": 2.4291,
      "theoretical_loss": 3.5273391049887213,
      "tokens_seen": 1446768640
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002836810431293882,
      "loss": 2.6043,
      "theoretical_loss": 3.527324919665242,
      "tokens_seen": 1446834176
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028367101303911735,
      "loss": 2.4816,
      "theoretical_loss": 3.52731073516419,
      "tokens_seen": 1446899712
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002836609829488466,
      "loss": 2.534,
      "theoretical_loss": 3.527296551485481,
      "tokens_seen": 1446965248
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002836509528585757,
      "loss": 2.5662,
      "theoretical_loss": 3.5272823686290296,
      "tokens_seen": 1447030784
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028364092276830495,
      "loss": 2.6041,
      "theoretical_loss": 3.5272681865947515,
      "tokens_seen": 1447096320
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002836308926780341,
      "loss": 2.6108,
      "theoretical_loss": 3.527254005382562,
      "tokens_seen": 1447161856
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002836208625877633,
      "loss": 2.2632,
      "theoretical_loss": 3.5272398249923755,
      "tokens_seen": 1447227392
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002836108324974925,
      "loss": 2.6465,
      "theoretical_loss": 3.5272256454241075,
      "tokens_seen": 1447292928
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002836008024072217,
      "loss": 2.7067,
      "theoretical_loss": 3.527211466677673,
      "tokens_seen": 1447358464
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028359077231695086,
      "loss": 2.3787,
      "theoretical_loss": 3.527197288752987,
      "tokens_seen": 1447424000
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002835807422266801,
      "loss": 2.5942,
      "theoretical_loss": 3.527183111649965,
      "tokens_seen": 1447489536
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002835707121364092,
      "loss": 2.8501,
      "theoretical_loss": 3.5271689353685223,
      "tokens_seen": 1447555072
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028356068204613846,
      "loss": 2.8219,
      "theoretical_loss": 3.5271547599085737,
      "tokens_seen": 1447620608
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002835506519558676,
      "loss": 2.7338,
      "theoretical_loss": 3.527140585270035,
      "tokens_seen": 1447686144
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002835406218655968,
      "loss": 2.8349,
      "theoretical_loss": 3.5271264114528202,
      "tokens_seen": 1447751680
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.000283530591775326,
      "loss": 2.5583,
      "theoretical_loss": 3.527112238456846,
      "tokens_seen": 1447817216
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002835205616850552,
      "loss": 2.5847,
      "theoretical_loss": 3.527098066282027,
      "tokens_seen": 1447882752
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028351053159478436,
      "loss": 2.5197,
      "theoretical_loss": 3.5270838949282783,
      "tokens_seen": 1447948288
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028350050150451354,
      "loss": 2.5117,
      "theoretical_loss": 3.5270697243955156,
      "tokens_seen": 1448013824
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002834904714142427,
      "loss": 2.6788,
      "theoretical_loss": 3.5270555546836535,
      "tokens_seen": 1448079360
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028348044132397196,
      "loss": 2.8404,
      "theoretical_loss": 3.527041385792608,
      "tokens_seen": 1448144896
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002834704112337011,
      "loss": 2.6377,
      "theoretical_loss": 3.527027217722294,
      "tokens_seen": 1448210432
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002834603811434303,
      "loss": 2.5585,
      "theoretical_loss": 3.527013050472627,
      "tokens_seen": 1448275968
    },
    {
      "epoch": 4.08,
      "objective/train/docs_used": 1591064,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.98321795463562,
      "objective/train/theoretical_loss": 3.5269988840435222,
      "objective/train/tokens_used": 1468801504,
      "theoretical_loss": 3.5269988840435222,
      "tokens_seen": 1448341504
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028345035105315945,
      "loss": 2.9032,
      "theoretical_loss": 3.5269988840435222,
      "tokens_seen": 1448341504
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002834403209628887,
      "loss": 2.719,
      "theoretical_loss": 3.5269847184348952,
      "tokens_seen": 1448407040
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028343029087261786,
      "loss": 2.6729,
      "theoretical_loss": 3.526970553646661,
      "tokens_seen": 1448472576
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028342026078234705,
      "loss": 2.483,
      "theoretical_loss": 3.526956389678735,
      "tokens_seen": 1448538112
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002834102306920762,
      "loss": 2.5919,
      "theoretical_loss": 3.526942226531033,
      "tokens_seen": 1448603648
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028340020060180546,
      "loss": 2.6517,
      "theoretical_loss": 3.52692806420347,
      "tokens_seen": 1448669184
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002833901705115346,
      "loss": 2.5691,
      "theoretical_loss": 3.526913902695962,
      "tokens_seen": 1448734720
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002833801404212638,
      "loss": 2.924,
      "theoretical_loss": 3.526899742008423,
      "tokens_seen": 1448800256
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028337011033099295,
      "loss": 2.8697,
      "theoretical_loss": 3.5268855821407703,
      "tokens_seen": 1448865792
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002833600802407222,
      "loss": 2.475,
      "theoretical_loss": 3.526871423092918,
      "tokens_seen": 1448931328
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028335005015045137,
      "loss": 2.416,
      "theoretical_loss": 3.5268572648647822,
      "tokens_seen": 1448996864
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028334002006018055,
      "loss": 2.7226,
      "theoretical_loss": 3.526843107456278,
      "tokens_seen": 1449062400
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028332998996990973,
      "loss": 2.537,
      "theoretical_loss": 3.526828950867321,
      "tokens_seen": 1449127936
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002833199598796389,
      "loss": 2.5673,
      "theoretical_loss": 3.526814795097827,
      "tokens_seen": 1449193472
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002833099297893681,
      "loss": 2.6473,
      "theoretical_loss": 3.5268006401477114,
      "tokens_seen": 1449259008
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002832998996990973,
      "loss": 2.5081,
      "theoretical_loss": 3.5267864860168894,
      "tokens_seen": 1449324544
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028328986960882645,
      "loss": 2.6657,
      "theoretical_loss": 3.526772332705277,
      "tokens_seen": 1449390080
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002832798395185557,
      "loss": 2.6562,
      "theoretical_loss": 3.5267581802127888,
      "tokens_seen": 1449455616
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002832698094282848,
      "loss": 2.634,
      "theoretical_loss": 3.5267440285393414,
      "tokens_seen": 1449521152
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028325977933801405,
      "loss": 2.645,
      "theoretical_loss": 3.52672987768485,
      "tokens_seen": 1449586688
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028324974924774323,
      "loss": 2.623,
      "theoretical_loss": 3.5267157276492305,
      "tokens_seen": 1449652224
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002832397191574724,
      "loss": 2.714,
      "theoretical_loss": 3.526701578432398,
      "tokens_seen": 1449717760
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002832296890672016,
      "loss": 2.6843,
      "theoretical_loss": 3.5266874300342685,
      "tokens_seen": 1449783296
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028321965897693083,
      "loss": 2.6982,
      "theoretical_loss": 3.526673282454757,
      "tokens_seen": 1449848832
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028320962888665996,
      "loss": 2.6372,
      "theoretical_loss": 3.5266591356937793,
      "tokens_seen": 1449914368
    },
    {
      "epoch": 4.08,
      "objective/train/docs_used": 1591697,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.471153736114502,
      "objective/train/theoretical_loss": 3.5266449897512517,
      "objective/train/tokens_used": 1470439904,
      "theoretical_loss": 3.5266449897512517,
      "tokens_seen": 1449979904
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002831995987963892,
      "loss": 2.5672,
      "theoretical_loss": 3.5266449897512517,
      "tokens_seen": 1449979904
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002831895687061183,
      "loss": 2.8299,
      "theoretical_loss": 3.52663084462709,
      "tokens_seen": 1450045440
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028317953861584755,
      "loss": 2.8545,
      "theoretical_loss": 3.5266167003212088,
      "tokens_seen": 1450110976
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028316950852557674,
      "loss": 2.716,
      "theoretical_loss": 3.5266025568335246,
      "tokens_seen": 1450176512
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002831594784353059,
      "loss": 2.8401,
      "theoretical_loss": 3.5265884141639523,
      "tokens_seen": 1450242048
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002831494483450351,
      "loss": 2.4752,
      "theoretical_loss": 3.5265742723124087,
      "tokens_seen": 1450307584
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002831394182547643,
      "loss": 2.6708,
      "theoretical_loss": 3.5265601312788086,
      "tokens_seen": 1450373120
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028312938816449346,
      "loss": 2.5739,
      "theoretical_loss": 3.5265459910630685,
      "tokens_seen": 1450438656
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002831193580742227,
      "loss": 2.3183,
      "theoretical_loss": 3.526531851665103,
      "tokens_seen": 1450504192
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002831093279839518,
      "loss": 2.6056,
      "theoretical_loss": 3.5265177130848295,
      "tokens_seen": 1450569728
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028309929789368106,
      "loss": 2.3986,
      "theoretical_loss": 3.5265035753221623,
      "tokens_seen": 1450635264
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002830892678034102,
      "loss": 2.254,
      "theoretical_loss": 3.526489438377018,
      "tokens_seen": 1450700800
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002830792377131394,
      "loss": 2.3774,
      "theoretical_loss": 3.5264753022493123,
      "tokens_seen": 1450766336
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002830692076228686,
      "loss": 2.6581,
      "theoretical_loss": 3.5264611669389603,
      "tokens_seen": 1450831872
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002830591775325978,
      "loss": 2.7144,
      "theoretical_loss": 3.5264470324458785,
      "tokens_seen": 1450897408
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028304914744232696,
      "loss": 2.722,
      "theoretical_loss": 3.5264328987699827,
      "tokens_seen": 1450962944
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002830391173520562,
      "loss": 2.7777,
      "theoretical_loss": 3.526418765911189,
      "tokens_seen": 1451028480
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002830290872617854,
      "loss": 2.5871,
      "theoretical_loss": 3.526404633869413,
      "tokens_seen": 1451094016
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028301905717151456,
      "loss": 2.676,
      "theoretical_loss": 3.5263905026445705,
      "tokens_seen": 1451159552
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028300902708124374,
      "loss": 2.4169,
      "theoretical_loss": 3.5263763722365766,
      "tokens_seen": 1451225088
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002829989969909729,
      "loss": 2.6727,
      "theoretical_loss": 3.526362242645349,
      "tokens_seen": 1451290624
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028298896690070216,
      "loss": 2.5742,
      "theoretical_loss": 3.526348113870802,
      "tokens_seen": 1451356160
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002829789368104313,
      "loss": 2.6444,
      "theoretical_loss": 3.526333985912852,
      "tokens_seen": 1451421696
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002829689067201605,
      "loss": 2.5713,
      "theoretical_loss": 3.5263198587714157,
      "tokens_seen": 1451487232
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028295887662988965,
      "loss": 2.6759,
      "theoretical_loss": 3.5263057324464078,
      "tokens_seen": 1451552768
    },
    {
      "epoch": 4.08,
      "objective/train/docs_used": 1592735,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.422314405441284,
      "objective/train/theoretical_loss": 3.526291606937745,
      "objective/train/tokens_used": 1472078304,
      "theoretical_loss": 3.526291606937745,
      "tokens_seen": 1451618304
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002829488465396189,
      "loss": 2.6869,
      "theoretical_loss": 3.526291606937745,
      "tokens_seen": 1451618304
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028293881644934806,
      "loss": 2.4801,
      "theoretical_loss": 3.5262774822453435,
      "tokens_seen": 1451683840
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028292878635907725,
      "loss": 2.5524,
      "theoretical_loss": 3.5262633583691185,
      "tokens_seen": 1451749376
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002829187562688064,
      "loss": 2.6287,
      "theoretical_loss": 3.5262492353089865,
      "tokens_seen": 1451814912
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028290872617853566,
      "loss": 2.9081,
      "theoretical_loss": 3.5262351130648635,
      "tokens_seen": 1451880448
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002828986960882648,
      "loss": 2.4654,
      "theoretical_loss": 3.5262209916366656,
      "tokens_seen": 1451945984
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.000282888665997994,
      "loss": 2.6926,
      "theoretical_loss": 3.5262068710243084,
      "tokens_seen": 1452011520
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028287863590772315,
      "loss": 2.866,
      "theoretical_loss": 3.5261927512277085,
      "tokens_seen": 1452077056
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002828686058174524,
      "loss": 2.7792,
      "theoretical_loss": 3.5261786322467814,
      "tokens_seen": 1452142592
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028285857572718157,
      "loss": 2.631,
      "theoretical_loss": 3.5261645140814437,
      "tokens_seen": 1452208128
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028284854563691075,
      "loss": 2.7531,
      "theoretical_loss": 3.526150396731611,
      "tokens_seen": 1452273664
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028283851554663993,
      "loss": 2.6413,
      "theoretical_loss": 3.5261362801972,
      "tokens_seen": 1452339200
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002828284854563691,
      "loss": 2.46,
      "theoretical_loss": 3.526122164478126,
      "tokens_seen": 1452404736
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002828184553660983,
      "loss": 2.7587,
      "theoretical_loss": 3.526108049574306,
      "tokens_seen": 1452470272
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002828084252758275,
      "loss": 2.7158,
      "theoretical_loss": 3.5260939354856555,
      "tokens_seen": 1452535808
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028279839518555665,
      "loss": 2.7085,
      "theoretical_loss": 3.5260798222120906,
      "tokens_seen": 1452601344
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002827883650952859,
      "loss": 2.5908,
      "theoretical_loss": 3.526065709753528,
      "tokens_seen": 1452666880
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.000282778335005015,
      "loss": 2.9158,
      "theoretical_loss": 3.5260515981098832,
      "tokens_seen": 1452732416
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028276830491474425,
      "loss": 2.6908,
      "theoretical_loss": 3.5260374872810734,
      "tokens_seen": 1452797952
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028275827482447343,
      "loss": 2.6313,
      "theoretical_loss": 3.5260233772670135,
      "tokens_seen": 1452863488
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002827482447342026,
      "loss": 2.5825,
      "theoretical_loss": 3.5260092680676207,
      "tokens_seen": 1452929024
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002827382146439318,
      "loss": 2.6314,
      "theoretical_loss": 3.525995159682811,
      "tokens_seen": 1452994560
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028272818455366103,
      "loss": 2.7025,
      "theoretical_loss": 3.525981052112501,
      "tokens_seen": 1453060096
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028271815446339016,
      "loss": 2.6858,
      "theoretical_loss": 3.5259669453566054,
      "tokens_seen": 1453125632
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002827081243731194,
      "loss": 2.5014,
      "theoretical_loss": 3.525952839415042,
      "tokens_seen": 1453191168
    },
    {
      "epoch": 4.08,
      "objective/train/docs_used": 1593470,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.433061122894287,
      "objective/train/theoretical_loss": 3.5259387342877266,
      "objective/train/tokens_used": 1473716704,
      "theoretical_loss": 3.5259387342877266,
      "tokens_seen": 1453256704
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002826980942828485,
      "loss": 2.6465,
      "theoretical_loss": 3.5259387342877266,
      "tokens_seen": 1453256704
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028268806419257775,
      "loss": 2.4855,
      "theoretical_loss": 3.5259246299745755,
      "tokens_seen": 1453322240
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028267803410230694,
      "loss": 2.6825,
      "theoretical_loss": 3.525910526475505,
      "tokens_seen": 1453387776
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002826680040120361,
      "loss": 2.7234,
      "theoretical_loss": 3.5258964237904316,
      "tokens_seen": 1453453312
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002826579739217653,
      "loss": 2.2644,
      "theoretical_loss": 3.525882321919271,
      "tokens_seen": 1453518848
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002826479438314945,
      "loss": 2.8498,
      "theoretical_loss": 3.5258682208619403,
      "tokens_seen": 1453584384
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028263791374122366,
      "loss": 2.4475,
      "theoretical_loss": 3.525854120618355,
      "tokens_seen": 1453649920
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002826278836509529,
      "loss": 2.6162,
      "theoretical_loss": 3.525840021188432,
      "tokens_seen": 1453715456
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.000282617853560682,
      "loss": 2.4738,
      "theoretical_loss": 3.525825922572088,
      "tokens_seen": 1453780992
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028260782347041126,
      "loss": 2.4866,
      "theoretical_loss": 3.5258118247692387,
      "tokens_seen": 1453846528
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002825977933801404,
      "loss": 2.4781,
      "theoretical_loss": 3.5257977277798007,
      "tokens_seen": 1453912064
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002825877632898696,
      "loss": 2.5941,
      "theoretical_loss": 3.5257836316036903,
      "tokens_seen": 1453977600
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002825777331995988,
      "loss": 2.4183,
      "theoretical_loss": 3.5257695362408246,
      "tokens_seen": 1454043136
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.000282567703109328,
      "loss": 2.6699,
      "theoretical_loss": 3.5257554416911194,
      "tokens_seen": 1454108672
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028255767301905716,
      "loss": 2.4834,
      "theoretical_loss": 3.525741347954491,
      "tokens_seen": 1454174208
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002825476429287864,
      "loss": 2.6883,
      "theoretical_loss": 3.5257272550308563,
      "tokens_seen": 1454239744
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002825376128385155,
      "loss": 2.677,
      "theoretical_loss": 3.5257131629201313,
      "tokens_seen": 1454305280
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028252758274824476,
      "loss": 2.5145,
      "theoretical_loss": 3.5256990716222334,
      "tokens_seen": 1454370816
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002825175526579739,
      "loss": 2.429,
      "theoretical_loss": 3.5256849811370783,
      "tokens_seen": 1454436352
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002825075225677031,
      "loss": 2.6504,
      "theoretical_loss": 3.5256708914645825,
      "tokens_seen": 1454501888
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002824974924774323,
      "loss": 2.5089,
      "theoretical_loss": 3.5256568026046624,
      "tokens_seen": 1454567424
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002824874623871615,
      "loss": 2.8249,
      "theoretical_loss": 3.525642714557235,
      "tokens_seen": 1454632960
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028247743229689067,
      "loss": 2.4911,
      "theoretical_loss": 3.525628627322217,
      "tokens_seen": 1454698496
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028246740220661985,
      "loss": 2.5383,
      "theoretical_loss": 3.5256145408995243,
      "tokens_seen": 1454764032
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028245737211634903,
      "loss": 2.5971,
      "theoretical_loss": 3.525600455289074,
      "tokens_seen": 1454829568
    },
    {
      "epoch": 4.08,
      "objective/train/docs_used": 1594935,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5903680324554443,
      "objective/train/theoretical_loss": 3.5255863704907826,
      "objective/train/tokens_used": 1475355104,
      "theoretical_loss": 3.5255863704907826,
      "tokens_seen": 1454895104
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028244734202607826,
      "loss": 2.7843,
      "theoretical_loss": 3.5255863704907826,
      "tokens_seen": 1454895104
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002824373119358074,
      "loss": 2.6224,
      "theoretical_loss": 3.525572286504566,
      "tokens_seen": 1454960640
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002824272818455366,
      "loss": 2.5467,
      "theoretical_loss": 3.525558203330342,
      "tokens_seen": 1455026176
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002824172517552658,
      "loss": 2.4688,
      "theoretical_loss": 3.525544120968026,
      "tokens_seen": 1455091712
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.000282407221664995,
      "loss": 2.5837,
      "theoretical_loss": 3.5255300394175357,
      "tokens_seen": 1455157248
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028239719157472417,
      "loss": 2.6213,
      "theoretical_loss": 3.5255159586787865,
      "tokens_seen": 1455222784
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028238716148445335,
      "loss": 2.6744,
      "theoretical_loss": 3.5255018787516965,
      "tokens_seen": 1455288320
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028237713139418253,
      "loss": 2.8926,
      "theoretical_loss": 3.5254877996361813,
      "tokens_seen": 1455353856
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028236710130391177,
      "loss": 2.6875,
      "theoretical_loss": 3.5254737213321583,
      "tokens_seen": 1455419392
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002823570712136409,
      "loss": 2.7662,
      "theoretical_loss": 3.525459643839543,
      "tokens_seen": 1455484928
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028234704112337013,
      "loss": 2.8077,
      "theoretical_loss": 3.525445567158254,
      "tokens_seen": 1455550464
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028233701103309926,
      "loss": 2.5677,
      "theoretical_loss": 3.5254314912882063,
      "tokens_seen": 1455616000
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002823269809428285,
      "loss": 2.7994,
      "theoretical_loss": 3.5254174162293177,
      "tokens_seen": 1455681536
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002823169508525577,
      "loss": 2.6619,
      "theoretical_loss": 3.5254033419815043,
      "tokens_seen": 1455747072
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028230692076228685,
      "loss": 2.8364,
      "theoretical_loss": 3.525389268544683,
      "tokens_seen": 1455812608
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028229689067201604,
      "loss": 2.9088,
      "theoretical_loss": 3.5253751959187705,
      "tokens_seen": 1455878144
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002822868605817452,
      "loss": 2.4792,
      "theoretical_loss": 3.525361124103684,
      "tokens_seen": 1455943680
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028227683049147445,
      "loss": 2.486,
      "theoretical_loss": 3.5253470530993396,
      "tokens_seen": 1456009216
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028226680040120363,
      "loss": 2.7525,
      "theoretical_loss": 3.5253329829056543,
      "tokens_seen": 1456074752
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002822567703109328,
      "loss": 2.7045,
      "theoretical_loss": 3.525318913522546,
      "tokens_seen": 1456140288
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.000282246740220662,
      "loss": 2.6482,
      "theoretical_loss": 3.52530484494993,
      "tokens_seen": 1456205824
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028223671013039123,
      "loss": 2.884,
      "theoretical_loss": 3.525290777187724,
      "tokens_seen": 1456271360
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028222668004012036,
      "loss": 2.6727,
      "theoretical_loss": 3.525276710235844,
      "tokens_seen": 1456336896
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002822166499498496,
      "loss": 2.4325,
      "theoretical_loss": 3.5252626440942088,
      "tokens_seen": 1456402432
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002822066198595787,
      "loss": 2.681,
      "theoretical_loss": 3.525248578762733,
      "tokens_seen": 1456467968
    },
    {
      "epoch": 4.08,
      "objective/train/docs_used": 1595425,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.1949830055236816,
      "objective/train/theoretical_loss": 3.525234514241334,
      "objective/train/tokens_used": 1476993504,
      "theoretical_loss": 3.525234514241334,
      "tokens_seen": 1456533504
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028219658976930795,
      "loss": 2.3501,
      "theoretical_loss": 3.525234514241334,
      "tokens_seen": 1456533504
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028218655967903714,
      "loss": 2.496,
      "theoretical_loss": 3.52522045052993,
      "tokens_seen": 1456599040
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002821765295887663,
      "loss": 2.7187,
      "theoretical_loss": 3.5252063876284367,
      "tokens_seen": 1456664576
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002821664994984955,
      "loss": 2.5332,
      "theoretical_loss": 3.5251923255367714,
      "tokens_seen": 1456730112
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002821564694082247,
      "loss": 2.4664,
      "theoretical_loss": 3.5251782642548513,
      "tokens_seen": 1456795648
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028214643931795386,
      "loss": 2.754,
      "theoretical_loss": 3.5251642037825928,
      "tokens_seen": 1456861184
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002821364092276831,
      "loss": 2.6015,
      "theoretical_loss": 3.525150144119913,
      "tokens_seen": 1456926720
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002821263791374122,
      "loss": 2.8958,
      "theoretical_loss": 3.5251360852667286,
      "tokens_seen": 1456992256
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028211634904714146,
      "loss": 2.6523,
      "theoretical_loss": 3.5251220272229578,
      "tokens_seen": 1457057792
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002821063189568706,
      "loss": 2.6786,
      "theoretical_loss": 3.5251079699885164,
      "tokens_seen": 1457123328
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002820962888665998,
      "loss": 2.7648,
      "theoretical_loss": 3.525093913563322,
      "tokens_seen": 1457188864
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.000282086258776329,
      "loss": 2.8179,
      "theoretical_loss": 3.525079857947291,
      "tokens_seen": 1457254400
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002820762286860582,
      "loss": 2.5025,
      "theoretical_loss": 3.5250658031403406,
      "tokens_seen": 1457319936
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028206619859578736,
      "loss": 2.604,
      "theoretical_loss": 3.5250517491423885,
      "tokens_seen": 1457385472
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002820561685055166,
      "loss": 2.5271,
      "theoretical_loss": 3.5250376959533516,
      "tokens_seen": 1457451008
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002820461384152457,
      "loss": 2.5158,
      "theoretical_loss": 3.5250236435731463,
      "tokens_seen": 1457516544
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028203610832497496,
      "loss": 2.5655,
      "theoretical_loss": 3.52500959200169,
      "tokens_seen": 1457582080
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002820260782347041,
      "loss": 2.622,
      "theoretical_loss": 3.5249955412389005,
      "tokens_seen": 1457647616
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002820160481444333,
      "loss": 2.4582,
      "theoretical_loss": 3.5249814912846933,
      "tokens_seen": 1457713152
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002820060180541625,
      "loss": 2.8293,
      "theoretical_loss": 3.5249674421389874,
      "tokens_seen": 1457778688
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002819959879638917,
      "loss": 2.8033,
      "theoretical_loss": 3.524953393801699,
      "tokens_seen": 1457844224
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028198595787362087,
      "loss": 2.7273,
      "theoretical_loss": 3.5249393462727445,
      "tokens_seen": 1457909760
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028197592778335005,
      "loss": 2.7569,
      "theoretical_loss": 3.5249252995520424,
      "tokens_seen": 1457975296
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028196589769307923,
      "loss": 2.706,
      "theoretical_loss": 3.5249112536395093,
      "tokens_seen": 1458040832
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028195586760280846,
      "loss": 2.5933,
      "theoretical_loss": 3.5248972085350623,
      "tokens_seen": 1458106368
    },
    {
      "epoch": 4.08,
      "objective/train/docs_used": 1596012,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.374768018722534,
      "objective/train/theoretical_loss": 3.5248831642386182,
      "objective/train/tokens_used": 1478631904,
      "theoretical_loss": 3.5248831642386182,
      "tokens_seen": 1458171904
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002819458375125376,
      "loss": 2.6555,
      "theoretical_loss": 3.5248831642386182,
      "tokens_seen": 1458171904
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002819358074222668,
      "loss": 2.6308,
      "theoretical_loss": 3.5248691207500955,
      "tokens_seen": 1458237440
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.000281925777331996,
      "loss": 2.3286,
      "theoretical_loss": 3.52485507806941,
      "tokens_seen": 1458302976
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002819157472417252,
      "loss": 2.4707,
      "theoretical_loss": 3.52484103619648,
      "tokens_seen": 1458368512
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028190571715145437,
      "loss": 2.646,
      "theoretical_loss": 3.524826995131222,
      "tokens_seen": 1458434048
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028189568706118355,
      "loss": 2.6066,
      "theoretical_loss": 3.5248129548735534,
      "tokens_seen": 1458499584
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028188565697091273,
      "loss": 2.616,
      "theoretical_loss": 3.524798915423392,
      "tokens_seen": 1458565120
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028187562688064197,
      "loss": 2.5804,
      "theoretical_loss": 3.5247848767806538,
      "tokens_seen": 1458630656
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002818655967903711,
      "loss": 2.7479,
      "theoretical_loss": 3.524770838945258,
      "tokens_seen": 1458696192
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028185556670010033,
      "loss": 2.4494,
      "theoretical_loss": 3.5247568019171203,
      "tokens_seen": 1458761728
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028184553660982946,
      "loss": 2.7117,
      "theoretical_loss": 3.5247427656961587,
      "tokens_seen": 1458827264
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002818355065195587,
      "loss": 2.6518,
      "theoretical_loss": 3.5247287302822903,
      "tokens_seen": 1458892800
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002818254764292879,
      "loss": 2.5601,
      "theoretical_loss": 3.5247146956754323,
      "tokens_seen": 1458958336
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028181544633901705,
      "loss": 2.1958,
      "theoretical_loss": 3.524700661875503,
      "tokens_seen": 1459023872
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028180541624874624,
      "loss": 2.5015,
      "theoretical_loss": 3.524686628882418,
      "tokens_seen": 1459089408
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002817953861584754,
      "loss": 2.5364,
      "theoretical_loss": 3.524672596696097,
      "tokens_seen": 1459154944
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002817853560682046,
      "loss": 2.6715,
      "theoretical_loss": 3.524658565316455,
      "tokens_seen": 1459220480
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028177532597793383,
      "loss": 2.4276,
      "theoretical_loss": 3.524644534743411,
      "tokens_seen": 1459286016
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028176529588766296,
      "loss": 2.3955,
      "theoretical_loss": 3.524630504976882,
      "tokens_seen": 1459351552
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002817552657973922,
      "loss": 2.4853,
      "theoretical_loss": 3.524616476016785,
      "tokens_seen": 1459417088
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002817452357071214,
      "loss": 2.2738,
      "theoretical_loss": 3.524602447863038,
      "tokens_seen": 1459482624
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028173520561685056,
      "loss": 2.4406,
      "theoretical_loss": 3.524588420515559,
      "tokens_seen": 1459548160
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028172517552657974,
      "loss": 2.3551,
      "theoretical_loss": 3.5245743939742633,
      "tokens_seen": 1459613696
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002817151454363089,
      "loss": 2.4789,
      "theoretical_loss": 3.5245603682390705,
      "tokens_seen": 1459679232
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002817051153460381,
      "loss": 2.8512,
      "theoretical_loss": 3.524546343309897,
      "tokens_seen": 1459744768
    },
    {
      "epoch": 4.08,
      "objective/train/docs_used": 1597246,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.754894256591797,
      "objective/train/theoretical_loss": 3.524532319186661,
      "objective/train/tokens_used": 1480270304,
      "theoretical_loss": 3.524532319186661,
      "tokens_seen": 1459810304
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028169508525576734,
      "loss": 2.6619,
      "theoretical_loss": 3.524532319186661,
      "tokens_seen": 1459810304
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028168505516549646,
      "loss": 2.5364,
      "theoretical_loss": 3.52451829586928,
      "tokens_seen": 1459875840
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002816750250752257,
      "loss": 2.5277,
      "theoretical_loss": 3.52450427335767,
      "tokens_seen": 1459941376
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002816649949849548,
      "loss": 2.5471,
      "theoretical_loss": 3.5244902516517507,
      "tokens_seen": 1460006912
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028165496489468406,
      "loss": 2.6017,
      "theoretical_loss": 3.5244762307514383,
      "tokens_seen": 1460072448
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028164493480441324,
      "loss": 2.6548,
      "theoretical_loss": 3.524462210656651,
      "tokens_seen": 1460137984
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002816349047141424,
      "loss": 2.3979,
      "theoretical_loss": 3.5244481913673056,
      "tokens_seen": 1460203520
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002816248746238716,
      "loss": 2.5545,
      "theoretical_loss": 3.5244341728833204,
      "tokens_seen": 1460269056
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002816148445336008,
      "loss": 2.6222,
      "theoretical_loss": 3.524420155204613,
      "tokens_seen": 1460334592
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028160481444332997,
      "loss": 2.2757,
      "theoretical_loss": 3.5244061383311003,
      "tokens_seen": 1460400128
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002815947843530592,
      "loss": 2.5474,
      "theoretical_loss": 3.5243921222627006,
      "tokens_seen": 1460465664
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028158475426278833,
      "loss": 2.491,
      "theoretical_loss": 3.524378106999331,
      "tokens_seen": 1460531200
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028157472417251756,
      "loss": 2.7228,
      "theoretical_loss": 3.5243640925409103,
      "tokens_seen": 1460596736
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028156469408224674,
      "loss": 2.6113,
      "theoretical_loss": 3.5243500788873545,
      "tokens_seen": 1460662272
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002815546639919759,
      "loss": 2.8053,
      "theoretical_loss": 3.5243360660385825,
      "tokens_seen": 1460727808
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002815446339017051,
      "loss": 2.6721,
      "theoretical_loss": 3.5243220539945113,
      "tokens_seen": 1460793344
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002815346038114343,
      "loss": 2.3533,
      "theoretical_loss": 3.5243080427550586,
      "tokens_seen": 1460858880
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002815245737211635,
      "loss": 2.5673,
      "theoretical_loss": 3.524294032320143,
      "tokens_seen": 1460924416
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002815145436308927,
      "loss": 2.4373,
      "theoretical_loss": 3.5242800226896813,
      "tokens_seen": 1460989952
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002815045135406219,
      "loss": 2.5728,
      "theoretical_loss": 3.524266013863591,
      "tokens_seen": 1461055488
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028149448345035107,
      "loss": 2.708,
      "theoretical_loss": 3.5242520058417908,
      "tokens_seen": 1461121024
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028148445336008025,
      "loss": 2.3159,
      "theoretical_loss": 3.524237998624198,
      "tokens_seen": 1461186560
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028147442326980943,
      "loss": 2.5539,
      "theoretical_loss": 3.5242239922107306,
      "tokens_seen": 1461252096
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028146439317953866,
      "loss": 2.1944,
      "theoretical_loss": 3.5242099866013055,
      "tokens_seen": 1461317632
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002814543630892678,
      "loss": 2.387,
      "theoretical_loss": 3.5241959817958413,
      "tokens_seen": 1461383168
    },
    {
      "epoch": 4.08,
      "objective/train/docs_used": 1598035,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9441568851470947,
      "objective/train/theoretical_loss": 3.524181977794256,
      "objective/train/tokens_used": 1481908704,
      "theoretical_loss": 3.524181977794256,
      "tokens_seen": 1461448704
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.000281444332998997,
      "loss": 2.757,
      "theoretical_loss": 3.524181977794256,
      "tokens_seen": 1461448704
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002814343029087262,
      "loss": 2.4984,
      "theoretical_loss": 3.5241679745964665,
      "tokens_seen": 1461514240
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002814242728184554,
      "loss": 2.6137,
      "theoretical_loss": 3.524153972202391,
      "tokens_seen": 1461579776
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028141424272818457,
      "loss": 2.5324,
      "theoretical_loss": 3.524139970611948,
      "tokens_seen": 1461645312
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028140421263791375,
      "loss": 2.6002,
      "theoretical_loss": 3.5241259698250547,
      "tokens_seen": 1461710848
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028139418254764293,
      "loss": 2.7401,
      "theoretical_loss": 3.5241119698416288,
      "tokens_seen": 1461776384
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028138415245737217,
      "loss": 2.7361,
      "theoretical_loss": 3.5240979706615887,
      "tokens_seen": 1461841920
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002813741223671013,
      "loss": 2.6632,
      "theoretical_loss": 3.5240839722848523,
      "tokens_seen": 1461907456
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028136409227683053,
      "loss": 2.4013,
      "theoretical_loss": 3.5240699747113364,
      "tokens_seen": 1461972992
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028135406218655966,
      "loss": 2.5163,
      "theoretical_loss": 3.524055977940961,
      "tokens_seen": 1462038528
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002813440320962889,
      "loss": 2.5318,
      "theoretical_loss": 3.5240419819736415,
      "tokens_seen": 1462104064
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002813340020060181,
      "loss": 2.6787,
      "theoretical_loss": 3.5240279868092976,
      "tokens_seen": 1462169600
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028132397191574725,
      "loss": 2.6692,
      "theoretical_loss": 3.524013992447847,
      "tokens_seen": 1462235136
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028131394182547644,
      "loss": 2.4938,
      "theoretical_loss": 3.5239999988892077,
      "tokens_seen": 1462300672
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002813039117352056,
      "loss": 2.5738,
      "theoretical_loss": 3.5239860061332964,
      "tokens_seen": 1462366208
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002812938816449348,
      "loss": 2.5126,
      "theoretical_loss": 3.523972014180033,
      "tokens_seen": 1462431744
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028128385155466403,
      "loss": 2.4717,
      "theoretical_loss": 3.5239580230293335,
      "tokens_seen": 1462497280
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028127382146439316,
      "loss": 2.6733,
      "theoretical_loss": 3.5239440326811176,
      "tokens_seen": 1462562816
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002812637913741224,
      "loss": 2.6682,
      "theoretical_loss": 3.5239300431353024,
      "tokens_seen": 1462628352
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002812537612838516,
      "loss": 2.5902,
      "theoretical_loss": 3.5239160543918064,
      "tokens_seen": 1462693888
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028124373119358076,
      "loss": 2.415,
      "theoretical_loss": 3.523902066450548,
      "tokens_seen": 1462759424
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028123370110330994,
      "loss": 2.5721,
      "theoretical_loss": 3.523888079311444,
      "tokens_seen": 1462824960
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002812236710130391,
      "loss": 2.4526,
      "theoretical_loss": 3.5238740929744132,
      "tokens_seen": 1462890496
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002812136409227683,
      "loss": 2.51,
      "theoretical_loss": 3.523860107439374,
      "tokens_seen": 1462956032
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028120361083249754,
      "loss": 2.5591,
      "theoretical_loss": 3.523846122706243,
      "tokens_seen": 1463021568
    },
    {
      "epoch": 4.08,
      "objective/train/docs_used": 1603030,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7301270961761475,
      "objective/train/theoretical_loss": 3.5238321387749405,
      "objective/train/tokens_used": 1483547104,
      "theoretical_loss": 3.5238321387749405,
      "tokens_seen": 1463087104
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028119358074222666,
      "loss": 2.5086,
      "theoretical_loss": 3.5238321387749405,
      "tokens_seen": 1463087104
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002811835506519559,
      "loss": 2.6898,
      "theoretical_loss": 3.523818155645383,
      "tokens_seen": 1463152640
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.000281173520561685,
      "loss": 2.3801,
      "theoretical_loss": 3.5238041733174894,
      "tokens_seen": 1463218176
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028116349047141426,
      "loss": 2.4418,
      "theoretical_loss": 3.523790191791178,
      "tokens_seen": 1463283712
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028115346038114344,
      "loss": 2.4984,
      "theoretical_loss": 3.523776211066366,
      "tokens_seen": 1463349248
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002811434302908726,
      "loss": 2.51,
      "theoretical_loss": 3.5237622311429724,
      "tokens_seen": 1463414784
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002811334002006018,
      "loss": 2.6372,
      "theoretical_loss": 3.5237482520209147,
      "tokens_seen": 1463480320
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.000281123370110331,
      "loss": 2.7331,
      "theoretical_loss": 3.523734273700112,
      "tokens_seen": 1463545856
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028111334002006017,
      "loss": 2.3877,
      "theoretical_loss": 3.5237202961804814,
      "tokens_seen": 1463611392
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002811033099297894,
      "loss": 2.3393,
      "theoretical_loss": 3.523706319461942,
      "tokens_seen": 1463676928
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028109327983951853,
      "loss": 2.6241,
      "theoretical_loss": 3.523692343544412,
      "tokens_seen": 1463742464
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028108324974924776,
      "loss": 2.5906,
      "theoretical_loss": 3.5236783684278086,
      "tokens_seen": 1463808000
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028107321965897695,
      "loss": 2.1827,
      "theoretical_loss": 3.5236643941120516,
      "tokens_seen": 1463873536
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002810631895687061,
      "loss": 2.8091,
      "theoretical_loss": 3.5236504205970576,
      "tokens_seen": 1463939072
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002810531594784353,
      "loss": 2.621,
      "theoretical_loss": 3.523636447882746,
      "tokens_seen": 1464004608
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002810431293881645,
      "loss": 2.3639,
      "theoretical_loss": 3.5236224759690353,
      "tokens_seen": 1464070144
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028103309929789367,
      "loss": 2.708,
      "theoretical_loss": 3.5236085048558428,
      "tokens_seen": 1464135680
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002810230692076229,
      "loss": 2.3876,
      "theoretical_loss": 3.5235945345430872,
      "tokens_seen": 1464201216
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028101303911735203,
      "loss": 2.7517,
      "theoretical_loss": 3.523580565030687,
      "tokens_seen": 1464266752
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028100300902708127,
      "loss": 2.5516,
      "theoretical_loss": 3.5235665963185605,
      "tokens_seen": 1464332288
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002809929789368104,
      "loss": 2.5998,
      "theoretical_loss": 3.523552628406626,
      "tokens_seen": 1464397824
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028098294884653963,
      "loss": 2.6177,
      "theoretical_loss": 3.5235386612948014,
      "tokens_seen": 1464463360
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002809729187562688,
      "loss": 2.5494,
      "theoretical_loss": 3.5235246949830064,
      "tokens_seen": 1464528896
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.000280962888665998,
      "loss": 2.74,
      "theoretical_loss": 3.5235107294711576,
      "tokens_seen": 1464594432
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028095285857572717,
      "loss": 2.7292,
      "theoretical_loss": 3.523496764759175,
      "tokens_seen": 1464659968
    },
    {
      "epoch": 4.08,
      "objective/train/docs_used": 1603030,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.211653470993042,
      "objective/train/theoretical_loss": 3.523482800846976,
      "objective/train/tokens_used": 1483791840,
      "theoretical_loss": 3.523482800846976,
      "tokens_seen": 1464725504
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028094282848545635,
      "loss": 2.2415,
      "theoretical_loss": 3.523482800846976,
      "tokens_seen": 1464725504
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028093279839518553,
      "loss": 2.4344,
      "theoretical_loss": 3.5234688377344785,
      "tokens_seen": 1464791040
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028092276830491477,
      "loss": 2.6155,
      "theoretical_loss": 3.5234548754216024,
      "tokens_seen": 1464856576
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002809127382146439,
      "loss": 2.6634,
      "theoretical_loss": 3.5234409139082654,
      "tokens_seen": 1464922112
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028090270812437313,
      "loss": 2.5183,
      "theoretical_loss": 3.5234269531943863,
      "tokens_seen": 1464987648
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002808926780341023,
      "loss": 2.6951,
      "theoretical_loss": 3.5234129932798828,
      "tokens_seen": 1465053184
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002808826479438315,
      "loss": 2.7359,
      "theoretical_loss": 3.5233990341646737,
      "tokens_seen": 1465118720
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002808726178535607,
      "loss": 2.5248,
      "theoretical_loss": 3.5233850758486778,
      "tokens_seen": 1465184256
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028086258776328986,
      "loss": 2.6991,
      "theoretical_loss": 3.5233711183318137,
      "tokens_seen": 1465249792
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028085255767301904,
      "loss": 2.4007,
      "theoretical_loss": 3.523357161613999,
      "tokens_seen": 1465315328
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002808425275827483,
      "loss": 2.5959,
      "theoretical_loss": 3.523343205695153,
      "tokens_seen": 1465380864
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002808324974924774,
      "loss": 2.7473,
      "theoretical_loss": 3.5233292505751943,
      "tokens_seen": 1465446400
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028082246740220664,
      "loss": 2.5896,
      "theoretical_loss": 3.523315296254041,
      "tokens_seen": 1465511936
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028081243731193576,
      "loss": 2.8018,
      "theoretical_loss": 3.523301342731612,
      "tokens_seen": 1465577472
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.000280802407221665,
      "loss": 2.252,
      "theoretical_loss": 3.523287390007826,
      "tokens_seen": 1465643008
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002807923771313942,
      "loss": 2.6345,
      "theoretical_loss": 3.523273438082601,
      "tokens_seen": 1465708544
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028078234704112336,
      "loss": 2.6385,
      "theoretical_loss": 3.5232594869558556,
      "tokens_seen": 1465774080
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002807723169508526,
      "loss": 2.603,
      "theoretical_loss": 3.5232455366275093,
      "tokens_seen": 1465839616
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002807622868605818,
      "loss": 2.5463,
      "theoretical_loss": 3.52323158709748,
      "tokens_seen": 1465905152
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028075225677031096,
      "loss": 2.6577,
      "theoretical_loss": 3.5232176383656864,
      "tokens_seen": 1465970688
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028074222668004014,
      "loss": 2.6319,
      "theoretical_loss": 3.523203690432047,
      "tokens_seen": 1466036224
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002807321965897693,
      "loss": 2.1931,
      "theoretical_loss": 3.5231897432964807,
      "tokens_seen": 1466101760
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002807221664994985,
      "loss": 2.6428,
      "theoretical_loss": 3.523175796958906,
      "tokens_seen": 1466167296
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028071213640922774,
      "loss": 2.5598,
      "theoretical_loss": 3.5231618514192427,
      "tokens_seen": 1466232832
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028070210631895686,
      "loss": 2.7383,
      "theoretical_loss": 3.5231479066774076,
      "tokens_seen": 1466298368
    },
    {
      "epoch": 4.08,
      "objective/train/docs_used": 1603030,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4567415714263916,
      "objective/train/theoretical_loss": 3.5231339627333202,
      "objective/train/tokens_used": 1483791840,
      "theoretical_loss": 3.5231339627333202,
      "tokens_seen": 1466363904
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002806920762286861,
      "loss": 2.4262,
      "theoretical_loss": 3.5231339627333202,
      "tokens_seen": 1466363904
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002806820461384152,
      "loss": 2.5823,
      "theoretical_loss": 3.5231200195868997,
      "tokens_seen": 1466429440
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028067201604814446,
      "loss": 2.5947,
      "theoretical_loss": 3.523106077238064,
      "tokens_seen": 1466494976
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028066198595787364,
      "loss": 2.4495,
      "theoretical_loss": 3.523092135686732,
      "tokens_seen": 1466560512
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002806519558676028,
      "loss": 2.3593,
      "theoretical_loss": 3.5230781949328236,
      "tokens_seen": 1466626048
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.000280641925777332,
      "loss": 2.6452,
      "theoretical_loss": 3.523064254976256,
      "tokens_seen": 1466691584
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002806318956870612,
      "loss": 2.5566,
      "theoretical_loss": 3.523050315816949,
      "tokens_seen": 1466757120
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028062186559679037,
      "loss": 2.4694,
      "theoretical_loss": 3.5230363774548206,
      "tokens_seen": 1466822656
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002806118355065196,
      "loss": 2.4054,
      "theoretical_loss": 3.5230224398897905,
      "tokens_seen": 1466888192
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028060180541624873,
      "loss": 2.8066,
      "theoretical_loss": 3.5230085031217766,
      "tokens_seen": 1466953728
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028059177532597796,
      "loss": 2.6842,
      "theoretical_loss": 3.5229945671506986,
      "tokens_seen": 1467019264
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028058174523570715,
      "loss": 2.6991,
      "theoretical_loss": 3.5229806319764743,
      "tokens_seen": 1467084800
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002805717151454363,
      "loss": 2.5365,
      "theoretical_loss": 3.5229666975990233,
      "tokens_seen": 1467150336
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002805616850551655,
      "loss": 2.4685,
      "theoretical_loss": 3.5229527640182647,
      "tokens_seen": 1467215872
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002805516549648947,
      "loss": 2.8453,
      "theoretical_loss": 3.522938831234116,
      "tokens_seen": 1467281408
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028054162487462387,
      "loss": 2.7806,
      "theoretical_loss": 3.522924899246498,
      "tokens_seen": 1467346944
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.0002805315947843531,
      "loss": 2.5334,
      "theoretical_loss": 3.5229109680553283,
      "tokens_seen": 1467412480
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028052156469408223,
      "loss": 2.609,
      "theoretical_loss": 3.5228970376605258,
      "tokens_seen": 1467478016
    },
    {
      "epoch": 4.08,
      "learning_rate": 0.00028051153460381147,
      "loss": 2.3968,
      "theoretical_loss": 3.52288310806201,
      "tokens_seen": 1467543552
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.0002805015045135406,
      "loss": 2.8006,
      "theoretical_loss": 3.5228691792596996,
      "tokens_seen": 1467609088
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.00028049147442326983,
      "loss": 2.5682,
      "theoretical_loss": 3.5228552512535134,
      "tokens_seen": 1467674624
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.000280481444332999,
      "loss": 2.7127,
      "theoretical_loss": 3.5228413240433705,
      "tokens_seen": 1467740160
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.0002804714142427282,
      "loss": 2.2954,
      "theoretical_loss": 3.5228273976291895,
      "tokens_seen": 1467805696
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.00028046138415245737,
      "loss": 2.8381,
      "theoretical_loss": 3.52281347201089,
      "tokens_seen": 1467871232
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.00028045135406218655,
      "loss": 2.4445,
      "theoretical_loss": 3.5227995471883906,
      "tokens_seen": 1467936768
    },
    {
      "epoch": 4.09,
      "objective/train/docs_used": 1603030,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5689315795898438,
      "objective/train/theoretical_loss": 3.52278562316161,
      "objective/train/tokens_used": 1483791840,
      "theoretical_loss": 3.52278562316161,
      "tokens_seen": 1468002304
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.00028044132397191574,
      "loss": 2.6411,
      "theoretical_loss": 3.52278562316161,
      "tokens_seen": 1468002304
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.00028043129388164497,
      "loss": 2.5643,
      "theoretical_loss": 3.522771699930468,
      "tokens_seen": 1468067840
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.0002804212637913741,
      "loss": 2.7135,
      "theoretical_loss": 3.522757777494883,
      "tokens_seen": 1468133376
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.00028041123370110333,
      "loss": 2.7565,
      "theoretical_loss": 3.522743855854774,
      "tokens_seen": 1468198912
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.0002804012036108325,
      "loss": 2.6198,
      "theoretical_loss": 3.5227299350100605,
      "tokens_seen": 1468264448
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.0002803911735205617,
      "loss": 2.6975,
      "theoretical_loss": 3.522716014960661,
      "tokens_seen": 1468329984
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.0002803811434302909,
      "loss": 2.4973,
      "theoretical_loss": 3.522702095706496,
      "tokens_seen": 1468395520
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.00028037111334002006,
      "loss": 2.5945,
      "theoretical_loss": 3.522688177247482,
      "tokens_seen": 1468461056
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.00028036108324974924,
      "loss": 2.6434,
      "theoretical_loss": 3.5226742595835407,
      "tokens_seen": 1468526592
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.0002803510531594785,
      "loss": 2.4662,
      "theoretical_loss": 3.5226603427145893,
      "tokens_seen": 1468592128
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.0002803410230692076,
      "loss": 2.6721,
      "theoretical_loss": 3.5226464266405477,
      "tokens_seen": 1468657664
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.00028033099297893684,
      "loss": 2.4547,
      "theoretical_loss": 3.5226325113613353,
      "tokens_seen": 1468723200
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.00028032096288866596,
      "loss": 2.518,
      "theoretical_loss": 3.522618596876871,
      "tokens_seen": 1468788736
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.0002803109327983952,
      "loss": 2.4428,
      "theoretical_loss": 3.522604683187074,
      "tokens_seen": 1468854272
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.0002803009027081244,
      "loss": 2.8113,
      "theoretical_loss": 3.5225907702918633,
      "tokens_seen": 1468919808
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.00028029087261785356,
      "loss": 2.336,
      "theoretical_loss": 3.5225768581911576,
      "tokens_seen": 1468985344
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.00028028084252758274,
      "loss": 2.7046,
      "theoretical_loss": 3.522562946884877,
      "tokens_seen": 1469050880
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.000280270812437312,
      "loss": 2.4133,
      "theoretical_loss": 3.5225490363729404,
      "tokens_seen": 1469116416
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.0002802607823470411,
      "loss": 2.4705,
      "theoretical_loss": 3.522535126655267,
      "tokens_seen": 1469181952
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.00028025075225677034,
      "loss": 2.4143,
      "theoretical_loss": 3.522521217731776,
      "tokens_seen": 1469247488
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.00028024072216649947,
      "loss": 2.7378,
      "theoretical_loss": 3.5225073096023864,
      "tokens_seen": 1469313024
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.0002802306920762287,
      "loss": 2.5421,
      "theoretical_loss": 3.5224934022670173,
      "tokens_seen": 1469378560
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.0002802206619859579,
      "loss": 2.4663,
      "theoretical_loss": 3.5224794957255887,
      "tokens_seen": 1469444096
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.00028021063189568706,
      "loss": 2.6527,
      "theoretical_loss": 3.5224655899780197,
      "tokens_seen": 1469509632
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.00028020060180541624,
      "loss": 2.5792,
      "theoretical_loss": 3.5224516850242287,
      "tokens_seen": 1469575168
    },
    {
      "epoch": 4.09,
      "objective/train/docs_used": 1603030,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.003769636154175,
      "objective/train/theoretical_loss": 3.522437780864136,
      "objective/train/tokens_used": 1483791840,
      "theoretical_loss": 3.522437780864136,
      "tokens_seen": 1469640704
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.0002801905717151454,
      "loss": 2.6305,
      "theoretical_loss": 3.522437780864136,
      "tokens_seen": 1469640704
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.0002801805416248746,
      "loss": 2.6325,
      "theoretical_loss": 3.5224238774976606,
      "tokens_seen": 1469706240
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.00028017051153460384,
      "loss": 2.5469,
      "theoretical_loss": 3.5224099749247215,
      "tokens_seen": 1469771776
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.00028016048144433297,
      "loss": 2.7021,
      "theoretical_loss": 3.5223960731452384,
      "tokens_seen": 1469837312
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.0002801504513540622,
      "loss": 2.4954,
      "theoretical_loss": 3.522382172159131,
      "tokens_seen": 1469902848
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.00028014042126379133,
      "loss": 2.6064,
      "theoretical_loss": 3.522368271966317,
      "tokens_seen": 1469968384
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.00028013039117352057,
      "loss": 2.6808,
      "theoretical_loss": 3.522354372566718,
      "tokens_seen": 1470033920
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.00028012036108324975,
      "loss": 2.8744,
      "theoretical_loss": 3.522340473960252,
      "tokens_seen": 1470099456
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.00028011033099297893,
      "loss": 2.5823,
      "theoretical_loss": 3.5223265761468388,
      "tokens_seen": 1470164992
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.0002801003009027081,
      "loss": 2.361,
      "theoretical_loss": 3.5223126791263977,
      "tokens_seen": 1470230528
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.00028009027081243735,
      "loss": 2.6317,
      "theoretical_loss": 3.522298782898848,
      "tokens_seen": 1470296064
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.00028008024072216647,
      "loss": 2.5516,
      "theoretical_loss": 3.5222848874641093,
      "tokens_seen": 1470361600
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.0002800702106318957,
      "loss": 2.6678,
      "theoretical_loss": 3.5222709928221008,
      "tokens_seen": 1470427136
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.00028006018054162483,
      "loss": 2.4722,
      "theoretical_loss": 3.522257098972742,
      "tokens_seen": 1470492672
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.00028005015045135407,
      "loss": 2.779,
      "theoretical_loss": 3.522243205915953,
      "tokens_seen": 1470558208
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.00028004012036108325,
      "loss": 2.6158,
      "theoretical_loss": 3.5222293136516525,
      "tokens_seen": 1470623744
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.00028003009027081243,
      "loss": 2.41,
      "theoretical_loss": 3.52221542217976,
      "tokens_seen": 1470689280
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.00028002006018054167,
      "loss": 2.4115,
      "theoretical_loss": 3.5222015315001958,
      "tokens_seen": 1470754816
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.0002800100300902708,
      "loss": 2.6044,
      "theoretical_loss": 3.522187641612878,
      "tokens_seen": 1470820352
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.00028000000000000003,
      "loss": 2.3933,
      "theoretical_loss": 3.5221737525177277,
      "tokens_seen": 1470885888
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.0002799899699097292,
      "loss": 2.5452,
      "theoretical_loss": 3.5221598642146636,
      "tokens_seen": 1470951424
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.0002799799398194584,
      "loss": 2.4104,
      "theoretical_loss": 3.522145976703605,
      "tokens_seen": 1471016960
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.0002799699097291876,
      "loss": 2.4964,
      "theoretical_loss": 3.522132089984472,
      "tokens_seen": 1471082496
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.00027995987963891675,
      "loss": 2.5139,
      "theoretical_loss": 3.522118204057184,
      "tokens_seen": 1471148032
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.00027994984954864594,
      "loss": 2.5464,
      "theoretical_loss": 3.5221043189216603,
      "tokens_seen": 1471213568
    },
    {
      "epoch": 4.09,
      "objective/train/docs_used": 1603030,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.143112897872925,
      "objective/train/theoretical_loss": 3.5220904345778203,
      "objective/train/tokens_used": 1483791840,
      "theoretical_loss": 3.5220904345778203,
      "tokens_seen": 1471279104
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.00027993981945837517,
      "loss": 2.7808,
      "theoretical_loss": 3.5220904345778203,
      "tokens_seen": 1471279104
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.0002799297893681043,
      "loss": 2.6289,
      "theoretical_loss": 3.5220765510255845,
      "tokens_seen": 1471344640
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.00027991975927783353,
      "loss": 2.6422,
      "theoretical_loss": 3.522062668264872,
      "tokens_seen": 1471410176
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.0002799097291875627,
      "loss": 2.5013,
      "theoretical_loss": 3.5220487862956027,
      "tokens_seen": 1471475712
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.0002798996990972919,
      "loss": 2.5409,
      "theoretical_loss": 3.522034905117695,
      "tokens_seen": 1471541248
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.0002798896690070211,
      "loss": 2.5736,
      "theoretical_loss": 3.52202102473107,
      "tokens_seen": 1471606784
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.00027987963891675026,
      "loss": 2.7528,
      "theoretical_loss": 3.5220071451356474,
      "tokens_seen": 1471672320
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.00027986960882647944,
      "loss": 2.2941,
      "theoretical_loss": 3.5219932663313456,
      "tokens_seen": 1471737856
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.0002798595787362087,
      "loss": 2.6457,
      "theoretical_loss": 3.5219793883180857,
      "tokens_seen": 1471803392
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.0002798495486459378,
      "loss": 2.5519,
      "theoretical_loss": 3.5219655110957864,
      "tokens_seen": 1471868928
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.00027983951855566704,
      "loss": 2.6568,
      "theoretical_loss": 3.521951634664368,
      "tokens_seen": 1471934464
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.00027982948846539616,
      "loss": 2.4972,
      "theoretical_loss": 3.521937759023749,
      "tokens_seen": 1472000000
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.0002798194583751254,
      "loss": 2.4951,
      "theoretical_loss": 3.5219238841738507,
      "tokens_seen": 1472065536
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.0002798094282848546,
      "loss": 2.6323,
      "theoretical_loss": 3.521910010114593,
      "tokens_seen": 1472131072
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.00027979939819458376,
      "loss": 2.7124,
      "theoretical_loss": 3.5218961368458936,
      "tokens_seen": 1472196608
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.00027978936810431294,
      "loss": 2.4036,
      "theoretical_loss": 3.521882264367674,
      "tokens_seen": 1472262144
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.0002797793380140422,
      "loss": 2.3475,
      "theoretical_loss": 3.5218683926798535,
      "tokens_seen": 1472327680
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.0002797693079237713,
      "loss": 2.6737,
      "theoretical_loss": 3.521854521782352,
      "tokens_seen": 1472393216
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.00027975927783350054,
      "loss": 2.61,
      "theoretical_loss": 3.5218406516750886,
      "tokens_seen": 1472458752
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.00027974924774322967,
      "loss": 2.4702,
      "theoretical_loss": 3.5218267823579845,
      "tokens_seen": 1472524288
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.0002797392176529589,
      "loss": 2.5781,
      "theoretical_loss": 3.521812913830958,
      "tokens_seen": 1472589824
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.0002797291875626881,
      "loss": 2.704,
      "theoretical_loss": 3.52179904609393,
      "tokens_seen": 1472655360
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.00027971915747241726,
      "loss": 2.5015,
      "theoretical_loss": 3.52178517914682,
      "tokens_seen": 1472720896
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.00027970912738214644,
      "loss": 2.4233,
      "theoretical_loss": 3.521771312989548,
      "tokens_seen": 1472786432
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.0002796990972918756,
      "loss": 2.7242,
      "theoretical_loss": 3.5217574476220337,
      "tokens_seen": 1472851968
    },
    {
      "epoch": 4.09,
      "objective/train/docs_used": 1603030,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5638537406921387,
      "objective/train/theoretical_loss": 3.5217435830441968,
      "objective/train/tokens_used": 1483791840,
      "theoretical_loss": 3.5217435830441968,
      "tokens_seen": 1472917504
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.0002796890672016048,
      "loss": 2.5489,
      "theoretical_loss": 3.5217435830441968,
      "tokens_seen": 1472917504
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.00027967903711133404,
      "loss": 2.5476,
      "theoretical_loss": 3.521729719255957,
      "tokens_seen": 1472983040
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.00027966900702106317,
      "loss": 2.3642,
      "theoretical_loss": 3.5217158562572353,
      "tokens_seen": 1473048576
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.0002796589769307924,
      "loss": 2.4547,
      "theoretical_loss": 3.521701994047951,
      "tokens_seen": 1473114112
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.00027964894684052153,
      "loss": 2.5487,
      "theoretical_loss": 3.5216881326280234,
      "tokens_seen": 1473179648
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.00027963891675025077,
      "loss": 2.819,
      "theoretical_loss": 3.521674271997373,
      "tokens_seen": 1473245184
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.00027962888665997995,
      "loss": 2.6273,
      "theoretical_loss": 3.52166041215592,
      "tokens_seen": 1473310720
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.00027961885656970913,
      "loss": 2.664,
      "theoretical_loss": 3.521646553103584,
      "tokens_seen": 1473376256
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.0002796088264794383,
      "loss": 2.4287,
      "theoretical_loss": 3.521632694840285,
      "tokens_seen": 1473441792
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.00027959879638916755,
      "loss": 2.3851,
      "theoretical_loss": 3.521618837365943,
      "tokens_seen": 1473507328
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.00027958876629889667,
      "loss": 2.4027,
      "theoretical_loss": 3.521604980680478,
      "tokens_seen": 1473572864
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.0002795787362086259,
      "loss": 2.6463,
      "theoretical_loss": 3.5215911247838103,
      "tokens_seen": 1473638400
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.00027956870611835503,
      "loss": 2.3324,
      "theoretical_loss": 3.5215772696758596,
      "tokens_seen": 1473703936
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.00027955867602808427,
      "loss": 2.2111,
      "theoretical_loss": 3.521563415356546,
      "tokens_seen": 1473769472
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.00027954864593781345,
      "loss": 2.3678,
      "theoretical_loss": 3.5215495618257897,
      "tokens_seen": 1473835008
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.00027953861584754263,
      "loss": 2.5052,
      "theoretical_loss": 3.52153570908351,
      "tokens_seen": 1473900544
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.0002795285857572718,
      "loss": 2.6186,
      "theoretical_loss": 3.5215218571296285,
      "tokens_seen": 1473966080
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.000279518555667001,
      "loss": 2.4767,
      "theoretical_loss": 3.5215080059640638,
      "tokens_seen": 1474031616
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.0002795085255767302,
      "loss": 2.3363,
      "theoretical_loss": 3.5214941555867365,
      "tokens_seen": 1474097152
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.0002794984954864594,
      "loss": 2.7143,
      "theoretical_loss": 3.5214803059975672,
      "tokens_seen": 1474162688
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.00027948846539618854,
      "loss": 2.4726,
      "theoretical_loss": 3.521466457196475,
      "tokens_seen": 1474228224
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.0002794784353059178,
      "loss": 2.7101,
      "theoretical_loss": 3.521452609183381,
      "tokens_seen": 1474293760
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.0002794684052156469,
      "loss": 2.31,
      "theoretical_loss": 3.521438761958205,
      "tokens_seen": 1474359296
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.00027945837512537614,
      "loss": 2.6026,
      "theoretical_loss": 3.521424915520867,
      "tokens_seen": 1474424832
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.0002794483450351053,
      "loss": 2.5403,
      "theoretical_loss": 3.521411069871287,
      "tokens_seen": 1474490368
    },
    {
      "debugging/Self-BLEU-5": 0.40022323577174246,
      "debugging/distinct-1-grams": 0.792449235368197,
      "debugging/distinct-2-grams": 0.964039917418843,
      "debugging/entropy-1-grams": 5.830051148996791,
      "debugging/entropy-2-grams": 6.574183943619101,
      "debugging/length": 506.7,
      "debugging/num_segments": 10,
      "debugging/score": 0.007128608811555618,
      "debugging/score_std": 0.008951280671415335,
      "epoch": 4.09,
      "objective/train/docs_used": 1603030,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.455482244491577,
      "objective/train/theoretical_loss": 3.5213972250093857,
      "objective/train/tokens_used": 1483791840,
      "theoretical_loss": 3.5213972250093857,
      "tokens_seen": 1474555904
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.0002794383149448345,
      "loss": 2.5027,
      "theoretical_loss": 3.5213972250093857,
      "tokens_seen": 1474555904
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.0002794282848545637,
      "loss": 2.409,
      "theoretical_loss": 3.5213833809350827,
      "tokens_seen": 1474621440
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.0002794182547642929,
      "loss": 2.7487,
      "theoretical_loss": 3.5213695376482983,
      "tokens_seen": 1474686976
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.00027940822467402204,
      "loss": 2.3847,
      "theoretical_loss": 3.5213556951489533,
      "tokens_seen": 1474752512
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.0002793981945837513,
      "loss": 2.4294,
      "theoretical_loss": 3.521341853436968,
      "tokens_seen": 1474818048
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.0002793881644934804,
      "loss": 2.7022,
      "theoretical_loss": 3.521328012512261,
      "tokens_seen": 1474883584
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.00027937813440320964,
      "loss": 2.4663,
      "theoretical_loss": 3.5213141723747543,
      "tokens_seen": 1474949120
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.0002793681043129388,
      "loss": 2.4916,
      "theoretical_loss": 3.5213003330243673,
      "tokens_seen": 1475014656
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.000279358074222668,
      "loss": 2.5263,
      "theoretical_loss": 3.521286494461021,
      "tokens_seen": 1475080192
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.0002793480441323972,
      "loss": 2.5372,
      "theoretical_loss": 3.521272656684635,
      "tokens_seen": 1475145728
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.00027933801404212636,
      "loss": 2.4316,
      "theoretical_loss": 3.5212588196951295,
      "tokens_seen": 1475211264
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.00027932798395185554,
      "loss": 2.3126,
      "theoretical_loss": 3.5212449834924255,
      "tokens_seen": 1475276800
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.0002793179538615848,
      "loss": 2.3813,
      "theoretical_loss": 3.5212311480764424,
      "tokens_seen": 1475342336
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.0002793079237713139,
      "loss": 2.8092,
      "theoretical_loss": 3.521217313447101,
      "tokens_seen": 1475407872
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.00027929789368104314,
      "loss": 2.5193,
      "theoretical_loss": 3.5212034796043223,
      "tokens_seen": 1475473408
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.00027928786359077227,
      "loss": 2.5998,
      "theoretical_loss": 3.5211896465480255,
      "tokens_seen": 1475538944
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.0002792778335005015,
      "loss": 2.4218,
      "theoretical_loss": 3.5211758142781315,
      "tokens_seen": 1475604480
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.00027926780341023074,
      "loss": 2.5907,
      "theoretical_loss": 3.5211619827945606,
      "tokens_seen": 1475670016
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.00027925777331995987,
      "loss": 2.3143,
      "theoretical_loss": 3.5211481520972336,
      "tokens_seen": 1475735552
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.0002792477432296891,
      "loss": 2.8273,
      "theoretical_loss": 3.52113432218607,
      "tokens_seen": 1475801088
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.0002792377131394183,
      "loss": 2.5817,
      "theoretical_loss": 3.521120493060991,
      "tokens_seen": 1475866624
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.00027922768304914746,
      "loss": 2.3576,
      "theoretical_loss": 3.5211066647219162,
      "tokens_seen": 1475932160
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.00027921765295887664,
      "loss": 2.3519,
      "theoretical_loss": 3.521092837168767,
      "tokens_seen": 1475997696
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.0002792076228686058,
      "loss": 2.62,
      "theoretical_loss": 3.521079010401463,
      "tokens_seen": 1476063232
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.000279197592778335,
      "loss": 2.4679,
      "theoretical_loss": 3.5210651844199257,
      "tokens_seen": 1476128768
    },
    {
      "epoch": 4.09,
      "objective/train/docs_used": 1603030,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.234206199645996,
      "objective/train/theoretical_loss": 3.5210513592240744,
      "objective/train/tokens_used": 1483791840,
      "theoretical_loss": 3.5210513592240744,
      "tokens_seen": 1476194304
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.00027918756268806424,
      "loss": 2.6296,
      "theoretical_loss": 3.5210513592240744,
      "tokens_seen": 1476194304
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.00027917753259779337,
      "loss": 2.5854,
      "theoretical_loss": 3.5210375348138303,
      "tokens_seen": 1476259840
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.0002791675025075226,
      "loss": 2.4903,
      "theoretical_loss": 3.5210237111891134,
      "tokens_seen": 1476325376
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.00027915747241725173,
      "loss": 2.7127,
      "theoretical_loss": 3.5210098883498446,
      "tokens_seen": 1476390912
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.00027914744232698097,
      "loss": 2.8816,
      "theoretical_loss": 3.520996066295944,
      "tokens_seen": 1476456448
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.00027913741223671015,
      "loss": 2.6881,
      "theoretical_loss": 3.520982245027333,
      "tokens_seen": 1476521984
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.00027912738214643933,
      "loss": 2.446,
      "theoretical_loss": 3.5209684245439314,
      "tokens_seen": 1476587520
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.0002791173520561685,
      "loss": 2.7502,
      "theoretical_loss": 3.5209546048456595,
      "tokens_seen": 1476653056
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.00027910732196589775,
      "loss": 2.6565,
      "theoretical_loss": 3.520940785932438,
      "tokens_seen": 1476718592
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.00027909729187562687,
      "loss": 2.3348,
      "theoretical_loss": 3.520926967804188,
      "tokens_seen": 1476784128
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.0002790872617853561,
      "loss": 2.6025,
      "theoretical_loss": 3.52091315046083,
      "tokens_seen": 1476849664
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.00027907723169508523,
      "loss": 2.6091,
      "theoretical_loss": 3.5208993339022845,
      "tokens_seen": 1476915200
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.00027906720160481447,
      "loss": 2.4213,
      "theoretical_loss": 3.5208855181284715,
      "tokens_seen": 1476980736
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.00027905717151454365,
      "loss": 2.6044,
      "theoretical_loss": 3.5208717031393117,
      "tokens_seen": 1477046272
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.00027904714142427283,
      "loss": 2.4268,
      "theoretical_loss": 3.520857888934727,
      "tokens_seen": 1477111808
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.000279037111334002,
      "loss": 2.4377,
      "theoretical_loss": 3.5208440755146366,
      "tokens_seen": 1477177344
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.0002790270812437312,
      "loss": 2.4946,
      "theoretical_loss": 3.5208302628789614,
      "tokens_seen": 1477242880
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.0002790170511534604,
      "loss": 2.3929,
      "theoretical_loss": 3.520816451027623,
      "tokens_seen": 1477308416
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.0002790070210631896,
      "loss": 2.605,
      "theoretical_loss": 3.520802639960541,
      "tokens_seen": 1477373952
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.00027899699097291874,
      "loss": 2.2421,
      "theoretical_loss": 3.5207888296776364,
      "tokens_seen": 1477439488
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.000278986960882648,
      "loss": 2.5311,
      "theoretical_loss": 3.52077502017883,
      "tokens_seen": 1477505024
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.0002789769307923771,
      "loss": 2.6109,
      "theoretical_loss": 3.520761211464043,
      "tokens_seen": 1477570560
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.00027896690070210634,
      "loss": 2.516,
      "theoretical_loss": 3.520747403533195,
      "tokens_seen": 1477636096
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.0002789568706118355,
      "loss": 2.4437,
      "theoretical_loss": 3.5207335963862074,
      "tokens_seen": 1477701632
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.0002789468405215647,
      "loss": 2.8298,
      "theoretical_loss": 3.5207197900230014,
      "tokens_seen": 1477767168
    },
    {
      "epoch": 4.09,
      "objective/train/docs_used": 1603030,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.0272233486175537,
      "objective/train/theoretical_loss": 3.5207059844434965,
      "objective/train/tokens_used": 1483791840,
      "theoretical_loss": 3.5207059844434965,
      "tokens_seen": 1477832704
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.0002789368104312939,
      "loss": 2.803,
      "theoretical_loss": 3.5207059844434965,
      "tokens_seen": 1477832704
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.0002789267803410231,
      "loss": 2.6148,
      "theoretical_loss": 3.5206921796476145,
      "tokens_seen": 1477898240
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.00027891675025075224,
      "loss": 2.6943,
      "theoretical_loss": 3.520678375635276,
      "tokens_seen": 1477963776
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.0002789067201604815,
      "loss": 2.4597,
      "theoretical_loss": 3.5206645724064014,
      "tokens_seen": 1478029312
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.0002788966900702106,
      "loss": 2.5187,
      "theoretical_loss": 3.5206507699609118,
      "tokens_seen": 1478094848
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.00027888665997993984,
      "loss": 2.762,
      "theoretical_loss": 3.520636968298728,
      "tokens_seen": 1478160384
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.000278876629889669,
      "loss": 2.8494,
      "theoretical_loss": 3.5206231674197705,
      "tokens_seen": 1478225920
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.0002788665997993982,
      "loss": 2.6576,
      "theoretical_loss": 3.520609367323961,
      "tokens_seen": 1478291456
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.0002788565697091274,
      "loss": 2.6144,
      "theoretical_loss": 3.520595568011219,
      "tokens_seen": 1478356992
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.00027884653961885656,
      "loss": 2.4924,
      "theoretical_loss": 3.520581769481467,
      "tokens_seen": 1478422528
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.00027883650952858574,
      "loss": 2.482,
      "theoretical_loss": 3.520567971734624,
      "tokens_seen": 1478488064
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.000278826479438315,
      "loss": 2.7568,
      "theoretical_loss": 3.5205541747706124,
      "tokens_seen": 1478553600
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.0002788164493480441,
      "loss": 2.5784,
      "theoretical_loss": 3.5205403785893523,
      "tokens_seen": 1478619136
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.00027880641925777334,
      "loss": 2.6064,
      "theoretical_loss": 3.5205265831907653,
      "tokens_seen": 1478684672
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.00027879638916750247,
      "loss": 2.2802,
      "theoretical_loss": 3.520512788574772,
      "tokens_seen": 1478750208
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.0002787863590772317,
      "loss": 2.5806,
      "theoretical_loss": 3.5204989947412924,
      "tokens_seen": 1478815744
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.0002787763289869609,
      "loss": 2.5576,
      "theoretical_loss": 3.5204852016902484,
      "tokens_seen": 1478881280
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.00027876629889669007,
      "loss": 2.5985,
      "theoretical_loss": 3.520471409421561,
      "tokens_seen": 1478946816
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.00027875626880641925,
      "loss": 2.5239,
      "theoretical_loss": 3.5204576179351506,
      "tokens_seen": 1479012352
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.0002787462387161485,
      "loss": 2.4756,
      "theoretical_loss": 3.520443827230939,
      "tokens_seen": 1479077888
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.0002787362086258776,
      "loss": 2.3017,
      "theoretical_loss": 3.520430037308846,
      "tokens_seen": 1479143424
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.00027872617853560685,
      "loss": 2.4728,
      "theoretical_loss": 3.520416248168794,
      "tokens_seen": 1479208960
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.00027871614844533597,
      "loss": 2.5853,
      "theoretical_loss": 3.5204024598107027,
      "tokens_seen": 1479274496
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.0002787061183550652,
      "loss": 2.6454,
      "theoretical_loss": 3.5203886722344935,
      "tokens_seen": 1479340032
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.0002786960882647944,
      "loss": 2.5003,
      "theoretical_loss": 3.5203748854400883,
      "tokens_seen": 1479405568
    },
    {
      "epoch": 4.09,
      "objective/train/docs_used": 1603030,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.675086259841919,
      "objective/train/theoretical_loss": 3.520361099427407,
      "objective/train/tokens_used": 1483791840,
      "theoretical_loss": 3.520361099427407,
      "tokens_seen": 1479471104
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.00027868605817452357,
      "loss": 2.5011,
      "theoretical_loss": 3.520361099427407,
      "tokens_seen": 1479471104
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.00027867602808425275,
      "loss": 2.553,
      "theoretical_loss": 3.5203473141963713,
      "tokens_seen": 1479536640
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.00027866599799398193,
      "loss": 2.6432,
      "theoretical_loss": 3.5203335297469014,
      "tokens_seen": 1479602176
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.0002786559679037111,
      "loss": 2.724,
      "theoretical_loss": 3.52031974607892,
      "tokens_seen": 1479667712
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.00027864593781344035,
      "loss": 2.5352,
      "theoretical_loss": 3.520305963192347,
      "tokens_seen": 1479733248
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.0002786359077231695,
      "loss": 2.5688,
      "theoretical_loss": 3.520292181087103,
      "tokens_seen": 1479798784
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.0002786258776328987,
      "loss": 2.5477,
      "theoretical_loss": 3.520278399763111,
      "tokens_seen": 1479864320
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.0002786158475426279,
      "loss": 2.4729,
      "theoretical_loss": 3.5202646192202898,
      "tokens_seen": 1479929856
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.00027860581745235707,
      "loss": 2.7487,
      "theoretical_loss": 3.520250839458562,
      "tokens_seen": 1479995392
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.00027859578736208625,
      "loss": 2.4694,
      "theoretical_loss": 3.520237060477849,
      "tokens_seen": 1480060928
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.00027858575727181544,
      "loss": 2.6762,
      "theoretical_loss": 3.520223282278071,
      "tokens_seen": 1480126464
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.0002785757271815446,
      "loss": 2.7235,
      "theoretical_loss": 3.520209504859149,
      "tokens_seen": 1480192000
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.00027856569709127385,
      "loss": 2.3901,
      "theoretical_loss": 3.5201957282210055,
      "tokens_seen": 1480257536
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.000278555667001003,
      "loss": 2.3777,
      "theoretical_loss": 3.5201819523635605,
      "tokens_seen": 1480323072
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.0002785456369107322,
      "loss": 2.2904,
      "theoretical_loss": 3.5201681772867355,
      "tokens_seen": 1480388608
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.00027853560682046134,
      "loss": 2.3948,
      "theoretical_loss": 3.5201544029904523,
      "tokens_seen": 1480454144
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.0002785255767301906,
      "loss": 2.6037,
      "theoretical_loss": 3.520140629474631,
      "tokens_seen": 1480519680
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.0002785155466399198,
      "loss": 2.7134,
      "theoretical_loss": 3.5201268567391937,
      "tokens_seen": 1480585216
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.00027850551654964894,
      "loss": 2.8329,
      "theoretical_loss": 3.520113084784062,
      "tokens_seen": 1480650752
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.0002784954864593782,
      "loss": 2.4178,
      "theoretical_loss": 3.5200993136091556,
      "tokens_seen": 1480716288
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.0002784854563691073,
      "loss": 2.4676,
      "theoretical_loss": 3.520085543214397,
      "tokens_seen": 1480781824
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.00027847542627883654,
      "loss": 2.1964,
      "theoretical_loss": 3.5200717735997076,
      "tokens_seen": 1480847360
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.0002784653961885657,
      "loss": 2.7747,
      "theoretical_loss": 3.5200580047650076,
      "tokens_seen": 1480912896
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.0002784553660982949,
      "loss": 2.5658,
      "theoretical_loss": 3.52004423671022,
      "tokens_seen": 1480978432
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.0002784453360080241,
      "loss": 2.6829,
      "theoretical_loss": 3.520030469435264,
      "tokens_seen": 1481043968
    },
    {
      "epoch": 4.09,
      "objective/train/docs_used": 1603030,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.576517105102539,
      "objective/train/theoretical_loss": 3.520016702940063,
      "objective/train/tokens_used": 1483791840,
      "theoretical_loss": 3.520016702940063,
      "tokens_seen": 1481109504
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.0002784353059177533,
      "loss": 2.6277,
      "theoretical_loss": 3.520016702940063,
      "tokens_seen": 1481109504
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.00027842527582748244,
      "loss": 2.4534,
      "theoretical_loss": 3.520002937224537,
      "tokens_seen": 1481175040
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.0002784152457372117,
      "loss": 2.5497,
      "theoretical_loss": 3.519989172288608,
      "tokens_seen": 1481240576
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.0002784052156469408,
      "loss": 2.5595,
      "theoretical_loss": 3.519975408132197,
      "tokens_seen": 1481306112
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.00027839518555667004,
      "loss": 2.4755,
      "theoretical_loss": 3.519961644755225,
      "tokens_seen": 1481371648
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.0002783851554663992,
      "loss": 2.5273,
      "theoretical_loss": 3.5199478821576142,
      "tokens_seen": 1481437184
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.0002783751253761284,
      "loss": 2.6981,
      "theoretical_loss": 3.5199341203392853,
      "tokens_seen": 1481502720
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.0002783650952858576,
      "loss": 2.6444,
      "theoretical_loss": 3.5199203593001602,
      "tokens_seen": 1481568256
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.00027835506519558676,
      "loss": 2.3927,
      "theoretical_loss": 3.5199065990401603,
      "tokens_seen": 1481633792
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.00027834503510531594,
      "loss": 2.4405,
      "theoretical_loss": 3.5198928395592075,
      "tokens_seen": 1481699328
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.0002783350050150452,
      "loss": 2.536,
      "theoretical_loss": 3.519879080857222,
      "tokens_seen": 1481764864
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.0002783249749247743,
      "loss": 2.4936,
      "theoretical_loss": 3.5198653229341255,
      "tokens_seen": 1481830400
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.00027831494483450354,
      "loss": 2.2565,
      "theoretical_loss": 3.51985156578984,
      "tokens_seen": 1481895936
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.00027830491474423267,
      "loss": 2.3814,
      "theoretical_loss": 3.5198378094242875,
      "tokens_seen": 1481961472
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.0002782948846539619,
      "loss": 2.4476,
      "theoretical_loss": 3.5198240538373877,
      "tokens_seen": 1482027008
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.0002782848545636911,
      "loss": 2.4944,
      "theoretical_loss": 3.5198102990290643,
      "tokens_seen": 1482092544
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.00027827482447342027,
      "loss": 2.4673,
      "theoretical_loss": 3.519796544999237,
      "tokens_seen": 1482158080
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.00027826479438314945,
      "loss": 2.3833,
      "theoretical_loss": 3.5197827917478284,
      "tokens_seen": 1482223616
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.0002782547642928787,
      "loss": 2.6754,
      "theoretical_loss": 3.5197690392747596,
      "tokens_seen": 1482289152
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.0002782447342026078,
      "loss": 2.2025,
      "theoretical_loss": 3.519755287579952,
      "tokens_seen": 1482354688
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.00027823470411233705,
      "loss": 2.522,
      "theoretical_loss": 3.519741536663327,
      "tokens_seen": 1482420224
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.00027822467402206617,
      "loss": 2.5283,
      "theoretical_loss": 3.519727786524807,
      "tokens_seen": 1482485760
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.0002782146439317954,
      "loss": 2.3373,
      "theoretical_loss": 3.519714037164313,
      "tokens_seen": 1482551296
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.0002782046138415246,
      "loss": 2.4523,
      "theoretical_loss": 3.5197002885817668,
      "tokens_seen": 1482616832
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.00027819458375125377,
      "loss": 2.8188,
      "theoretical_loss": 3.5196865407770894,
      "tokens_seen": 1482682368
    },
    {
      "epoch": 4.09,
      "objective/train/docs_used": 1603030,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.83261775970459,
      "objective/train/theoretical_loss": 3.5196727937502033,
      "objective/train/tokens_used": 1483791840,
      "theoretical_loss": 3.5196727937502033,
      "tokens_seen": 1482747904
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.00027818455366098295,
      "loss": 2.7207,
      "theoretical_loss": 3.5196727937502033,
      "tokens_seen": 1482747904
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.00027817452357071213,
      "loss": 2.5233,
      "theoretical_loss": 3.5196590475010296,
      "tokens_seen": 1482813440
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.0002781644934804413,
      "loss": 2.5227,
      "theoretical_loss": 3.51964530202949,
      "tokens_seen": 1482878976
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.00027815446339017055,
      "loss": 2.5866,
      "theoretical_loss": 3.519631557335506,
      "tokens_seen": 1482944512
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.0002781444332998997,
      "loss": 2.5737,
      "theoretical_loss": 3.519617813419,
      "tokens_seen": 1483010048
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.0002781344032096289,
      "loss": 2.4936,
      "theoretical_loss": 3.5196040702798923,
      "tokens_seen": 1483075584
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.0002781243731193581,
      "loss": 2.74,
      "theoretical_loss": 3.519590327918106,
      "tokens_seen": 1483141120
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.0002781143430290873,
      "loss": 2.5662,
      "theoretical_loss": 3.519576586333562,
      "tokens_seen": 1483206656
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.00027810431293881645,
      "loss": 2.6936,
      "theoretical_loss": 3.519562845526182,
      "tokens_seen": 1483272192
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.00027809428284854564,
      "loss": 2.5848,
      "theoretical_loss": 3.519549105495888,
      "tokens_seen": 1483337728
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.0002780842527582748,
      "loss": 2.735,
      "theoretical_loss": 3.5195353662426014,
      "tokens_seen": 1483403264
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.00027807422266800405,
      "loss": 2.4056,
      "theoretical_loss": 3.5195216277662444,
      "tokens_seen": 1483468800
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.0002780641925777332,
      "loss": 2.3115,
      "theoretical_loss": 3.5195078900667385,
      "tokens_seen": 1483534336
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.0002780541624874624,
      "loss": 2.5639,
      "theoretical_loss": 3.5194941531440054,
      "tokens_seen": 1483599872
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.00027804413239719154,
      "loss": 2.5111,
      "theoretical_loss": 3.5194804169979665,
      "tokens_seen": 1483665408
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.0002780341023069208,
      "loss": 2.6556,
      "theoretical_loss": 3.5194666816285447,
      "tokens_seen": 1483730944
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.00027802407221664996,
      "loss": 2.8688,
      "theoretical_loss": 3.519452947035661,
      "tokens_seen": 1483796480
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.00027801404212637914,
      "loss": 3.5425,
      "theoretical_loss": 3.5194359944683216,
      "tokens_seen": 1483877376
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002780040120361083,
      "loss": 2.8675,
      "theoretical_loss": 3.5194222616102326,
      "tokens_seen": 1483942912
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002779939819458375,
      "loss": 2.6917,
      "theoretical_loss": 3.5194085295284285,
      "tokens_seen": 1484008448
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002779839518555667,
      "loss": 2.8906,
      "theoretical_loss": 3.519394798222832,
      "tokens_seen": 1484073984
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002779739217652959,
      "loss": 2.6199,
      "theoretical_loss": 3.519381067693365,
      "tokens_seen": 1484139520
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.00027796389167502504,
      "loss": 2.701,
      "theoretical_loss": 3.519367337939948,
      "tokens_seen": 1484205056
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002779538615847543,
      "loss": 2.5548,
      "theoretical_loss": 3.519353608962504,
      "tokens_seen": 1484270592
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.00027794383149448346,
      "loss": 2.709,
      "theoretical_loss": 3.5193398807609544,
      "tokens_seen": 1484336128
    },
    {
      "epoch": 5.0,
      "objective/train/docs_used": 1667839,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.698960781097412,
      "objective/train/theoretical_loss": 3.5193295851189266,
      "objective/train/tokens_used": 1504845280,
      "theoretical_loss": 3.5193295851189266,
      "tokens_seen": 1484385280
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.00027793380140421264,
      "loss": 2.6594,
      "theoretical_loss": 3.519326153335222,
      "tokens_seen": 1484401664
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002779237713139418,
      "loss": 2.679,
      "theoretical_loss": 3.5193124266852274,
      "tokens_seen": 1484467200
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.000277913741223671,
      "loss": 2.6059,
      "theoretical_loss": 3.519298700810893,
      "tokens_seen": 1484532736
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002779037111334002,
      "loss": 2.5911,
      "theoretical_loss": 3.519284975712141,
      "tokens_seen": 1484598272
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002778936810431294,
      "loss": 2.542,
      "theoretical_loss": 3.5192712513888935,
      "tokens_seen": 1484663808
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.00027788365095285855,
      "loss": 2.634,
      "theoretical_loss": 3.5192575278410723,
      "tokens_seen": 1484729344
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002778736208625878,
      "loss": 2.5404,
      "theoretical_loss": 3.519243805068599,
      "tokens_seen": 1484794880
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002778635907723169,
      "loss": 2.4077,
      "theoretical_loss": 3.5192300830713954,
      "tokens_seen": 1484860416
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.00027785356068204614,
      "loss": 2.8491,
      "theoretical_loss": 3.5192163618493844,
      "tokens_seen": 1484925952
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002778435305917753,
      "loss": 2.6214,
      "theoretical_loss": 3.5192026414024875,
      "tokens_seen": 1484991488
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002778335005015045,
      "loss": 2.8468,
      "theoretical_loss": 3.5191889217306267,
      "tokens_seen": 1485057024
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002778234704112337,
      "loss": 2.8204,
      "theoretical_loss": 3.5191752028337238,
      "tokens_seen": 1485122560
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.00027781344032096287,
      "loss": 2.645,
      "theoretical_loss": 3.5191614847117005,
      "tokens_seen": 1485188096
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.00027780341023069205,
      "loss": 2.8198,
      "theoretical_loss": 3.51914776736448,
      "tokens_seen": 1485253632
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002777933801404213,
      "loss": 2.6427,
      "theoretical_loss": 3.519134050791984,
      "tokens_seen": 1485319168
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002777833500501504,
      "loss": 2.6776,
      "theoretical_loss": 3.519120334994134,
      "tokens_seen": 1485384704
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.00027777331995987965,
      "loss": 2.641,
      "theoretical_loss": 3.5191066199708523,
      "tokens_seen": 1485450240
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002777632898696089,
      "loss": 2.4597,
      "theoretical_loss": 3.5190929057220615,
      "tokens_seen": 1485515776
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.000277753259779338,
      "loss": 2.6771,
      "theoretical_loss": 3.5190791922476823,
      "tokens_seen": 1485581312
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.00027774322968906725,
      "loss": 2.6651,
      "theoretical_loss": 3.5190654795476384,
      "tokens_seen": 1485646848
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.00027773319959879637,
      "loss": 2.4468,
      "theoretical_loss": 3.5190517676218516,
      "tokens_seen": 1485712384
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002777231695085256,
      "loss": 2.6605,
      "theoretical_loss": 3.5190380564702433,
      "tokens_seen": 1485777920
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002777131394182548,
      "loss": 2.7455,
      "theoretical_loss": 3.519024346092736,
      "tokens_seen": 1485843456
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.00027770310932798397,
      "loss": 2.6425,
      "theoretical_loss": 3.519010636489252,
      "tokens_seen": 1485908992
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.00027769307923771315,
      "loss": 2.6573,
      "theoretical_loss": 3.518996927659714,
      "tokens_seen": 1485974528
    },
    {
      "epoch": 5.0,
      "objective/train/docs_used": 1672845,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.550705909729004,
      "objective/train/theoretical_loss": 3.5189866465454154,
      "objective/train/tokens_used": 1506483680,
      "theoretical_loss": 3.5189866465454154,
      "tokens_seen": 1486023680
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.00027768304914744233,
      "loss": 2.6352,
      "theoretical_loss": 3.518983219604043,
      "tokens_seen": 1486040064
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002776730190571715,
      "loss": 2.702,
      "theoretical_loss": 3.518969512322162,
      "tokens_seen": 1486105600
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.00027766298896690075,
      "loss": 2.739,
      "theoretical_loss": 3.518955805813993,
      "tokens_seen": 1486171136
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002776529588766299,
      "loss": 2.5802,
      "theoretical_loss": 3.518942100079458,
      "tokens_seen": 1486236672
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002776429287863591,
      "loss": 2.5728,
      "theoretical_loss": 3.5189283951184795,
      "tokens_seen": 1486302208
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002776328986960883,
      "loss": 2.5793,
      "theoretical_loss": 3.5189146909309796,
      "tokens_seen": 1486367744
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002776228686058175,
      "loss": 2.6295,
      "theoretical_loss": 3.5189009875168806,
      "tokens_seen": 1486433280
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.00027761283851554665,
      "loss": 2.6874,
      "theoretical_loss": 3.518887284876105,
      "tokens_seen": 1486498816
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.00027760280842527584,
      "loss": 2.6939,
      "theoretical_loss": 3.518873583008574,
      "tokens_seen": 1486564352
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.000277592778335005,
      "loss": 2.7765,
      "theoretical_loss": 3.518859881914212,
      "tokens_seen": 1486629888
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.00027758274824473425,
      "loss": 2.6764,
      "theoretical_loss": 3.518846181592939,
      "tokens_seen": 1486695424
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002775727181544634,
      "loss": 2.7328,
      "theoretical_loss": 3.518832482044678,
      "tokens_seen": 1486760960
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002775626880641926,
      "loss": 2.4887,
      "theoretical_loss": 3.518818783269352,
      "tokens_seen": 1486826496
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.00027755265797392174,
      "loss": 2.6795,
      "theoretical_loss": 3.5188050852668833,
      "tokens_seen": 1486892032
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.000277542627883651,
      "loss": 2.791,
      "theoretical_loss": 3.5187913880371937,
      "tokens_seen": 1486957568
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.00027753259779338016,
      "loss": 2.6841,
      "theoretical_loss": 3.5187776915802056,
      "tokens_seen": 1487023104
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.00027752256770310934,
      "loss": 2.5651,
      "theoretical_loss": 3.518763995895841,
      "tokens_seen": 1487088640
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002775125376128385,
      "loss": 2.7094,
      "theoretical_loss": 3.5187503009840233,
      "tokens_seen": 1487154176
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002775025075225677,
      "loss": 2.6081,
      "theoretical_loss": 3.5187366068446746,
      "tokens_seen": 1487219712
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002774924774322969,
      "loss": 2.62,
      "theoretical_loss": 3.518722913477716,
      "tokens_seen": 1487285248
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002774824473420261,
      "loss": 2.5542,
      "theoretical_loss": 3.518709220883072,
      "tokens_seen": 1487350784
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.00027747241725175524,
      "loss": 2.6896,
      "theoretical_loss": 3.518695529060663,
      "tokens_seen": 1487416320
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002774623871614845,
      "loss": 2.788,
      "theoretical_loss": 3.518681838010413,
      "tokens_seen": 1487481856
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.00027745235707121366,
      "loss": 2.6174,
      "theoretical_loss": 3.5186681477322437,
      "tokens_seen": 1487547392
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.00027744232698094284,
      "loss": 2.8734,
      "theoretical_loss": 3.5186544582260773,
      "tokens_seen": 1487612928
    },
    {
      "epoch": 5.0,
      "objective/train/docs_used": 1677864,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8672678470611572,
      "objective/train/theoretical_loss": 3.518644191603033,
      "objective/train/tokens_used": 1508122080,
      "theoretical_loss": 3.518644191603033,
      "tokens_seen": 1487662080
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.000277432296890672,
      "loss": 2.636,
      "theoretical_loss": 3.518640769491837,
      "tokens_seen": 1487678464
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002774222668004012,
      "loss": 2.7245,
      "theoretical_loss": 3.5186270815294445,
      "tokens_seen": 1487744000
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002774122367101304,
      "loss": 2.6811,
      "theoretical_loss": 3.5186133943388227,
      "tokens_seen": 1487809536
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002774022066198596,
      "loss": 2.6277,
      "theoretical_loss": 3.5185997079198943,
      "tokens_seen": 1487875072
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.00027739217652958875,
      "loss": 2.4533,
      "theoretical_loss": 3.518586022272581,
      "tokens_seen": 1487940608
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.000277382146439318,
      "loss": 2.6915,
      "theoretical_loss": 3.5185723373968063,
      "tokens_seen": 1488006144
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002773721163490471,
      "loss": 2.6598,
      "theoretical_loss": 3.5185586532924917,
      "tokens_seen": 1488071680
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.00027736208625877634,
      "loss": 2.5223,
      "theoretical_loss": 3.518544969959561,
      "tokens_seen": 1488137216
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002773520561685055,
      "loss": 2.6451,
      "theoretical_loss": 3.518531287397936,
      "tokens_seen": 1488202752
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002773420260782347,
      "loss": 2.6403,
      "theoretical_loss": 3.5185176056075393,
      "tokens_seen": 1488268288
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002773319959879639,
      "loss": 2.7058,
      "theoretical_loss": 3.518503924588293,
      "tokens_seen": 1488333824
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.00027732196589769307,
      "loss": 2.7885,
      "theoretical_loss": 3.5184902443401205,
      "tokens_seen": 1488399360
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.00027731193580742225,
      "loss": 2.7033,
      "theoretical_loss": 3.518476564862944,
      "tokens_seen": 1488464896
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002773019057171515,
      "loss": 2.6601,
      "theoretical_loss": 3.5184628861566867,
      "tokens_seen": 1488530432
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002772918756268806,
      "loss": 2.7296,
      "theoretical_loss": 3.51844920822127,
      "tokens_seen": 1488595968
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.00027728184553660985,
      "loss": 2.7617,
      "theoretical_loss": 3.5184355310566175,
      "tokens_seen": 1488661504
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.00027727181544633903,
      "loss": 2.736,
      "theoretical_loss": 3.518421854662652,
      "tokens_seen": 1488727040
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002772617853560682,
      "loss": 2.7205,
      "theoretical_loss": 3.5184081790392945,
      "tokens_seen": 1488792576
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002772517552657974,
      "loss": 2.8174,
      "theoretical_loss": 3.51839450418647,
      "tokens_seen": 1488858112
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.00027724172517552657,
      "loss": 2.6738,
      "theoretical_loss": 3.5183808301040997,
      "tokens_seen": 1488923648
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.00027723169508525575,
      "loss": 2.6079,
      "theoretical_loss": 3.518367156792106,
      "tokens_seen": 1488989184
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.000277221664994985,
      "loss": 2.7517,
      "theoretical_loss": 3.518353484250413,
      "tokens_seen": 1489054720
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002772116349047141,
      "loss": 2.6961,
      "theoretical_loss": 3.518339812478942,
      "tokens_seen": 1489120256
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.00027720160481444335,
      "loss": 2.6361,
      "theoretical_loss": 3.5183261414776172,
      "tokens_seen": 1489185792
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002771915747241725,
      "loss": 2.7121,
      "theoretical_loss": 3.51831247124636,
      "tokens_seen": 1489251328
    },
    {
      "epoch": 5.0,
      "objective/train/docs_used": 1682798,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.633384943008423,
      "objective/train/theoretical_loss": 3.5183022190782274,
      "objective/train/tokens_used": 1509760480,
      "theoretical_loss": 3.5183022190782274,
      "tokens_seen": 1489300480
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002771815446339017,
      "loss": 2.6501,
      "theoretical_loss": 3.518298801785093,
      "tokens_seen": 1489316864
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002771715145436309,
      "loss": 2.6853,
      "theoretical_loss": 3.51828513309374,
      "tokens_seen": 1489382400
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002771614844533601,
      "loss": 2.7524,
      "theoretical_loss": 3.5182714651722238,
      "tokens_seen": 1489447936
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.00027715145436308926,
      "loss": 2.6166,
      "theoretical_loss": 3.5182577980204663,
      "tokens_seen": 1489513472
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002771414242728185,
      "loss": 2.7881,
      "theoretical_loss": 3.5182441316383906,
      "tokens_seen": 1489579008
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002771313941825476,
      "loss": 2.5965,
      "theoretical_loss": 3.5182304660259196,
      "tokens_seen": 1489644544
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.00027712136409227685,
      "loss": 2.7613,
      "theoretical_loss": 3.518216801182976,
      "tokens_seen": 1489710080
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.000277111334002006,
      "loss": 2.791,
      "theoretical_loss": 3.5182031371094826,
      "tokens_seen": 1489775616
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002771013039117352,
      "loss": 2.5291,
      "theoretical_loss": 3.5181894738053625,
      "tokens_seen": 1489841152
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002770912738214644,
      "loss": 2.6356,
      "theoretical_loss": 3.518175811270538,
      "tokens_seen": 1489906688
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002770812437311936,
      "loss": 2.7813,
      "theoretical_loss": 3.518162149504933,
      "tokens_seen": 1489972224
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.00027707121364092276,
      "loss": 2.685,
      "theoretical_loss": 3.5181484885084693,
      "tokens_seen": 1490037760
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.00027706118355065194,
      "loss": 2.5715,
      "theoretical_loss": 3.5181348282810703,
      "tokens_seen": 1490103296
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002770511534603811,
      "loss": 2.4107,
      "theoretical_loss": 3.518121168822658,
      "tokens_seen": 1490168832
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.00027704112337011036,
      "loss": 2.5857,
      "theoretical_loss": 3.5181075101331567,
      "tokens_seen": 1490234368
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002770310932798395,
      "loss": 2.7753,
      "theoretical_loss": 3.5180938522124885,
      "tokens_seen": 1490299904
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002770210631895687,
      "loss": 2.6862,
      "theoretical_loss": 3.518080195060576,
      "tokens_seen": 1490365440
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002770110330992979,
      "loss": 2.7125,
      "theoretical_loss": 3.518066538677343,
      "tokens_seen": 1490430976
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002770010030090271,
      "loss": 2.6126,
      "theoretical_loss": 3.518052883062712,
      "tokens_seen": 1490496512
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002769909729187563,
      "loss": 2.5852,
      "theoretical_loss": 3.5180392282166055,
      "tokens_seen": 1490562048
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.00027698094282848544,
      "loss": 2.5653,
      "theoretical_loss": 3.5180255741389477,
      "tokens_seen": 1490627584
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002769709127382147,
      "loss": 2.7681,
      "theoretical_loss": 3.51801192082966,
      "tokens_seen": 1490693120
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.00027696088264794386,
      "loss": 2.6062,
      "theoretical_loss": 3.5179982682886664,
      "tokens_seen": 1490758656
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.00027695085255767304,
      "loss": 2.831,
      "theoretical_loss": 3.5179846165158897,
      "tokens_seen": 1490824192
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002769408224674022,
      "loss": 2.7401,
      "theoretical_loss": 3.5179709655112528,
      "tokens_seen": 1490889728
    },
    {
      "epoch": 5.0,
      "objective/train/docs_used": 1687806,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.763873815536499,
      "objective/train/theoretical_loss": 3.517960727761821,
      "objective/train/tokens_used": 1511398880,
      "theoretical_loss": 3.517960727761821,
      "tokens_seen": 1490938880
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002769307923771314,
      "loss": 2.7617,
      "theoretical_loss": 3.5179573152746793,
      "tokens_seen": 1490955264
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002769207622868606,
      "loss": 2.5945,
      "theoretical_loss": 3.517943665806091,
      "tokens_seen": 1491020800
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002769107321965898,
      "loss": 2.783,
      "theoretical_loss": 3.517930017105412,
      "tokens_seen": 1491086336
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.00027690070210631895,
      "loss": 2.6221,
      "theoretical_loss": 3.5179163691725646,
      "tokens_seen": 1491151872
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002768906720160482,
      "loss": 2.5724,
      "theoretical_loss": 3.517902722007473,
      "tokens_seen": 1491217408
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002768806419257773,
      "loss": 2.7295,
      "theoretical_loss": 3.517889075610059,
      "tokens_seen": 1491282944
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.00027687061183550655,
      "loss": 2.7662,
      "theoretical_loss": 3.5178754299802457,
      "tokens_seen": 1491348480
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002768605817452357,
      "loss": 2.7277,
      "theoretical_loss": 3.517861785117957,
      "tokens_seen": 1491414016
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002768505516549649,
      "loss": 2.7318,
      "theoretical_loss": 3.5178481410231166,
      "tokens_seen": 1491479552
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002768405215646941,
      "loss": 2.7809,
      "theoretical_loss": 3.517834497695646,
      "tokens_seen": 1491545088
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.00027683049147442327,
      "loss": 2.6275,
      "theoretical_loss": 3.517820855135469,
      "tokens_seen": 1491610624
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.00027682046138415245,
      "loss": 2.5591,
      "theoretical_loss": 3.5178072133425093,
      "tokens_seen": 1491676160
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002768104312938817,
      "loss": 2.908,
      "theoretical_loss": 3.517793572316689,
      "tokens_seen": 1491741696
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002768004012036108,
      "loss": 2.5793,
      "theoretical_loss": 3.517779932057932,
      "tokens_seen": 1491807232
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.00027679037111334005,
      "loss": 2.6962,
      "theoretical_loss": 3.5177662925661615,
      "tokens_seen": 1491872768
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.00027678034102306923,
      "loss": 2.4658,
      "theoretical_loss": 3.5177526538413,
      "tokens_seen": 1491938304
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002767703109327984,
      "loss": 2.5116,
      "theoretical_loss": 3.5177390158832713,
      "tokens_seen": 1492003840
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002767602808425276,
      "loss": 2.6111,
      "theoretical_loss": 3.517725378691998,
      "tokens_seen": 1492069376
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.00027675025075225677,
      "loss": 2.6248,
      "theoretical_loss": 3.5177117422674047,
      "tokens_seen": 1492134912
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.00027674022066198595,
      "loss": 2.7078,
      "theoretical_loss": 3.517698106609413,
      "tokens_seen": 1492200448
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002767301905717152,
      "loss": 2.6051,
      "theoretical_loss": 3.5176844717179474,
      "tokens_seen": 1492265984
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002767201604814443,
      "loss": 2.6811,
      "theoretical_loss": 3.51767083759293,
      "tokens_seen": 1492331520
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.00027671013039117355,
      "loss": 2.5908,
      "theoretical_loss": 3.5176572042342853,
      "tokens_seen": 1492397056
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002767001003009027,
      "loss": 2.817,
      "theoretical_loss": 3.517643571641935,
      "tokens_seen": 1492462592
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002766900702106319,
      "loss": 2.8072,
      "theoretical_loss": 3.517629939815804,
      "tokens_seen": 1492528128
    },
    {
      "epoch": 5.0,
      "objective/train/docs_used": 1692905,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.619441032409668,
      "objective/train/theoretical_loss": 3.5176197164489906,
      "objective/train/tokens_used": 1513037280,
      "theoretical_loss": 3.5176197164489906,
      "tokens_seen": 1492577280
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002766800401203611,
      "loss": 2.6709,
      "theoretical_loss": 3.517616308755815,
      "tokens_seen": 1492593664
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002766700100300903,
      "loss": 2.7228,
      "theoretical_loss": 3.5176026784618912,
      "tokens_seen": 1492659200
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.00027665997993981946,
      "loss": 2.8684,
      "theoretical_loss": 3.5175890489339556,
      "tokens_seen": 1492724736
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002766499498495487,
      "loss": 2.6712,
      "theoretical_loss": 3.517575420171932,
      "tokens_seen": 1492790272
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002766399197592778,
      "loss": 2.5039,
      "theoretical_loss": 3.5175617921757434,
      "tokens_seen": 1492855808
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.00027662988966900705,
      "loss": 2.7792,
      "theoretical_loss": 3.517548164945313,
      "tokens_seen": 1492921344
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002766198595787362,
      "loss": 2.8762,
      "theoretical_loss": 3.517534538480565,
      "tokens_seen": 1492986880
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002766098294884654,
      "loss": 2.7011,
      "theoretical_loss": 3.517520912781422,
      "tokens_seen": 1493052416
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002765997993981946,
      "loss": 2.6816,
      "theoretical_loss": 3.517507287847808,
      "tokens_seen": 1493117952
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002765897693079238,
      "loss": 2.6194,
      "theoretical_loss": 3.517493663679646,
      "tokens_seen": 1493183488
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.00027657973921765296,
      "loss": 2.918,
      "theoretical_loss": 3.517480040276859,
      "tokens_seen": 1493249024
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.00027656970912738214,
      "loss": 2.5906,
      "theoretical_loss": 3.517466417639371,
      "tokens_seen": 1493314560
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002765596790371113,
      "loss": 2.7555,
      "theoretical_loss": 3.517452795767105,
      "tokens_seen": 1493380096
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.00027654964894684056,
      "loss": 2.8005,
      "theoretical_loss": 3.5174391746599856,
      "tokens_seen": 1493445632
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002765396188565697,
      "loss": 2.5952,
      "theoretical_loss": 3.517425554317935,
      "tokens_seen": 1493511168
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002765295887662989,
      "loss": 2.6816,
      "theoretical_loss": 3.5174119347408768,
      "tokens_seen": 1493576704
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.00027651955867602805,
      "loss": 2.6175,
      "theoretical_loss": 3.5173983159287348,
      "tokens_seen": 1493642240
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002765095285857573,
      "loss": 2.6861,
      "theoretical_loss": 3.5173846978814316,
      "tokens_seen": 1493707776
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.00027649949849548646,
      "loss": 2.6073,
      "theoretical_loss": 3.5173710805988927,
      "tokens_seen": 1493773312
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.00027648946840521564,
      "loss": 2.5486,
      "theoretical_loss": 3.51735746408104,
      "tokens_seen": 1493838848
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002764794383149448,
      "loss": 2.7331,
      "theoretical_loss": 3.517343848327797,
      "tokens_seen": 1493904384
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.00027646940822467406,
      "loss": 2.6075,
      "theoretical_loss": 3.517330233339088,
      "tokens_seen": 1493969920
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002764593781344032,
      "loss": 2.7378,
      "theoretical_loss": 3.517316619114836,
      "tokens_seen": 1494035456
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002764493480441324,
      "loss": 2.7399,
      "theoretical_loss": 3.517303005654964,
      "tokens_seen": 1494100992
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.00027643931795386155,
      "loss": 2.465,
      "theoretical_loss": 3.517289392959397,
      "tokens_seen": 1494166528
    },
    {
      "epoch": 5.0,
      "objective/train/docs_used": 1697937,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4400620460510254,
      "objective/train/theoretical_loss": 3.5172791839392508,
      "objective/train/tokens_used": 1514675680,
      "theoretical_loss": 3.5172791839392508,
      "tokens_seen": 1494215680
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002764292878635908,
      "loss": 2.6827,
      "theoretical_loss": 3.517275781028058,
      "tokens_seen": 1494232064
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.00027641925777331997,
      "loss": 2.6565,
      "theoretical_loss": 3.51726216986087,
      "tokens_seen": 1494297600
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.00027640922768304915,
      "loss": 2.4964,
      "theoretical_loss": 3.517248559457757,
      "tokens_seen": 1494363136
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.00027639919759277833,
      "loss": 2.8277,
      "theoretical_loss": 3.5172349498186426,
      "tokens_seen": 1494428672
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002763891675025075,
      "loss": 2.545,
      "theoretical_loss": 3.5172213409434505,
      "tokens_seen": 1494494208
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002763791374122367,
      "loss": 2.6756,
      "theoretical_loss": 3.517207732832104,
      "tokens_seen": 1494559744
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002763691073219659,
      "loss": 2.6207,
      "theoretical_loss": 3.5171941254845267,
      "tokens_seen": 1494625280
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.00027635907723169505,
      "loss": 2.6007,
      "theoretical_loss": 3.517180518900643,
      "tokens_seen": 1494690816
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002763490471414243,
      "loss": 2.6077,
      "theoretical_loss": 3.517166913080376,
      "tokens_seen": 1494756352
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002763390170511534,
      "loss": 2.7704,
      "theoretical_loss": 3.5171533080236497,
      "tokens_seen": 1494821888
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.00027632898696088265,
      "loss": 2.7405,
      "theoretical_loss": 3.517139703730387,
      "tokens_seen": 1494887424
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.00027631895687061183,
      "loss": 2.606,
      "theoretical_loss": 3.5171261002005125,
      "tokens_seen": 1494952960
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.000276308926780341,
      "loss": 2.5175,
      "theoretical_loss": 3.517112497433949,
      "tokens_seen": 1495018496
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002762988966900702,
      "loss": 2.661,
      "theoretical_loss": 3.517098895430621,
      "tokens_seen": 1495084032
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.00027628886659979943,
      "loss": 2.6579,
      "theoretical_loss": 3.5170852941904522,
      "tokens_seen": 1495149568
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.00027627883650952856,
      "loss": 2.5546,
      "theoretical_loss": 3.5170716937133655,
      "tokens_seen": 1495215104
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002762688064192578,
      "loss": 2.6988,
      "theoretical_loss": 3.5170580939992853,
      "tokens_seen": 1495280640
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.00027625877632898697,
      "loss": 2.7503,
      "theoretical_loss": 3.5170444950481357,
      "tokens_seen": 1495346176
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.00027624874623871615,
      "loss": 2.5672,
      "theoretical_loss": 3.51703089685984,
      "tokens_seen": 1495411712
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002762387161484454,
      "loss": 2.8523,
      "theoretical_loss": 3.517017299434322,
      "tokens_seen": 1495477248
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002762286860581745,
      "loss": 2.598,
      "theoretical_loss": 3.517003702771505,
      "tokens_seen": 1495542784
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.00027621865596790375,
      "loss": 2.5702,
      "theoretical_loss": 3.5169901068713134,
      "tokens_seen": 1495608320
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002762086258776329,
      "loss": 2.7542,
      "theoretical_loss": 3.516976511733671,
      "tokens_seen": 1495673856
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002761985957873621,
      "loss": 2.7126,
      "theoretical_loss": 3.516962917358502,
      "tokens_seen": 1495739392
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002761885656970913,
      "loss": 2.6053,
      "theoretical_loss": 3.516949323745729,
      "tokens_seen": 1495804928
    },
    {
      "epoch": 5.0,
      "objective/train/docs_used": 1700816,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9013986587524414,
      "objective/train/theoretical_loss": 3.5169391290364267,
      "objective/train/tokens_used": 1516314080,
      "theoretical_loss": 3.5169391290364267,
      "tokens_seen": 1495854080
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002761785356068205,
      "loss": 2.8365,
      "theoretical_loss": 3.516935730895277,
      "tokens_seen": 1495870464
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.00027616850551654966,
      "loss": 2.5379,
      "theoretical_loss": 3.51692213880707,
      "tokens_seen": 1495936000
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002761584754262789,
      "loss": 2.7247,
      "theoretical_loss": 3.51690854748103,
      "tokens_seen": 1496001536
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.000276148445336008,
      "loss": 2.5445,
      "theoretical_loss": 3.516894956917083,
      "tokens_seen": 1496067072
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.00027613841524573725,
      "loss": 2.6178,
      "theoretical_loss": 3.516881367115152,
      "tokens_seen": 1496132608
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002761283851554664,
      "loss": 2.7579,
      "theoretical_loss": 3.5168677780751616,
      "tokens_seen": 1496198144
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002761183550651956,
      "loss": 2.7081,
      "theoretical_loss": 3.5168541897970345,
      "tokens_seen": 1496263680
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002761083249749248,
      "loss": 2.5498,
      "theoretical_loss": 3.516840602280695,
      "tokens_seen": 1496329216
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.000276098294884654,
      "loss": 2.5387,
      "theoretical_loss": 3.5168270155260677,
      "tokens_seen": 1496394752
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.00027608826479438316,
      "loss": 2.6254,
      "theoretical_loss": 3.5168134295330757,
      "tokens_seen": 1496460288
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.00027607823470411234,
      "loss": 2.6312,
      "theoretical_loss": 3.5167998443016435,
      "tokens_seen": 1496525824
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002760682046138415,
      "loss": 2.6842,
      "theoretical_loss": 3.516786259831695,
      "tokens_seen": 1496591360
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.00027605817452357076,
      "loss": 2.6796,
      "theoretical_loss": 3.516772676123154,
      "tokens_seen": 1496656896
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002760481444332999,
      "loss": 2.6596,
      "theoretical_loss": 3.5167590931759447,
      "tokens_seen": 1496722432
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002760381143430291,
      "loss": 2.7718,
      "theoretical_loss": 3.5167455109899906,
      "tokens_seen": 1496787968
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.00027602808425275825,
      "loss": 2.7429,
      "theoretical_loss": 3.516731929565216,
      "tokens_seen": 1496853504
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002760180541624875,
      "loss": 2.7515,
      "theoretical_loss": 3.5167183489015454,
      "tokens_seen": 1496919040
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.00027600802407221666,
      "loss": 2.7387,
      "theoretical_loss": 3.5167047689989026,
      "tokens_seen": 1496984576
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.00027599799398194584,
      "loss": 2.6507,
      "theoretical_loss": 3.5166911898572106,
      "tokens_seen": 1497050112
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.000275987963891675,
      "loss": 2.7042,
      "theoretical_loss": 3.5166776114763953,
      "tokens_seen": 1497115648
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.00027597793380140426,
      "loss": 2.6221,
      "theoretical_loss": 3.516664033856379,
      "tokens_seen": 1497181184
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002759679037111334,
      "loss": 2.7729,
      "theoretical_loss": 3.5166504569970867,
      "tokens_seen": 1497246720
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002759578736208626,
      "loss": 2.7716,
      "theoretical_loss": 3.5166368808984423,
      "tokens_seen": 1497312256
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.00027594784353059175,
      "loss": 2.6233,
      "theoretical_loss": 3.51662330556037,
      "tokens_seen": 1497377792
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.000275937813440321,
      "loss": 2.721,
      "theoretical_loss": 3.516609730982794,
      "tokens_seen": 1497443328
    },
    {
      "epoch": 5.0,
      "objective/train/docs_used": 1707842,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.416391372680664,
      "objective/train/theoretical_loss": 3.5165995505486416,
      "objective/train/tokens_used": 1517952480,
      "theoretical_loss": 3.5165995505486416,
      "tokens_seen": 1497492480
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.00027592778335005017,
      "loss": 2.5181,
      "theoretical_loss": 3.516596157165638,
      "tokens_seen": 1497508864
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.00027591775325977935,
      "loss": 2.5366,
      "theoretical_loss": 3.5165825841088267,
      "tokens_seen": 1497574400
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.00027590772316950853,
      "loss": 2.6365,
      "theoretical_loss": 3.5165690118122837,
      "tokens_seen": 1497639936
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002758976930792377,
      "loss": 2.67,
      "theoretical_loss": 3.5165554402759334,
      "tokens_seen": 1497705472
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002758876629889669,
      "loss": 2.5694,
      "theoretical_loss": 3.5165418694996995,
      "tokens_seen": 1497771008
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002758776328986961,
      "loss": 2.6292,
      "theoretical_loss": 3.5165282994835065,
      "tokens_seen": 1497836544
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.00027586760280842525,
      "loss": 2.4334,
      "theoretical_loss": 3.5165147302272795,
      "tokens_seen": 1497902080
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002758575727181545,
      "loss": 2.6236,
      "theoretical_loss": 3.5165011617309414,
      "tokens_seen": 1497967616
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002758475426278836,
      "loss": 2.4739,
      "theoretical_loss": 3.516487593994417,
      "tokens_seen": 1498033152
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.00027583751253761285,
      "loss": 2.5641,
      "theoretical_loss": 3.51647402701763,
      "tokens_seen": 1498098688
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.00027582748244734203,
      "loss": 2.6901,
      "theoretical_loss": 3.516460460800505,
      "tokens_seen": 1498164224
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002758174523570712,
      "loss": 2.6487,
      "theoretical_loss": 3.5164468953429666,
      "tokens_seen": 1498229760
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002758074222668004,
      "loss": 2.5901,
      "theoretical_loss": 3.5164333306449382,
      "tokens_seen": 1498295296
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.00027579739217652963,
      "loss": 2.5624,
      "theoretical_loss": 3.5164197667063455,
      "tokens_seen": 1498360832
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.00027578736208625876,
      "loss": 2.7026,
      "theoretical_loss": 3.5164062035271106,
      "tokens_seen": 1498426368
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.000275777331995988,
      "loss": 2.7093,
      "theoretical_loss": 3.51639264110716,
      "tokens_seen": 1498491904
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002757673019057171,
      "loss": 2.5889,
      "theoretical_loss": 3.5163790794464163,
      "tokens_seen": 1498557440
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.00027575727181544635,
      "loss": 2.5409,
      "theoretical_loss": 3.516365518544805,
      "tokens_seen": 1498622976
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.00027574724172517554,
      "loss": 2.592,
      "theoretical_loss": 3.5163519584022493,
      "tokens_seen": 1498688512
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002757372116349047,
      "loss": 2.6117,
      "theoretical_loss": 3.5163383990186743,
      "tokens_seen": 1498754048
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002757271815446339,
      "loss": 2.6196,
      "theoretical_loss": 3.516324840394004,
      "tokens_seen": 1498819584
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002757171514543631,
      "loss": 2.6298,
      "theoretical_loss": 3.5163112825281635,
      "tokens_seen": 1498885120
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.00027570712136409226,
      "loss": 2.6609,
      "theoretical_loss": 3.516297725421076,
      "tokens_seen": 1498950656
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002756970912738215,
      "loss": 2.7129,
      "theoretical_loss": 3.5162841690726667,
      "tokens_seen": 1499016192
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002756870611835506,
      "loss": 2.6601,
      "theoretical_loss": 3.5162706134828596,
      "tokens_seen": 1499081728
    },
    {
      "epoch": 5.0,
      "objective/train/docs_used": 1710663,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6549386978149414,
      "objective/train/theoretical_loss": 3.5162604472882912,
      "objective/train/tokens_used": 1519590880,
      "theoretical_loss": 3.5162604472882912,
      "tokens_seen": 1499130880
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.00027567703109327986,
      "loss": 2.6568,
      "theoretical_loss": 3.516257058651579,
      "tokens_seen": 1499147264
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.000275667001003009,
      "loss": 2.7675,
      "theoretical_loss": 3.5162435045787497,
      "tokens_seen": 1499212800
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002756569709127382,
      "loss": 2.6526,
      "theoretical_loss": 3.5162299512642954,
      "tokens_seen": 1499278336
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002756469408224674,
      "loss": 2.6224,
      "theoretical_loss": 3.516216398708141,
      "tokens_seen": 1499343872
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002756369107321966,
      "loss": 2.682,
      "theoretical_loss": 3.516202846910212,
      "tokens_seen": 1499409408
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.00027562688064192576,
      "loss": 2.7775,
      "theoretical_loss": 3.516189295870431,
      "tokens_seen": 1499474944
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.000275616850551655,
      "loss": 2.6704,
      "theoretical_loss": 3.5161757455887233,
      "tokens_seen": 1499540480
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002756068204613841,
      "loss": 2.8034,
      "theoretical_loss": 3.5161621960650136,
      "tokens_seen": 1499606016
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.00027559679037111336,
      "loss": 2.5323,
      "theoretical_loss": 3.516148647299225,
      "tokens_seen": 1499671552
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002755867602808425,
      "loss": 2.7179,
      "theoretical_loss": 3.5161350992912843,
      "tokens_seen": 1499737088
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002755767301905717,
      "loss": 2.5213,
      "theoretical_loss": 3.516121552041114,
      "tokens_seen": 1499802624
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002755667001003009,
      "loss": 2.5912,
      "theoretical_loss": 3.51610800554864,
      "tokens_seen": 1499868160
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002755566700100301,
      "loss": 2.6746,
      "theoretical_loss": 3.5160944598137855,
      "tokens_seen": 1499933696
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.00027554663991975927,
      "loss": 2.6959,
      "theoretical_loss": 3.5160809148364764,
      "tokens_seen": 1499999232
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.00027553660982948845,
      "loss": 2.5214,
      "theoretical_loss": 3.5160673706166357,
      "tokens_seen": 1500064768
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.00027552657973921763,
      "loss": 2.7854,
      "theoretical_loss": 3.5160538271541895,
      "tokens_seen": 1500130304
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.00027551654964894686,
      "loss": 2.6988,
      "theoretical_loss": 3.516040284449061,
      "tokens_seen": 1500195840
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.00027550651955867604,
      "loss": 2.6114,
      "theoretical_loss": 3.516026742501176,
      "tokens_seen": 1500261376
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002754964894684052,
      "loss": 2.5612,
      "theoretical_loss": 3.516013201310458,
      "tokens_seen": 1500326912
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027548645937813446,
      "loss": 2.5483,
      "theoretical_loss": 3.515999660876833,
      "tokens_seen": 1500392448
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002754764292878636,
      "loss": 2.7099,
      "theoretical_loss": 3.515986121200224,
      "tokens_seen": 1500457984
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002754663991975928,
      "loss": 2.6436,
      "theoretical_loss": 3.5159725822805568,
      "tokens_seen": 1500523520
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027545636910732195,
      "loss": 2.7573,
      "theoretical_loss": 3.5159590441177544,
      "tokens_seen": 1500589056
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002754463390170512,
      "loss": 2.7131,
      "theoretical_loss": 3.515945506711744,
      "tokens_seen": 1500654592
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027543630892678037,
      "loss": 2.5178,
      "theoretical_loss": 3.5159319700624483,
      "tokens_seen": 1500720128
    },
    {
      "epoch": 5.01,
      "objective/train/docs_used": 1715522,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.640793800354004,
      "objective/train/theoretical_loss": 3.515921818072025,
      "objective/train/tokens_used": 1521229280,
      "theoretical_loss": 3.515921818072025,
      "tokens_seen": 1500769280
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027542627883650955,
      "loss": 2.7359,
      "theoretical_loss": 3.5159184341697918,
      "tokens_seen": 1500785664
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027541624874623873,
      "loss": 2.7044,
      "theoretical_loss": 3.5159048990337007,
      "tokens_seen": 1500851200
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002754062186559679,
      "loss": 2.5357,
      "theoretical_loss": 3.515891364654099,
      "tokens_seen": 1500916736
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002753961885656971,
      "loss": 2.7109,
      "theoretical_loss": 3.5158778310309104,
      "tokens_seen": 1500982272
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002753861584754263,
      "loss": 2.6688,
      "theoretical_loss": 3.5158642981640607,
      "tokens_seen": 1501047808
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027537612838515545,
      "loss": 2.5025,
      "theoretical_loss": 3.5158507660534744,
      "tokens_seen": 1501113344
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002753660982948847,
      "loss": 2.4948,
      "theoretical_loss": 3.5158372346990765,
      "tokens_seen": 1501178880
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002753560682046138,
      "loss": 2.6206,
      "theoretical_loss": 3.515823704100791,
      "tokens_seen": 1501244416
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027534603811434305,
      "loss": 2.5264,
      "theoretical_loss": 3.5158101742585437,
      "tokens_seen": 1501309952
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027533600802407223,
      "loss": 2.7406,
      "theoretical_loss": 3.5157966451722578,
      "tokens_seen": 1501375488
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002753259779338014,
      "loss": 2.6902,
      "theoretical_loss": 3.515783116841859,
      "tokens_seen": 1501441024
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002753159478435306,
      "loss": 2.6181,
      "theoretical_loss": 3.515769589267273,
      "tokens_seen": 1501506560
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027530591775325983,
      "loss": 2.6366,
      "theoretical_loss": 3.515756062448423,
      "tokens_seen": 1501572096
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027529588766298896,
      "loss": 2.6837,
      "theoretical_loss": 3.5157425363852344,
      "tokens_seen": 1501637632
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002752858575727182,
      "loss": 2.7563,
      "theoretical_loss": 3.515729011077632,
      "tokens_seen": 1501703168
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002752758274824473,
      "loss": 2.6408,
      "theoretical_loss": 3.515715486525541,
      "tokens_seen": 1501768704
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027526579739217655,
      "loss": 2.487,
      "theoretical_loss": 3.515701962728885,
      "tokens_seen": 1501834240
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027525576730190574,
      "loss": 2.7103,
      "theoretical_loss": 3.5156884396875907,
      "tokens_seen": 1501899776
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002752457372116349,
      "loss": 2.5855,
      "theoretical_loss": 3.5156749174015816,
      "tokens_seen": 1501965312
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002752357071213641,
      "loss": 2.6207,
      "theoretical_loss": 3.515661395870783,
      "tokens_seen": 1502030848
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002752256770310933,
      "loss": 2.7456,
      "theoretical_loss": 3.5156478750951194,
      "tokens_seen": 1502096384
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027521564694082246,
      "loss": 2.5779,
      "theoretical_loss": 3.515634355074517,
      "tokens_seen": 1502161920
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002752056168505517,
      "loss": 2.5976,
      "theoretical_loss": 3.5156208358088987,
      "tokens_seen": 1502227456
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002751955867602808,
      "loss": 2.5067,
      "theoretical_loss": 3.5156073172981905,
      "tokens_seen": 1502292992
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027518555667001006,
      "loss": 2.6524,
      "theoretical_loss": 3.5155937995423177,
      "tokens_seen": 1502358528
    },
    {
      "epoch": 5.01,
      "objective/train/docs_used": 1720743,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.869342803955078,
      "objective/train/theoretical_loss": 3.515583661720728,
      "objective/train/tokens_used": 1522867680,
      "theoretical_loss": 3.515583661720728,
      "tokens_seen": 1502407680
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002751755265797392,
      "loss": 2.6944,
      "theoretical_loss": 3.5155802825412046,
      "tokens_seen": 1502424064
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002751654964894684,
      "loss": 2.6483,
      "theoretical_loss": 3.515566766294776,
      "tokens_seen": 1502489600
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002751554663991976,
      "loss": 2.5766,
      "theoretical_loss": 3.515553250802957,
      "tokens_seen": 1502555136
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002751454363089268,
      "loss": 2.8034,
      "theoretical_loss": 3.5155397360656733,
      "tokens_seen": 1502620672
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027513540621865596,
      "loss": 2.5673,
      "theoretical_loss": 3.5155262220828485,
      "tokens_seen": 1502686208
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002751253761283852,
      "loss": 2.6488,
      "theoretical_loss": 3.515512708854409,
      "tokens_seen": 1502751744
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002751153460381143,
      "loss": 2.5111,
      "theoretical_loss": 3.5154991963802793,
      "tokens_seen": 1502817280
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027510531594784356,
      "loss": 2.7109,
      "theoretical_loss": 3.5154856846603835,
      "tokens_seen": 1502882816
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002750952858575727,
      "loss": 2.5817,
      "theoretical_loss": 3.515472173694648,
      "tokens_seen": 1502948352
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002750852557673019,
      "loss": 2.5877,
      "theoretical_loss": 3.5154586634829967,
      "tokens_seen": 1503013888
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002750752256770311,
      "loss": 2.6702,
      "theoretical_loss": 3.5154451540253553,
      "tokens_seen": 1503079424
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002750651955867603,
      "loss": 2.7882,
      "theoretical_loss": 3.515431645321649,
      "tokens_seen": 1503144960
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027505516549648947,
      "loss": 2.5627,
      "theoretical_loss": 3.5154181373718023,
      "tokens_seen": 1503210496
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027504513540621865,
      "loss": 2.7719,
      "theoretical_loss": 3.5154046301757402,
      "tokens_seen": 1503276032
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027503510531594783,
      "loss": 2.8695,
      "theoretical_loss": 3.515391123733388,
      "tokens_seen": 1503341568
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027502507522567706,
      "loss": 2.4746,
      "theoretical_loss": 3.5153776180446714,
      "tokens_seen": 1503407104
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002750150451354062,
      "loss": 2.664,
      "theoretical_loss": 3.5153641131095146,
      "tokens_seen": 1503472640
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002750050150451354,
      "loss": 2.6918,
      "theoretical_loss": 3.515350608927843,
      "tokens_seen": 1503538176
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027499498495486455,
      "loss": 2.8506,
      "theoretical_loss": 3.515337105499582,
      "tokens_seen": 1503603712
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002749849548645938,
      "loss": 2.6836,
      "theoretical_loss": 3.5153236028246564,
      "tokens_seen": 1503669248
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027497492477432297,
      "loss": 2.752,
      "theoretical_loss": 3.5153101009029912,
      "tokens_seen": 1503734784
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027496489468405215,
      "loss": 2.6325,
      "theoretical_loss": 3.515296599734512,
      "tokens_seen": 1503800320
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027495486459378133,
      "loss": 2.7755,
      "theoretical_loss": 3.515283099319144,
      "tokens_seen": 1503865856
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027494483450351057,
      "loss": 2.7143,
      "theoretical_loss": 3.5152695996568117,
      "tokens_seen": 1503931392
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002749348044132397,
      "loss": 2.6559,
      "theoretical_loss": 3.515256100747441,
      "tokens_seen": 1503996928
    },
    {
      "epoch": 5.01,
      "objective/train/docs_used": 1725747,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6323916912078857,
      "objective/train/theoretical_loss": 3.515245977059499,
      "objective/train/tokens_used": 1524506080,
      "theoretical_loss": 3.515245977059499,
      "tokens_seen": 1504046080
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027492477432296893,
      "loss": 2.7248,
      "theoretical_loss": 3.515242602590957,
      "tokens_seen": 1504062464
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027491474423269806,
      "loss": 2.5724,
      "theoretical_loss": 3.5152291051872844,
      "tokens_seen": 1504128000
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002749047141424273,
      "loss": 2.6799,
      "theoretical_loss": 3.515215608536349,
      "tokens_seen": 1504193536
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027489468405215647,
      "loss": 2.515,
      "theoretical_loss": 3.5152021126380752,
      "tokens_seen": 1504259072
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027488465396188565,
      "loss": 2.4422,
      "theoretical_loss": 3.5151886174923894,
      "tokens_seen": 1504324608
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027487462387161483,
      "loss": 2.5653,
      "theoretical_loss": 3.5151751230992163,
      "tokens_seen": 1504390144
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.000274864593781344,
      "loss": 2.668,
      "theoretical_loss": 3.515161629458481,
      "tokens_seen": 1504455680
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002748545636910732,
      "loss": 2.6074,
      "theoretical_loss": 3.5151481365701085,
      "tokens_seen": 1504521216
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027484453360080243,
      "loss": 2.7049,
      "theoretical_loss": 3.515134644434025,
      "tokens_seen": 1504586752
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027483450351053156,
      "loss": 2.7778,
      "theoretical_loss": 3.5151211530501545,
      "tokens_seen": 1504652288
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002748244734202608,
      "loss": 2.4842,
      "theoretical_loss": 3.5151076624184237,
      "tokens_seen": 1504717824
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027481444332999,
      "loss": 2.7311,
      "theoretical_loss": 3.515094172538757,
      "tokens_seen": 1504783360
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027480441323971916,
      "loss": 2.7872,
      "theoretical_loss": 3.515080683411081,
      "tokens_seen": 1504848896
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027479438314944834,
      "loss": 2.5982,
      "theoretical_loss": 3.5150671950353187,
      "tokens_seen": 1504914432
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002747843530591775,
      "loss": 2.6789,
      "theoretical_loss": 3.515053707411397,
      "tokens_seen": 1504979968
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027477432296890675,
      "loss": 2.6325,
      "theoretical_loss": 3.5150402205392415,
      "tokens_seen": 1505045504
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027476429287863594,
      "loss": 2.5831,
      "theoretical_loss": 3.515026734418777,
      "tokens_seen": 1505111040
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002747542627883651,
      "loss": 2.6323,
      "theoretical_loss": 3.515013249049929,
      "tokens_seen": 1505176576
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002747442326980943,
      "loss": 2.5551,
      "theoretical_loss": 3.5149997644326225,
      "tokens_seen": 1505242112
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002747342026078235,
      "loss": 2.6337,
      "theoretical_loss": 3.5149862805667835,
      "tokens_seen": 1505307648
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027472417251755266,
      "loss": 2.5774,
      "theoretical_loss": 3.5149727974523373,
      "tokens_seen": 1505373184
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002747141424272819,
      "loss": 2.5681,
      "theoretical_loss": 3.514959315089209,
      "tokens_seen": 1505438720
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.000274704112337011,
      "loss": 2.6553,
      "theoretical_loss": 3.5149458334773245,
      "tokens_seen": 1505504256
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027469408224674026,
      "loss": 2.5902,
      "theoretical_loss": 3.5149323526166087,
      "tokens_seen": 1505569792
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002746840521564694,
      "loss": 2.5925,
      "theoretical_loss": 3.5149188725069873,
      "tokens_seen": 1505635328
    },
    {
      "epoch": 5.01,
      "objective/train/docs_used": 1730984,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6221024990081787,
      "objective/train/theoretical_loss": 3.514908762917632,
      "objective/train/tokens_used": 1526144480,
      "theoretical_loss": 3.514908762917632,
      "tokens_seen": 1505684480
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002746740220661986,
      "loss": 2.5532,
      "theoretical_loss": 3.5149053931483856,
      "tokens_seen": 1505700864
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002746639919759278,
      "loss": 2.5331,
      "theoretical_loss": 3.51489191454073,
      "tokens_seen": 1505766400
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.000274653961885657,
      "loss": 2.6332,
      "theoretical_loss": 3.5148784366839445,
      "tokens_seen": 1505831936
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027464393179538616,
      "loss": 2.7405,
      "theoretical_loss": 3.514864959577955,
      "tokens_seen": 1505897472
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002746339017051154,
      "loss": 2.7854,
      "theoretical_loss": 3.514851483222688,
      "tokens_seen": 1505963008
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002746238716148445,
      "loss": 2.6165,
      "theoretical_loss": 3.514838007618068,
      "tokens_seen": 1506028544
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027461384152457376,
      "loss": 2.7394,
      "theoretical_loss": 3.5148245327640213,
      "tokens_seen": 1506094080
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002746038114343029,
      "loss": 2.6271,
      "theoretical_loss": 3.514811058660473,
      "tokens_seen": 1506159616
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002745937813440321,
      "loss": 2.6569,
      "theoretical_loss": 3.514797585307348,
      "tokens_seen": 1506225152
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002745837512537613,
      "loss": 2.7429,
      "theoretical_loss": 3.514784112704573,
      "tokens_seen": 1506290688
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002745737211634905,
      "loss": 2.7054,
      "theoretical_loss": 3.514770640852073,
      "tokens_seen": 1506356224
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027456369107321967,
      "loss": 2.7931,
      "theoretical_loss": 3.5147571697497737,
      "tokens_seen": 1506421760
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027455366098294885,
      "loss": 2.6221,
      "theoretical_loss": 3.5147436993976005,
      "tokens_seen": 1506487296
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027454363089267803,
      "loss": 2.5807,
      "theoretical_loss": 3.5147302297954797,
      "tokens_seen": 1506552832
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027453360080240726,
      "loss": 2.5029,
      "theoretical_loss": 3.514716760943336,
      "tokens_seen": 1506618368
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002745235707121364,
      "loss": 2.5789,
      "theoretical_loss": 3.5147032928410953,
      "tokens_seen": 1506683904
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002745135406218656,
      "loss": 2.564,
      "theoretical_loss": 3.514689825488683,
      "tokens_seen": 1506749440
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027450351053159475,
      "loss": 2.6247,
      "theoretical_loss": 3.5146763588860255,
      "tokens_seen": 1506814976
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.000274493480441324,
      "loss": 2.6367,
      "theoretical_loss": 3.514662893033048,
      "tokens_seen": 1506880512
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027448345035105317,
      "loss": 2.6014,
      "theoretical_loss": 3.514649427929676,
      "tokens_seen": 1506946048
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027447342026078235,
      "loss": 2.6573,
      "theoretical_loss": 3.5146359635758353,
      "tokens_seen": 1507011584
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027446339017051153,
      "loss": 2.7415,
      "theoretical_loss": 3.514622499971452,
      "tokens_seen": 1507077120
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027445336008024077,
      "loss": 2.7173,
      "theoretical_loss": 3.5146090371164505,
      "tokens_seen": 1507142656
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002744433299899699,
      "loss": 2.6439,
      "theoretical_loss": 3.5145955750107585,
      "tokens_seen": 1507208192
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027443329989969913,
      "loss": 2.534,
      "theoretical_loss": 3.5145821136543,
      "tokens_seen": 1507273728
    },
    {
      "epoch": 5.01,
      "objective/train/docs_used": 1733768,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.598609209060669,
      "objective/train/theoretical_loss": 3.5145720181285967,
      "objective/train/tokens_used": 1527782880,
      "theoretical_loss": 3.5145720181285967,
      "tokens_seen": 1507322880
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027442326980942826,
      "loss": 2.6034,
      "theoretical_loss": 3.514568653047002,
      "tokens_seen": 1507339264
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002744132397191575,
      "loss": 2.5776,
      "theoretical_loss": 3.514555193188789,
      "tokens_seen": 1507404800
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027440320962888667,
      "loss": 2.5075,
      "theoretical_loss": 3.5145417340795877,
      "tokens_seen": 1507470336
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027439317953861585,
      "loss": 2.6137,
      "theoretical_loss": 3.5145282757193232,
      "tokens_seen": 1507535872
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027438314944834503,
      "loss": 2.6294,
      "theoretical_loss": 3.514514818107922,
      "tokens_seen": 1507601408
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002743731193580742,
      "loss": 2.5912,
      "theoretical_loss": 3.5145013612453093,
      "tokens_seen": 1507666944
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002743630892678034,
      "loss": 2.4158,
      "theoretical_loss": 3.514487905131411,
      "tokens_seen": 1507732480
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027435305917753263,
      "loss": 2.4401,
      "theoretical_loss": 3.514474449766153,
      "tokens_seen": 1507798016
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027434302908726176,
      "loss": 2.6716,
      "theoretical_loss": 3.514460995149461,
      "tokens_seen": 1507863552
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.000274332998996991,
      "loss": 2.51,
      "theoretical_loss": 3.5144475412812612,
      "tokens_seen": 1507929088
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002743229689067202,
      "loss": 2.5596,
      "theoretical_loss": 3.514434088161479,
      "tokens_seen": 1507994624
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027431293881644936,
      "loss": 2.5836,
      "theoretical_loss": 3.5144206357900405,
      "tokens_seen": 1508060160
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027430290872617854,
      "loss": 2.562,
      "theoretical_loss": 3.5144071841668714,
      "tokens_seen": 1508125696
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002742928786359077,
      "loss": 2.6443,
      "theoretical_loss": 3.514393733291897,
      "tokens_seen": 1508191232
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002742828485456369,
      "loss": 2.6106,
      "theoretical_loss": 3.5143802831650444,
      "tokens_seen": 1508256768
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027427281845536614,
      "loss": 2.4577,
      "theoretical_loss": 3.5143668337862386,
      "tokens_seen": 1508322304
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027426278836509526,
      "loss": 2.6082,
      "theoretical_loss": 3.514353385155406,
      "tokens_seen": 1508387840
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002742527582748245,
      "loss": 2.6183,
      "theoretical_loss": 3.514339937272472,
      "tokens_seen": 1508453376
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002742427281845536,
      "loss": 2.7128,
      "theoretical_loss": 3.514326490137363,
      "tokens_seen": 1508518912
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027423269809428286,
      "loss": 2.7367,
      "theoretical_loss": 3.514313043750005,
      "tokens_seen": 1508584448
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027422266800401204,
      "loss": 2.6525,
      "theoretical_loss": 3.514299598110323,
      "tokens_seen": 1508649984
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002742126379137412,
      "loss": 2.6169,
      "theoretical_loss": 3.5142861532182437,
      "tokens_seen": 1508715520
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002742026078234704,
      "loss": 2.7413,
      "theoretical_loss": 3.514272709073693,
      "tokens_seen": 1508781056
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002741925777331996,
      "loss": 2.5877,
      "theoretical_loss": 3.514259265676597,
      "tokens_seen": 1508846592
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027418254764292877,
      "loss": 2.5007,
      "theoretical_loss": 3.5142458230268816,
      "tokens_seen": 1508912128
    },
    {
      "epoch": 5.01,
      "objective/train/docs_used": 1734888,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6094367504119873,
      "objective/train/theoretical_loss": 3.514235741530019,
      "objective/train/tokens_used": 1529421280,
      "theoretical_loss": 3.514235741530019,
      "tokens_seen": 1508961280
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.000274172517552658,
      "loss": 2.504,
      "theoretical_loss": 3.5142323811244722,
      "tokens_seen": 1508977664
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027416248746238713,
      "loss": 2.7591,
      "theoretical_loss": 3.514218939969296,
      "tokens_seen": 1509043200
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027415245737211636,
      "loss": 2.4556,
      "theoretical_loss": 3.514205499561278,
      "tokens_seen": 1509108736
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027414242728184554,
      "loss": 2.7308,
      "theoretical_loss": 3.5141920599003447,
      "tokens_seen": 1509174272
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002741323971915747,
      "loss": 2.7473,
      "theoretical_loss": 3.5141786209864216,
      "tokens_seen": 1509239808
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002741223671013039,
      "loss": 2.4884,
      "theoretical_loss": 3.5141651828194354,
      "tokens_seen": 1509305344
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002741123370110331,
      "loss": 2.702,
      "theoretical_loss": 3.514151745399312,
      "tokens_seen": 1509370880
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027410230692076227,
      "loss": 2.7508,
      "theoretical_loss": 3.514138308725977,
      "tokens_seen": 1509436416
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002740922768304915,
      "loss": 2.5619,
      "theoretical_loss": 3.514124872799357,
      "tokens_seen": 1509501952
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027408224674022063,
      "loss": 2.8769,
      "theoretical_loss": 3.514111437619378,
      "tokens_seen": 1509567488
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027407221664994987,
      "loss": 2.7444,
      "theoretical_loss": 3.514098003185966,
      "tokens_seen": 1509633024
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.000274062186559679,
      "loss": 2.481,
      "theoretical_loss": 3.5140845694990466,
      "tokens_seen": 1509698560
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027405215646940823,
      "loss": 2.6672,
      "theoretical_loss": 3.5140711365585466,
      "tokens_seen": 1509764096
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002740421263791374,
      "loss": 2.6577,
      "theoretical_loss": 3.5140577043643924,
      "tokens_seen": 1509829632
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002740320962888666,
      "loss": 2.5711,
      "theoretical_loss": 3.514044272916509,
      "tokens_seen": 1509895168
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002740220661985958,
      "loss": 2.5343,
      "theoretical_loss": 3.514030842214824,
      "tokens_seen": 1509960704
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027401203610832495,
      "loss": 2.4917,
      "theoretical_loss": 3.5140174122592627,
      "tokens_seen": 1510026240
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002740020060180542,
      "loss": 2.7321,
      "theoretical_loss": 3.5140039830497507,
      "tokens_seen": 1510091776
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027399197592778337,
      "loss": 2.5898,
      "theoretical_loss": 3.513990554586216,
      "tokens_seen": 1510157312
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027398194583751255,
      "loss": 2.7621,
      "theoretical_loss": 3.513977126868583,
      "tokens_seen": 1510222848
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027397191574724173,
      "loss": 2.4087,
      "theoretical_loss": 3.5139636998967783,
      "tokens_seen": 1510288384
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027396188565697097,
      "loss": 2.6989,
      "theoretical_loss": 3.513950273670728,
      "tokens_seen": 1510353920
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002739518555667001,
      "loss": 2.6959,
      "theoretical_loss": 3.51393684819036,
      "tokens_seen": 1510419456
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027394182547642933,
      "loss": 2.5562,
      "theoretical_loss": 3.513923423455598,
      "tokens_seen": 1510484992
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027393179538615846,
      "loss": 2.4551,
      "theoretical_loss": 3.51390999946637,
      "tokens_seen": 1510550528
    },
    {
      "epoch": 5.01,
      "objective/train/docs_used": 1735676,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.696056842803955,
      "objective/train/theoretical_loss": 3.513899931963661,
      "objective/train/tokens_used": 1531059680,
      "theoretical_loss": 3.513899931963661,
      "tokens_seen": 1510599680
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002739217652958877,
      "loss": 2.5213,
      "theoretical_loss": 3.5138965762226015,
      "tokens_seen": 1510616064
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027391173520561687,
      "loss": 2.6409,
      "theoretical_loss": 3.513883153724219,
      "tokens_seen": 1510681600
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027390170511534605,
      "loss": 2.7535,
      "theoretical_loss": 3.513869731971149,
      "tokens_seen": 1510747136
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027389167502507524,
      "loss": 2.6657,
      "theoretical_loss": 3.5138563109633174,
      "tokens_seen": 1510812672
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002738816449348044,
      "loss": 2.759,
      "theoretical_loss": 3.5138428907006505,
      "tokens_seen": 1510878208
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002738716148445336,
      "loss": 2.4819,
      "theoretical_loss": 3.513829471183075,
      "tokens_seen": 1510943744
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027386158475426283,
      "loss": 2.7243,
      "theoretical_loss": 3.513816052410517,
      "tokens_seen": 1511009280
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027385155466399196,
      "loss": 2.7477,
      "theoretical_loss": 3.5138026343829027,
      "tokens_seen": 1511074816
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002738415245737212,
      "loss": 2.6833,
      "theoretical_loss": 3.5137892171001583,
      "tokens_seen": 1511140352
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002738314944834504,
      "loss": 2.567,
      "theoretical_loss": 3.5137758005622106,
      "tokens_seen": 1511205888
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027382146439317956,
      "loss": 2.689,
      "theoretical_loss": 3.513762384768986,
      "tokens_seen": 1511271424
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027381143430290874,
      "loss": 2.7233,
      "theoretical_loss": 3.5137489697204103,
      "tokens_seen": 1511336960
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002738014042126379,
      "loss": 2.6076,
      "theoretical_loss": 3.5137355554164103,
      "tokens_seen": 1511402496
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002737913741223671,
      "loss": 2.8242,
      "theoretical_loss": 3.513722141856912,
      "tokens_seen": 1511468032
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027378134403209634,
      "loss": 2.7138,
      "theoretical_loss": 3.5137087290418423,
      "tokens_seen": 1511533568
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027377131394182546,
      "loss": 2.6239,
      "theoretical_loss": 3.5136953169711274,
      "tokens_seen": 1511599104
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002737612838515547,
      "loss": 2.5643,
      "theoretical_loss": 3.5136819056446935,
      "tokens_seen": 1511664640
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002737512537612838,
      "loss": 2.8376,
      "theoretical_loss": 3.5136684950624675,
      "tokens_seen": 1511730176
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027374122367101306,
      "loss": 2.4982,
      "theoretical_loss": 3.5136550852243755,
      "tokens_seen": 1511795712
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027373119358074224,
      "loss": 2.5081,
      "theoretical_loss": 3.5136416761303444,
      "tokens_seen": 1511861248
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002737211634904714,
      "loss": 2.5411,
      "theoretical_loss": 3.5136282677803,
      "tokens_seen": 1511926784
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002737111334002006,
      "loss": 2.689,
      "theoretical_loss": 3.513614860174169,
      "tokens_seen": 1511992320
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002737011033099298,
      "loss": 2.6541,
      "theoretical_loss": 3.513601453311878,
      "tokens_seen": 1512057856
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027369107321965897,
      "loss": 2.6243,
      "theoretical_loss": 3.5135880471933536,
      "tokens_seen": 1512123392
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002736810431293882,
      "loss": 2.6046,
      "theoretical_loss": 3.513574641818522,
      "tokens_seen": 1512188928
    },
    {
      "epoch": 5.01,
      "objective/train/docs_used": 1736838,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.042192220687866,
      "objective/train/theoretical_loss": 3.5135645882754023,
      "objective/train/tokens_used": 1532698080,
      "theoretical_loss": 3.5135645882754023,
      "tokens_seen": 1512238080
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027367101303911733,
      "loss": 2.7713,
      "theoretical_loss": 3.51356123718731,
      "tokens_seen": 1512254464
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027366098294884656,
      "loss": 2.7605,
      "theoretical_loss": 3.513547833299644,
      "tokens_seen": 1512320000
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027365095285857574,
      "loss": 2.6,
      "theoretical_loss": 3.5135344301554503,
      "tokens_seen": 1512385536
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002736409227683049,
      "loss": 2.5412,
      "theoretical_loss": 3.5135210277546562,
      "tokens_seen": 1512451072
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002736308926780341,
      "loss": 2.7291,
      "theoretical_loss": 3.513507626097187,
      "tokens_seen": 1512516608
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002736208625877633,
      "loss": 2.4899,
      "theoretical_loss": 3.5134942251829706,
      "tokens_seen": 1512582144
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027361083249749247,
      "loss": 2.5796,
      "theoretical_loss": 3.513480825011933,
      "tokens_seen": 1512647680
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002736008024072217,
      "loss": 2.5228,
      "theoretical_loss": 3.513467425584001,
      "tokens_seen": 1512713216
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027359077231695083,
      "loss": 2.7755,
      "theoretical_loss": 3.5134540268991006,
      "tokens_seen": 1512778752
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027358074222668007,
      "loss": 2.5736,
      "theoretical_loss": 3.5134406289571585,
      "tokens_seen": 1512844288
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002735707121364092,
      "loss": 2.7156,
      "theoretical_loss": 3.5134272317581026,
      "tokens_seen": 1512909824
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027356068204613843,
      "loss": 2.702,
      "theoretical_loss": 3.513413835301858,
      "tokens_seen": 1512975360
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002735506519558676,
      "loss": 2.7033,
      "theoretical_loss": 3.513400439588352,
      "tokens_seen": 1513040896
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002735406218655968,
      "loss": 2.6876,
      "theoretical_loss": 3.513387044617511,
      "tokens_seen": 1513106432
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027353059177532597,
      "loss": 2.7163,
      "theoretical_loss": 3.5133736503892616,
      "tokens_seen": 1513171968
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027352056168505515,
      "loss": 2.7879,
      "theoretical_loss": 3.513360256903531,
      "tokens_seen": 1513237504
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027351053159478433,
      "loss": 2.6548,
      "theoretical_loss": 3.5133468641602454,
      "tokens_seen": 1513303040
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027350050150451357,
      "loss": 2.7286,
      "theoretical_loss": 3.513333472159332,
      "tokens_seen": 1513368576
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002734904714142427,
      "loss": 2.7595,
      "theoretical_loss": 3.5133200809007166,
      "tokens_seen": 1513434112
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027348044132397193,
      "loss": 2.5187,
      "theoretical_loss": 3.513306690384327,
      "tokens_seen": 1513499648
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002734704112337011,
      "loss": 2.6689,
      "theoretical_loss": 3.5132933006100893,
      "tokens_seen": 1513565184
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002734603811434303,
      "loss": 2.5274,
      "theoretical_loss": 3.5132799115779303,
      "tokens_seen": 1513630720
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002734503510531595,
      "loss": 2.8102,
      "theoretical_loss": 3.5132665232877764,
      "tokens_seen": 1513696256
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027344032096288866,
      "loss": 2.6089,
      "theoretical_loss": 3.5132531357395553,
      "tokens_seen": 1513761792
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027343029087261784,
      "loss": 2.6788,
      "theoretical_loss": 3.5132397489331932,
      "tokens_seen": 1513827328
    },
    {
      "epoch": 5.01,
      "objective/train/docs_used": 1737452,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.892695426940918,
      "objective/train/theoretical_loss": 3.5132297093152225,
      "objective/train/tokens_used": 1534336480,
      "theoretical_loss": 3.5132297093152225,
      "tokens_seen": 1513876480
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002734202607823471,
      "loss": 2.5524,
      "theoretical_loss": 3.5132263628686164,
      "tokens_seen": 1513892864
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002734102306920762,
      "loss": 2.61,
      "theoretical_loss": 3.513212977545753,
      "tokens_seen": 1513958400
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027340020060180544,
      "loss": 2.6958,
      "theoretical_loss": 3.5131995929645283,
      "tokens_seen": 1514023936
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027339017051153456,
      "loss": 2.6809,
      "theoretical_loss": 3.5131862091248696,
      "tokens_seen": 1514089472
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002733801404212638,
      "loss": 2.7049,
      "theoretical_loss": 3.5131728260267048,
      "tokens_seen": 1514155008
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.000273370110330993,
      "loss": 2.6297,
      "theoretical_loss": 3.513159443669959,
      "tokens_seen": 1514220544
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027336008024072216,
      "loss": 2.7736,
      "theoretical_loss": 3.51314606205456,
      "tokens_seen": 1514286080
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027335005015045134,
      "loss": 2.6571,
      "theoretical_loss": 3.5131326811804344,
      "tokens_seen": 1514351616
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002733400200601806,
      "loss": 2.559,
      "theoretical_loss": 3.5131193010475097,
      "tokens_seen": 1514417152
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002733299899699097,
      "loss": 2.8363,
      "theoretical_loss": 3.513105921655712,
      "tokens_seen": 1514482688
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027331995987963894,
      "loss": 2.7014,
      "theoretical_loss": 3.5130925430049684,
      "tokens_seen": 1514548224
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027330992978936807,
      "loss": 2.586,
      "theoretical_loss": 3.513079165095206,
      "tokens_seen": 1514613760
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002732998996990973,
      "loss": 2.4239,
      "theoretical_loss": 3.513065787926351,
      "tokens_seen": 1514679296
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002732898696088265,
      "loss": 2.4609,
      "theoretical_loss": 3.5130524114983315,
      "tokens_seen": 1514744832
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027327983951855566,
      "loss": 2.5722,
      "theoretical_loss": 3.5130390358110732,
      "tokens_seen": 1514810368
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002732698094282849,
      "loss": 2.5675,
      "theoretical_loss": 3.513025660864504,
      "tokens_seen": 1514875904
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.000273259779338014,
      "loss": 2.4711,
      "theoretical_loss": 3.51301228665855,
      "tokens_seen": 1514941440
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027324974924774326,
      "loss": 2.494,
      "theoretical_loss": 3.5129989131931394,
      "tokens_seen": 1515006976
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027323971915747244,
      "loss": 2.4075,
      "theoretical_loss": 3.5129855404681978,
      "tokens_seen": 1515072512
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002732296890672016,
      "loss": 2.7047,
      "theoretical_loss": 3.5129721684836523,
      "tokens_seen": 1515138048
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002732196589769308,
      "loss": 2.5743,
      "theoretical_loss": 3.5129587972394307,
      "tokens_seen": 1515203584
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027320962888666,
      "loss": 2.5521,
      "theoretical_loss": 3.5129454267354596,
      "tokens_seen": 1515269120
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027319959879638917,
      "loss": 2.7014,
      "theoretical_loss": 3.512932056971666,
      "tokens_seen": 1515334656
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002731895687061184,
      "loss": 2.6108,
      "theoretical_loss": 3.5129186879479772,
      "tokens_seen": 1515400192
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027317953861584753,
      "loss": 2.618,
      "theoretical_loss": 3.51290531966432,
      "tokens_seen": 1515465728
    },
    {
      "epoch": 5.01,
      "objective/train/docs_used": 1739078,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.715820789337158,
      "objective/train/theoretical_loss": 3.5128952939371785,
      "objective/train/tokens_used": 1535974880,
      "theoretical_loss": 3.5128952939371785,
      "tokens_seen": 1515514880
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027316950852557676,
      "loss": 2.6964,
      "theoretical_loss": 3.5128919521206208,
      "tokens_seen": 1515531264
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027315947843530594,
      "loss": 2.8865,
      "theoretical_loss": 3.512878585316807,
      "tokens_seen": 1515596800
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002731494483450351,
      "loss": 2.5913,
      "theoretical_loss": 3.5128652192528067,
      "tokens_seen": 1515662336
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002731394182547643,
      "loss": 2.4339,
      "theoretical_loss": 3.512851853928546,
      "tokens_seen": 1515727872
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002731293881644935,
      "loss": 2.5569,
      "theoretical_loss": 3.5128384893439524,
      "tokens_seen": 1515793408
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027311935807422267,
      "loss": 2.6283,
      "theoretical_loss": 3.512825125498952,
      "tokens_seen": 1515858944
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002731093279839519,
      "loss": 2.6183,
      "theoretical_loss": 3.512811762393473,
      "tokens_seen": 1515924480
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027309929789368103,
      "loss": 2.5518,
      "theoretical_loss": 3.512798400027442,
      "tokens_seen": 1515990016
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027308926780341027,
      "loss": 2.6761,
      "theoretical_loss": 3.5127850384007866,
      "tokens_seen": 1516055552
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002730792377131394,
      "loss": 2.5924,
      "theoretical_loss": 3.5127716775134328,
      "tokens_seen": 1516121088
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027306920762286863,
      "loss": 2.7086,
      "theoretical_loss": 3.5127583173653094,
      "tokens_seen": 1516186624
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002730591775325978,
      "loss": 2.7092,
      "theoretical_loss": 3.5127449579563423,
      "tokens_seen": 1516252160
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.000273049147442327,
      "loss": 2.6952,
      "theoretical_loss": 3.5127315992864587,
      "tokens_seen": 1516317696
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027303911735205617,
      "loss": 2.5468,
      "theoretical_loss": 3.5127182413555866,
      "tokens_seen": 1516383232
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027302908726178535,
      "loss": 2.581,
      "theoretical_loss": 3.5127048841636523,
      "tokens_seen": 1516448768
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027301905717151453,
      "loss": 2.3842,
      "theoretical_loss": 3.512691527710584,
      "tokens_seen": 1516514304
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027300902708124377,
      "loss": 2.5223,
      "theoretical_loss": 3.512678171996307,
      "tokens_seen": 1516579840
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002729989969909729,
      "loss": 2.5641,
      "theoretical_loss": 3.5126648170207515,
      "tokens_seen": 1516645376
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027298896690070213,
      "loss": 2.8428,
      "theoretical_loss": 3.512651462783842,
      "tokens_seen": 1516710912
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002729789368104313,
      "loss": 2.4741,
      "theoretical_loss": 3.512638109285507,
      "tokens_seen": 1516776448
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002729689067201605,
      "loss": 2.6423,
      "theoretical_loss": 3.512624756525673,
      "tokens_seen": 1516841984
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002729588766298897,
      "loss": 2.5588,
      "theoretical_loss": 3.512611404504268,
      "tokens_seen": 1516907520
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027294884653961886,
      "loss": 2.7271,
      "theoretical_loss": 3.512598053221219,
      "tokens_seen": 1516973056
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027293881644934804,
      "loss": 2.5997,
      "theoretical_loss": 3.5125847026764534,
      "tokens_seen": 1517038592
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002729287863590773,
      "loss": 2.5554,
      "theoretical_loss": 3.5125713528698985,
      "tokens_seen": 1517104128
    },
    {
      "epoch": 5.01,
      "objective/train/docs_used": 1739851,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.781825542449951,
      "objective/train/theoretical_loss": 3.5125613409993894,
      "objective/train/tokens_used": 1537613280,
      "theoretical_loss": 3.5125613409993894,
      "tokens_seen": 1517153280
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002729187562688064,
      "loss": 2.5735,
      "theoretical_loss": 3.512558003801481,
      "tokens_seen": 1517169664
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027290872617853564,
      "loss": 2.8424,
      "theoretical_loss": 3.512544655471129,
      "tokens_seen": 1517235200
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027289869608826476,
      "loss": 2.7819,
      "theoretical_loss": 3.5125313078787697,
      "tokens_seen": 1517300736
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.000272888665997994,
      "loss": 2.8765,
      "theoretical_loss": 3.51251796102433,
      "tokens_seen": 1517366272
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002728786359077232,
      "loss": 2.6052,
      "theoretical_loss": 3.512504614907737,
      "tokens_seen": 1517431808
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027286860581745236,
      "loss": 2.6188,
      "theoretical_loss": 3.512491269528919,
      "tokens_seen": 1517497344
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027285857572718154,
      "loss": 2.6225,
      "theoretical_loss": 3.5124779248878024,
      "tokens_seen": 1517562880
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002728485456369108,
      "loss": 2.6682,
      "theoretical_loss": 3.5124645809843154,
      "tokens_seen": 1517628416
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002728385155466399,
      "loss": 2.7252,
      "theoretical_loss": 3.512451237818385,
      "tokens_seen": 1517693952
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027282848545636914,
      "loss": 2.5136,
      "theoretical_loss": 3.512437895389938,
      "tokens_seen": 1517759488
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027281845536609827,
      "loss": 2.5185,
      "theoretical_loss": 3.512424553698903,
      "tokens_seen": 1517825024
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002728084252758275,
      "loss": 2.5597,
      "theoretical_loss": 3.5124112127452065,
      "tokens_seen": 1517890560
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002727983951855567,
      "loss": 2.5773,
      "theoretical_loss": 3.5123978725287763,
      "tokens_seen": 1517956096
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027278836509528586,
      "loss": 2.4945,
      "theoretical_loss": 3.5123845330495396,
      "tokens_seen": 1518021632
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027277833500501504,
      "loss": 2.591,
      "theoretical_loss": 3.5123711943074234,
      "tokens_seen": 1518087168
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002727683049147442,
      "loss": 2.6178,
      "theoretical_loss": 3.5123578563023568,
      "tokens_seen": 1518152704
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002727582748244734,
      "loss": 2.6907,
      "theoretical_loss": 3.5123445190342655,
      "tokens_seen": 1518218240
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027274824473420264,
      "loss": 2.376,
      "theoretical_loss": 3.5123311825030776,
      "tokens_seen": 1518283776
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027273821464393177,
      "loss": 2.5858,
      "theoretical_loss": 3.512317846708721,
      "tokens_seen": 1518349312
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.000272728184553661,
      "loss": 2.5271,
      "theoretical_loss": 3.5123045116511227,
      "tokens_seen": 1518414848
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027271815446339013,
      "loss": 2.6194,
      "theoretical_loss": 3.51229117733021,
      "tokens_seen": 1518480384
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027270812437311937,
      "loss": 2.5939,
      "theoretical_loss": 3.5122778437459106,
      "tokens_seen": 1518545920
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027269809428284855,
      "loss": 2.6146,
      "theoretical_loss": 3.5122645108981523,
      "tokens_seen": 1518611456
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027268806419257773,
      "loss": 2.6375,
      "theoretical_loss": 3.5122511787868627,
      "tokens_seen": 1518676992
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002726780341023069,
      "loss": 2.6602,
      "theoretical_loss": 3.512237847411969,
      "tokens_seen": 1518742528
    },
    {
      "epoch": 5.01,
      "objective/train/docs_used": 1741412,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7110002040863037,
      "objective/train/theoretical_loss": 3.512227849364015,
      "objective/train/tokens_used": 1539251680,
      "theoretical_loss": 3.512227849364015,
      "tokens_seen": 1518791680
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027266800401203614,
      "loss": 2.5932,
      "theoretical_loss": 3.512224516773399,
      "tokens_seen": 1518808064
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027265797392176527,
      "loss": 2.7002,
      "theoretical_loss": 3.51221118687108,
      "tokens_seen": 1518873600
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002726479438314945,
      "loss": 2.6501,
      "theoretical_loss": 3.5121978577049395,
      "tokens_seen": 1518939136
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027263791374122363,
      "loss": 2.6066,
      "theoretical_loss": 3.512184529274905,
      "tokens_seen": 1519004672
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027262788365095287,
      "loss": 2.6338,
      "theoretical_loss": 3.5121712015809052,
      "tokens_seen": 1519070208
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027261785356068205,
      "loss": 2.7052,
      "theoretical_loss": 3.5121578746228668,
      "tokens_seen": 1519135744
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027260782347041123,
      "loss": 2.6389,
      "theoretical_loss": 3.512144548400717,
      "tokens_seen": 1519201280
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002725977933801404,
      "loss": 2.7032,
      "theoretical_loss": 3.512131222914384,
      "tokens_seen": 1519266816
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002725877632898696,
      "loss": 2.6396,
      "theoretical_loss": 3.512117898163795,
      "tokens_seen": 1519332352
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002725777331995988,
      "loss": 2.6893,
      "theoretical_loss": 3.5121045741488786,
      "tokens_seen": 1519397888
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.000272567703109328,
      "loss": 2.7178,
      "theoretical_loss": 3.5120912508695614,
      "tokens_seen": 1519463424
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027255767301905714,
      "loss": 2.8325,
      "theoretical_loss": 3.5120779283257715,
      "tokens_seen": 1519528960
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027254764292878637,
      "loss": 2.5731,
      "theoretical_loss": 3.512064606517437,
      "tokens_seen": 1519594496
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002725376128385155,
      "loss": 2.5434,
      "theoretical_loss": 3.5120512854444845,
      "tokens_seen": 1519660032
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027252758274824473,
      "loss": 2.5804,
      "theoretical_loss": 3.512037965106843,
      "tokens_seen": 1519725568
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027251755265797397,
      "loss": 2.5707,
      "theoretical_loss": 3.5120246455044395,
      "tokens_seen": 1519791104
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002725075225677031,
      "loss": 2.6185,
      "theoretical_loss": 3.512011326637201,
      "tokens_seen": 1519856640
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027249749247743233,
      "loss": 2.5685,
      "theoretical_loss": 3.5119980085050573,
      "tokens_seen": 1519922176
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002724874623871615,
      "loss": 2.6181,
      "theoretical_loss": 3.5119846911079335,
      "tokens_seen": 1519987712
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002724774322968907,
      "loss": 2.5499,
      "theoretical_loss": 3.5119713744457597,
      "tokens_seen": 1520053248
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002724674022066199,
      "loss": 2.4947,
      "theoretical_loss": 3.511958058518462,
      "tokens_seen": 1520118784
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027245737211634906,
      "loss": 2.6411,
      "theoretical_loss": 3.511944743325969,
      "tokens_seen": 1520184320
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027244734202607824,
      "loss": 2.5907,
      "theoretical_loss": 3.511931428868208,
      "tokens_seen": 1520249856
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002724373119358075,
      "loss": 2.5607,
      "theoretical_loss": 3.5119181151451078,
      "tokens_seen": 1520315392
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002724272818455366,
      "loss": 2.4661,
      "theoretical_loss": 3.511904802156595,
      "tokens_seen": 1520380928
    },
    {
      "epoch": 5.01,
      "objective/train/docs_used": 1741916,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7754528522491455,
      "objective/train/theoretical_loss": 3.51189481789724,
      "objective/train/tokens_used": 1540890080,
      "theoretical_loss": 3.51189481789724,
      "tokens_seen": 1520430080
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027241725175526584,
      "loss": 2.504,
      "theoretical_loss": 3.511891489902598,
      "tokens_seen": 1520446464
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027240722166499496,
      "loss": 2.7549,
      "theoretical_loss": 3.511878178383044,
      "tokens_seen": 1520512000
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002723971915747242,
      "loss": 2.7554,
      "theoretical_loss": 3.511864867597862,
      "tokens_seen": 1520577536
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002723871614844534,
      "loss": 2.7106,
      "theoretical_loss": 3.5118515575469784,
      "tokens_seen": 1520643072
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027237713139418256,
      "loss": 2.6407,
      "theoretical_loss": 3.5118382482303225,
      "tokens_seen": 1520708608
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027236710130391174,
      "loss": 2.5378,
      "theoretical_loss": 3.511824939647821,
      "tokens_seen": 1520774144
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.000272357071213641,
      "loss": 2.7233,
      "theoretical_loss": 3.5118116317994024,
      "tokens_seen": 1520839680
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002723470411233701,
      "loss": 2.9172,
      "theoretical_loss": 3.5117983246849946,
      "tokens_seen": 1520905216
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027233701103309934,
      "loss": 2.4646,
      "theoretical_loss": 3.5117850183045247,
      "tokens_seen": 1520970752
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027232698094282847,
      "loss": 2.6007,
      "theoretical_loss": 3.511771712657922,
      "tokens_seen": 1521036288
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002723169508525577,
      "loss": 2.4467,
      "theoretical_loss": 3.5117584077451127,
      "tokens_seen": 1521101824
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002723069207622869,
      "loss": 2.5762,
      "theoretical_loss": 3.5117451035660263,
      "tokens_seen": 1521167360
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027229689067201606,
      "loss": 2.7585,
      "theoretical_loss": 3.51173180012059,
      "tokens_seen": 1521232896
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027228686058174524,
      "loss": 2.8096,
      "theoretical_loss": 3.5117184974087317,
      "tokens_seen": 1521298432
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002722768304914744,
      "loss": 2.6325,
      "theoretical_loss": 3.5117051954303795,
      "tokens_seen": 1521363968
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002722668004012036,
      "loss": 2.6678,
      "theoretical_loss": 3.511691894185461,
      "tokens_seen": 1521429504
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027225677031093284,
      "loss": 2.7784,
      "theoretical_loss": 3.5116785936739046,
      "tokens_seen": 1521495040
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027224674022066197,
      "loss": 2.8446,
      "theoretical_loss": 3.5116652938956383,
      "tokens_seen": 1521560576
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002722367101303912,
      "loss": 2.9045,
      "theoretical_loss": 3.51165199485059,
      "tokens_seen": 1521626112
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027222668004012033,
      "loss": 2.5861,
      "theoretical_loss": 3.511638696538687,
      "tokens_seen": 1521691648
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027221664994984957,
      "loss": 2.6447,
      "theoretical_loss": 3.5116253989598585,
      "tokens_seen": 1521757184
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027220661985957875,
      "loss": 2.5343,
      "theoretical_loss": 3.5116121021140323,
      "tokens_seen": 1521822720
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027219658976930793,
      "loss": 2.7088,
      "theoretical_loss": 3.5115988060011354,
      "tokens_seen": 1521888256
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002721865596790371,
      "loss": 2.4863,
      "theoretical_loss": 3.5115855106210967,
      "tokens_seen": 1521953792
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027217652958876635,
      "loss": 2.5216,
      "theoretical_loss": 3.5115722159738443,
      "tokens_seen": 1522019328
    },
    {
      "epoch": 5.01,
      "objective/train/docs_used": 1743369,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8488197326660156,
      "objective/train/theoretical_loss": 3.5115622454692526,
      "objective/train/tokens_used": 1542528480,
      "theoretical_loss": 3.5115622454692526,
      "tokens_seen": 1522068480
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027216649949849547,
      "loss": 2.6903,
      "theoretical_loss": 3.511558922059306,
      "tokens_seen": 1522084864
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002721564694082247,
      "loss": 2.5928,
      "theoretical_loss": 3.51154562887741,
      "tokens_seen": 1522150400
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027214643931795383,
      "loss": 2.532,
      "theoretical_loss": 3.5115323364280844,
      "tokens_seen": 1522215936
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027213640922768307,
      "loss": 2.502,
      "theoretical_loss": 3.5115190447112568,
      "tokens_seen": 1522281472
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027212637913741225,
      "loss": 2.5598,
      "theoretical_loss": 3.511505753726856,
      "tokens_seen": 1522347008
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027211634904714143,
      "loss": 2.5905,
      "theoretical_loss": 3.51149246347481,
      "tokens_seen": 1522412544
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002721063189568706,
      "loss": 2.7652,
      "theoretical_loss": 3.5114791739550464,
      "tokens_seen": 1522478080
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002720962888665998,
      "loss": 2.6179,
      "theoretical_loss": 3.511465885167494,
      "tokens_seen": 1522543616
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.000272086258776329,
      "loss": 2.6407,
      "theoretical_loss": 3.51145259711208,
      "tokens_seen": 1522609152
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002720762286860582,
      "loss": 2.5974,
      "theoretical_loss": 3.511439309788734,
      "tokens_seen": 1522674688
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027206619859578734,
      "loss": 2.6676,
      "theoretical_loss": 3.511426023197383,
      "tokens_seen": 1522740224
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027205616850551657,
      "loss": 2.4657,
      "theoretical_loss": 3.511412737337955,
      "tokens_seen": 1522805760
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002720461384152457,
      "loss": 2.5989,
      "theoretical_loss": 3.5113994522103793,
      "tokens_seen": 1522871296
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027203610832497493,
      "loss": 2.4463,
      "theoretical_loss": 3.5113861678145835,
      "tokens_seen": 1522936832
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002720260782347041,
      "loss": 2.5057,
      "theoretical_loss": 3.511372884150495,
      "tokens_seen": 1523002368
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002720160481444333,
      "loss": 2.6699,
      "theoretical_loss": 3.511359601218044,
      "tokens_seen": 1523067904
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002720060180541625,
      "loss": 2.723,
      "theoretical_loss": 3.5113463190171568,
      "tokens_seen": 1523133440
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002719959879638917,
      "loss": 2.7831,
      "theoretical_loss": 3.5113330375477623,
      "tokens_seen": 1523198976
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027198595787362084,
      "loss": 2.7578,
      "theoretical_loss": 3.5113197568097894,
      "tokens_seen": 1523264512
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002719759277833501,
      "loss": 2.4327,
      "theoretical_loss": 3.511306476803165,
      "tokens_seen": 1523330048
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002719658976930792,
      "loss": 2.6328,
      "theoretical_loss": 3.5112931975278183,
      "tokens_seen": 1523395584
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027195586760280844,
      "loss": 2.3377,
      "theoretical_loss": 3.5112799189836776,
      "tokens_seen": 1523461120
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002719458375125376,
      "loss": 2.5108,
      "theoretical_loss": 3.511266641170671,
      "tokens_seen": 1523526656
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002719358074222668,
      "loss": 2.4412,
      "theoretical_loss": 3.511253364088727,
      "tokens_seen": 1523592192
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.000271925777331996,
      "loss": 2.7561,
      "theoretical_loss": 3.5112400877377734,
      "tokens_seen": 1523657728
    },
    {
      "epoch": 5.01,
      "objective/train/docs_used": 1744115,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8299615383148193,
      "objective/train/theoretical_loss": 3.5112301309542273,
      "objective/train/tokens_used": 1544166880,
      "theoretical_loss": 3.5112301309542273,
      "tokens_seen": 1523706880
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027191574724172516,
      "loss": 2.7278,
      "theoretical_loss": 3.5112268121177386,
      "tokens_seen": 1523723264
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027190571715145434,
      "loss": 2.9239,
      "theoretical_loss": 3.511213537228551,
      "tokens_seen": 1523788800
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002718956870611836,
      "loss": 2.5451,
      "theoretical_loss": 3.5112002630701395,
      "tokens_seen": 1523854336
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002718856569709127,
      "loss": 2.5095,
      "theoretical_loss": 3.5111869896424324,
      "tokens_seen": 1523919872
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027187562688064194,
      "loss": 2.7109,
      "theoretical_loss": 3.5111737169453567,
      "tokens_seen": 1523985408
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027186559679037107,
      "loss": 2.753,
      "theoretical_loss": 3.5111604449788425,
      "tokens_seen": 1524050944
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002718555667001003,
      "loss": 2.6246,
      "theoretical_loss": 3.511147173742817,
      "tokens_seen": 1524116480
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002718455366098295,
      "loss": 2.4407,
      "theoretical_loss": 3.511133903237209,
      "tokens_seen": 1524182016
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027183550651955867,
      "loss": 2.674,
      "theoretical_loss": 3.511120633461947,
      "tokens_seen": 1524247552
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027182547642928785,
      "loss": 2.5761,
      "theoretical_loss": 3.5111073644169593,
      "tokens_seen": 1524313088
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002718154463390171,
      "loss": 2.6593,
      "theoretical_loss": 3.511094096102174,
      "tokens_seen": 1524378624
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002718054162487462,
      "loss": 2.5198,
      "theoretical_loss": 3.51108082851752,
      "tokens_seen": 1524444160
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027179538615847544,
      "loss": 2.6014,
      "theoretical_loss": 3.511067561662926,
      "tokens_seen": 1524509696
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027178535606820457,
      "loss": 2.7626,
      "theoretical_loss": 3.5110542955383197,
      "tokens_seen": 1524575232
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002717753259779338,
      "loss": 2.3878,
      "theoretical_loss": 3.51104103014363,
      "tokens_seen": 1524640768
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027176529588766304,
      "loss": 2.4231,
      "theoretical_loss": 3.5110277654787847,
      "tokens_seen": 1524706304
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027175526579739217,
      "loss": 2.6618,
      "theoretical_loss": 3.5110145015437135,
      "tokens_seen": 1524771840
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002717452357071214,
      "loss": 2.7225,
      "theoretical_loss": 3.511001238338344,
      "tokens_seen": 1524837376
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027173520561685053,
      "loss": 2.6813,
      "theoretical_loss": 3.5109879758626046,
      "tokens_seen": 1524902912
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027172517552657977,
      "loss": 2.6589,
      "theoretical_loss": 3.5109747141164243,
      "tokens_seen": 1524968448
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027171514543630895,
      "loss": 2.7244,
      "theoretical_loss": 3.5109614530997315,
      "tokens_seen": 1525033984
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027170511534603813,
      "loss": 2.5411,
      "theoretical_loss": 3.510948192812455,
      "tokens_seen": 1525099520
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002716950852557673,
      "loss": 2.6788,
      "theoretical_loss": 3.5109349332545223,
      "tokens_seen": 1525165056
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027168505516549655,
      "loss": 2.607,
      "theoretical_loss": 3.510921674425863,
      "tokens_seen": 1525230592
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027167502507522567,
      "loss": 2.4647,
      "theoretical_loss": 3.510908416326405,
      "tokens_seen": 1525296128
    },
    {
      "epoch": 5.01,
      "objective/train/docs_used": 1744854,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9327380657196045,
      "objective/train/theoretical_loss": 3.5108984732303075,
      "objective/train/tokens_used": 1545805280,
      "theoretical_loss": 3.5108984732303075,
      "tokens_seen": 1525345280
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002716649949849549,
      "loss": 2.6041,
      "theoretical_loss": 3.5108951589560773,
      "tokens_seen": 1525361664
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027165496489468403,
      "loss": 2.6702,
      "theoretical_loss": 3.5108819023148086,
      "tokens_seen": 1525427200
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027164493480441327,
      "loss": 2.725,
      "theoretical_loss": 3.5108686464025274,
      "tokens_seen": 1525492736
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027163490471414245,
      "loss": 2.749,
      "theoretical_loss": 3.5108553912191613,
      "tokens_seen": 1525558272
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027162487462387163,
      "loss": 2.7232,
      "theoretical_loss": 3.5108421367646403,
      "tokens_seen": 1525623808
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002716148445336008,
      "loss": 2.596,
      "theoretical_loss": 3.5108288830388923,
      "tokens_seen": 1525689344
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027160481444333,
      "loss": 2.5888,
      "theoretical_loss": 3.5108156300418463,
      "tokens_seen": 1525754880
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002715947843530592,
      "loss": 2.5965,
      "theoretical_loss": 3.5108023777734303,
      "tokens_seen": 1525820416
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002715847542627884,
      "loss": 2.5714,
      "theoretical_loss": 3.5107891262335738,
      "tokens_seen": 1525885952
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027157472417251754,
      "loss": 2.7426,
      "theoretical_loss": 3.5107758754222047,
      "tokens_seen": 1525951488
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002715646940822468,
      "loss": 2.4285,
      "theoretical_loss": 3.510762625339252,
      "tokens_seen": 1526017024
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002715546639919759,
      "loss": 2.4279,
      "theoretical_loss": 3.5107493759846444,
      "tokens_seen": 1526082560
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027154463390170514,
      "loss": 2.433,
      "theoretical_loss": 3.5107361273583106,
      "tokens_seen": 1526148096
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002715346038114343,
      "loss": 2.8518,
      "theoretical_loss": 3.5107228794601797,
      "tokens_seen": 1526213632
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002715245737211635,
      "loss": 2.5255,
      "theoretical_loss": 3.510709632290179,
      "tokens_seen": 1526279168
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002715145436308927,
      "loss": 2.759,
      "theoretical_loss": 3.510696385848239,
      "tokens_seen": 1526344704
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002715045135406219,
      "loss": 2.4789,
      "theoretical_loss": 3.5106831401342875,
      "tokens_seen": 1526410240
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027149448345035104,
      "loss": 2.3398,
      "theoretical_loss": 3.510669895148253,
      "tokens_seen": 1526475776
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002714844533600803,
      "loss": 2.5167,
      "theoretical_loss": 3.5106566508900645,
      "tokens_seen": 1526541312
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002714744232698094,
      "loss": 2.5327,
      "theoretical_loss": 3.510643407359651,
      "tokens_seen": 1526606848
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027146439317953864,
      "loss": 2.558,
      "theoretical_loss": 3.5106301645569413,
      "tokens_seen": 1526672384
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002714543630892678,
      "loss": 2.6568,
      "theoretical_loss": 3.510616922481864,
      "tokens_seen": 1526737920
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.000271444332998997,
      "loss": 2.6549,
      "theoretical_loss": 3.5106036811343477,
      "tokens_seen": 1526803456
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002714343029087262,
      "loss": 2.7285,
      "theoretical_loss": 3.5105904405143216,
      "tokens_seen": 1526868992
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027142427281845536,
      "loss": 2.4567,
      "theoretical_loss": 3.5105772006217144,
      "tokens_seen": 1526934528
    },
    {
      "epoch": 5.01,
      "objective/train/docs_used": 1746475,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7063658237457275,
      "objective/train/theoretical_loss": 3.5105672711795846,
      "objective/train/tokens_used": 1547443680,
      "theoretical_loss": 3.5105672711795846,
      "tokens_seen": 1526983680
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027141424272818454,
      "loss": 2.7586,
      "theoretical_loss": 3.5105639614564543,
      "tokens_seen": 1527000064
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002714042126379138,
      "loss": 2.5778,
      "theoretical_loss": 3.510550723018471,
      "tokens_seen": 1527065600
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002713941825476429,
      "loss": 2.6325,
      "theoretical_loss": 3.510537485307693,
      "tokens_seen": 1527131136
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027138415245737214,
      "loss": 2.6391,
      "theoretical_loss": 3.5105242483240486,
      "tokens_seen": 1527196672
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027137412236710127,
      "loss": 2.6166,
      "theoretical_loss": 3.510511012067468,
      "tokens_seen": 1527262208
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002713640922768305,
      "loss": 2.8506,
      "theoretical_loss": 3.510497776537879,
      "tokens_seen": 1527327744
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002713540621865597,
      "loss": 2.5764,
      "theoretical_loss": 3.5104845417352104,
      "tokens_seen": 1527393280
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027134403209628887,
      "loss": 2.6689,
      "theoretical_loss": 3.510471307659392,
      "tokens_seen": 1527458816
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027133400200601805,
      "loss": 2.7366,
      "theoretical_loss": 3.5104580743103515,
      "tokens_seen": 1527524352
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002713239719157473,
      "loss": 2.6548,
      "theoretical_loss": 3.5104448416880185,
      "tokens_seen": 1527589888
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002713139418254764,
      "loss": 2.5055,
      "theoretical_loss": 3.510431609792322,
      "tokens_seen": 1527655424
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027130391173520564,
      "loss": 2.6919,
      "theoretical_loss": 3.510418378623191,
      "tokens_seen": 1527720960
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027129388164493477,
      "loss": 2.437,
      "theoretical_loss": 3.510405148180554,
      "tokens_seen": 1527786496
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.000271283851554664,
      "loss": 2.5164,
      "theoretical_loss": 3.5103919184643404,
      "tokens_seen": 1527852032
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002712738214643932,
      "loss": 2.5583,
      "theoretical_loss": 3.5103786894744786,
      "tokens_seen": 1527917568
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027126379137412237,
      "loss": 2.5135,
      "theoretical_loss": 3.510365461210898,
      "tokens_seen": 1527983104
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027125376128385155,
      "loss": 2.7525,
      "theoretical_loss": 3.5103522336735278,
      "tokens_seen": 1528048640
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027124373119358073,
      "loss": 2.6378,
      "theoretical_loss": 3.510339006862296,
      "tokens_seen": 1528114176
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002712337011033099,
      "loss": 2.7063,
      "theoretical_loss": 3.5103257807771326,
      "tokens_seen": 1528179712
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027122367101303915,
      "loss": 2.4968,
      "theoretical_loss": 3.5103125554179666,
      "tokens_seen": 1528245248
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002712136409227683,
      "loss": 2.6722,
      "theoretical_loss": 3.5102993307847266,
      "tokens_seen": 1528310784
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002712036108324975,
      "loss": 2.5445,
      "theoretical_loss": 3.510286106877341,
      "tokens_seen": 1528376320
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002711935807422267,
      "loss": 2.4946,
      "theoretical_loss": 3.51027288369574,
      "tokens_seen": 1528441856
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027118355065195587,
      "loss": 2.6322,
      "theoretical_loss": 3.5102596612398522,
      "tokens_seen": 1528507392
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027117352056168505,
      "loss": 2.7352,
      "theoretical_loss": 3.510246439509607,
      "tokens_seen": 1528572928
    },
    {
      "epoch": 5.01,
      "objective/train/docs_used": 1747177,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.3864831924438477,
      "objective/train/theoretical_loss": 3.5102365236880826,
      "objective/train/tokens_used": 1549082080,
      "theoretical_loss": 3.5102365236880826,
      "tokens_seen": 1528622080
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027116349047141423,
      "loss": 2.586,
      "theoretical_loss": 3.5102332185049328,
      "tokens_seen": 1528638464
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002711534603811434,
      "loss": 2.5084,
      "theoretical_loss": 3.510219998225759,
      "tokens_seen": 1528704000
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027114343029087265,
      "loss": 2.6257,
      "theoretical_loss": 3.5102067786720146,
      "tokens_seen": 1528769536
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002711334002006018,
      "loss": 2.5572,
      "theoretical_loss": 3.510193559843629,
      "tokens_seen": 1528835072
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.000271123370110331,
      "loss": 2.5593,
      "theoretical_loss": 3.510180341740531,
      "tokens_seen": 1528900608
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027111334002006014,
      "loss": 2.6829,
      "theoretical_loss": 3.51016712436265,
      "tokens_seen": 1528966144
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002711033099297894,
      "loss": 2.7308,
      "theoretical_loss": 3.510153907709914,
      "tokens_seen": 1529031680
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027109327983951856,
      "loss": 2.76,
      "theoretical_loss": 3.5101406917822544,
      "tokens_seen": 1529097216
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027108324974924774,
      "loss": 2.4291,
      "theoretical_loss": 3.5101274765795987,
      "tokens_seen": 1529162752
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002710732196589769,
      "loss": 2.5715,
      "theoretical_loss": 3.510114262101876,
      "tokens_seen": 1529228288
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002710631895687061,
      "loss": 2.822,
      "theoretical_loss": 3.510101048349016,
      "tokens_seen": 1529293824
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002710531594784353,
      "loss": 2.5231,
      "theoretical_loss": 3.510087835320948,
      "tokens_seen": 1529359360
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002710431293881645,
      "loss": 2.6624,
      "theoretical_loss": 3.5100746230176005,
      "tokens_seen": 1529424896
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027103309929789364,
      "loss": 2.5751,
      "theoretical_loss": 3.5100614114389033,
      "tokens_seen": 1529490432
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002710230692076229,
      "loss": 2.7269,
      "theoretical_loss": 3.5100482005847855,
      "tokens_seen": 1529555968
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002710130391173521,
      "loss": 2.7465,
      "theoretical_loss": 3.5100349904551766,
      "tokens_seen": 1529621504
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027100300902708124,
      "loss": 2.4731,
      "theoretical_loss": 3.5100217810500047,
      "tokens_seen": 1529687040
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002709929789368105,
      "loss": 2.578,
      "theoretical_loss": 3.5100085723692,
      "tokens_seen": 1529752576
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002709829488465396,
      "loss": 2.4907,
      "theoretical_loss": 3.509995364412692,
      "tokens_seen": 1529818112
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027097291875626884,
      "loss": 2.592,
      "theoretical_loss": 3.509982157180409,
      "tokens_seen": 1529883648
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.000270962888665998,
      "loss": 2.54,
      "theoretical_loss": 3.5099689506722815,
      "tokens_seen": 1529949184
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002709528585757272,
      "loss": 2.6529,
      "theoretical_loss": 3.5099557448882375,
      "tokens_seen": 1530014720
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002709428284854564,
      "loss": 2.7304,
      "theoretical_loss": 3.509942539828207,
      "tokens_seen": 1530080256
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027093279839518556,
      "loss": 2.5196,
      "theoretical_loss": 3.509929335492119,
      "tokens_seen": 1530145792
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027092276830491474,
      "loss": 2.6669,
      "theoretical_loss": 3.5099161318799035,
      "tokens_seen": 1530211328
    },
    {
      "epoch": 5.01,
      "objective/train/docs_used": 1748466,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8011348247528076,
      "objective/train/theoretical_loss": 3.5099062296457397,
      "objective/train/tokens_used": 1550720480,
      "theoretical_loss": 3.5099062296457397,
      "tokens_seen": 1530260480
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.000270912738214644,
      "loss": 2.5529,
      "theoretical_loss": 3.5099029289914885,
      "tokens_seen": 1530276864
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002709027081243731,
      "loss": 2.4209,
      "theoretical_loss": 3.509889726826805,
      "tokens_seen": 1530342400
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027089267803410234,
      "loss": 2.7938,
      "theoretical_loss": 3.509876525385781,
      "tokens_seen": 1530407936
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027088264794383147,
      "loss": 2.5396,
      "theoretical_loss": 3.5098633246683457,
      "tokens_seen": 1530473472
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002708726178535607,
      "loss": 2.5344,
      "theoretical_loss": 3.50985012467443,
      "tokens_seen": 1530539008
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002708625877632899,
      "loss": 2.6562,
      "theoretical_loss": 3.509836925403962,
      "tokens_seen": 1530604544
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027085255767301907,
      "loss": 2.6576,
      "theoretical_loss": 3.509823726856871,
      "tokens_seen": 1530670080
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027084252758274825,
      "loss": 2.6448,
      "theoretical_loss": 3.5098105290330874,
      "tokens_seen": 1530735616
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002708324974924775,
      "loss": 2.7434,
      "theoretical_loss": 3.509797331932539,
      "tokens_seen": 1530801152
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002708224674022066,
      "loss": 2.6026,
      "theoretical_loss": 3.509784135555157,
      "tokens_seen": 1530866688
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027081243731193584,
      "loss": 2.6652,
      "theoretical_loss": 3.5097709399008696,
      "tokens_seen": 1530932224
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027080240722166497,
      "loss": 2.4217,
      "theoretical_loss": 3.5097577449696074,
      "tokens_seen": 1530997760
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002707923771313942,
      "loss": 2.6055,
      "theoretical_loss": 3.5097445507612983,
      "tokens_seen": 1531063296
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002707823470411234,
      "loss": 2.6107,
      "theoretical_loss": 3.5097313572758724,
      "tokens_seen": 1531128832
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027077231695085257,
      "loss": 2.6486,
      "theoretical_loss": 3.50971816451326,
      "tokens_seen": 1531194368
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027076228686058175,
      "loss": 2.6052,
      "theoretical_loss": 3.509704972473389,
      "tokens_seen": 1531259904
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027075225677031093,
      "loss": 2.7182,
      "theoretical_loss": 3.5096917811561905,
      "tokens_seen": 1531325440
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002707422266800401,
      "loss": 2.6964,
      "theoretical_loss": 3.5096785905615926,
      "tokens_seen": 1531390976
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027073219658976935,
      "loss": 2.5845,
      "theoretical_loss": 3.509665400689525,
      "tokens_seen": 1531456512
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002707221664994985,
      "loss": 2.6935,
      "theoretical_loss": 3.5096522115399185,
      "tokens_seen": 1531522048
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002707121364092277,
      "loss": 2.6935,
      "theoretical_loss": 3.509639023112701,
      "tokens_seen": 1531587584
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002707021063189569,
      "loss": 2.4785,
      "theoretical_loss": 3.5096258354078036,
      "tokens_seen": 1531653120
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027069207622868607,
      "loss": 2.4478,
      "theoretical_loss": 3.509612648425154,
      "tokens_seen": 1531718656
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027068204613841525,
      "loss": 2.5772,
      "theoretical_loss": 3.509599462164683,
      "tokens_seen": 1531784192
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027067201604814443,
      "loss": 2.6304,
      "theoretical_loss": 3.5095862766263206,
      "tokens_seen": 1531849728
    },
    {
      "epoch": 5.01,
      "objective/train/docs_used": 1749169,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.604491949081421,
      "objective/train/theoretical_loss": 3.509576387946389,
      "objective/train/tokens_used": 1552358880,
      "theoretical_loss": 3.509576387946389,
      "tokens_seen": 1531898880
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002706619859578736,
      "loss": 2.7063,
      "theoretical_loss": 3.5095730918099948,
      "tokens_seen": 1531915264
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027065195586760285,
      "loss": 2.7013,
      "theoretical_loss": 3.5095599077156363,
      "tokens_seen": 1531980800
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.000270641925777332,
      "loss": 2.6335,
      "theoretical_loss": 3.509546724343174,
      "tokens_seen": 1532046336
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002706318956870612,
      "loss": 2.4774,
      "theoretical_loss": 3.5095335416925386,
      "tokens_seen": 1532111872
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027062186559679034,
      "loss": 2.5814,
      "theoretical_loss": 3.5095203597636586,
      "tokens_seen": 1532177408
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002706118355065196,
      "loss": 2.6218,
      "theoretical_loss": 3.509507178556464,
      "tokens_seen": 1532242944
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027060180541624876,
      "loss": 2.5315,
      "theoretical_loss": 3.5094939980708846,
      "tokens_seen": 1532308480
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027059177532597794,
      "loss": 2.7335,
      "theoretical_loss": 3.5094808183068498,
      "tokens_seen": 1532374016
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002705817452357071,
      "loss": 2.503,
      "theoretical_loss": 3.509467639264289,
      "tokens_seen": 1532439552
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002705717151454363,
      "loss": 2.6337,
      "theoretical_loss": 3.5094544609431324,
      "tokens_seen": 1532505088
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002705616850551655,
      "loss": 2.5151,
      "theoretical_loss": 3.5094412833433095,
      "tokens_seen": 1532570624
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002705516549648947,
      "loss": 2.4761,
      "theoretical_loss": 3.5094281064647497,
      "tokens_seen": 1532636160
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027054162487462384,
      "loss": 2.6397,
      "theoretical_loss": 3.5094149303073827,
      "tokens_seen": 1532701696
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002705315947843531,
      "loss": 2.796,
      "theoretical_loss": 3.509401754871139,
      "tokens_seen": 1532767232
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027052156469408226,
      "loss": 2.5567,
      "theoretical_loss": 3.5093885801559472,
      "tokens_seen": 1532832768
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027051153460381144,
      "loss": 2.4442,
      "theoretical_loss": 3.5093754061617375,
      "tokens_seen": 1532898304
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002705015045135406,
      "loss": 2.3743,
      "theoretical_loss": 3.50936223288844,
      "tokens_seen": 1532963840
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002704914744232698,
      "loss": 2.8448,
      "theoretical_loss": 3.5093490603359836,
      "tokens_seen": 1533029376
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.000270481444332999,
      "loss": 2.6492,
      "theoretical_loss": 3.5093358885042987,
      "tokens_seen": 1533094912
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002704714142427282,
      "loss": 2.6417,
      "theoretical_loss": 3.5093227173933146,
      "tokens_seen": 1533160448
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.00027046138415245735,
      "loss": 2.6674,
      "theoretical_loss": 3.5093095470029616,
      "tokens_seen": 1533225984
    },
    {
      "epoch": 5.01,
      "learning_rate": 0.0002704513540621866,
      "loss": 2.7446,
      "theoretical_loss": 3.509296377333169,
      "tokens_seen": 1533291520
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002704413239719157,
      "loss": 2.5301,
      "theoretical_loss": 3.509283208383867,
      "tokens_seen": 1533357056
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00027043129388164494,
      "loss": 2.2968,
      "theoretical_loss": 3.509270040154985,
      "tokens_seen": 1533422592
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002704212637913741,
      "loss": 2.4793,
      "theoretical_loss": 3.5092568726464526,
      "tokens_seen": 1533488128
    },
    {
      "epoch": 5.02,
      "objective/train/docs_used": 1750709,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6057729721069336,
      "objective/train/theoretical_loss": 3.5092469974877414,
      "objective/train/tokens_used": 1553997280,
      "theoretical_loss": 3.5092469974877414,
      "tokens_seen": 1533537280
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002704112337011033,
      "loss": 2.634,
      "theoretical_loss": 3.509243705858201,
      "tokens_seen": 1533553664
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002704012036108325,
      "loss": 2.6113,
      "theoretical_loss": 3.509230539790158,
      "tokens_seen": 1533619200
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00027039117352056167,
      "loss": 2.5485,
      "theoretical_loss": 3.5092173744422546,
      "tokens_seen": 1533684736
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00027038114343029085,
      "loss": 2.6423,
      "theoretical_loss": 3.5092042098144205,
      "tokens_seen": 1533750272
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002703711133400201,
      "loss": 2.7032,
      "theoretical_loss": 3.5091910459065856,
      "tokens_seen": 1533815808
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002703610832497492,
      "loss": 2.5698,
      "theoretical_loss": 3.5091778827186797,
      "tokens_seen": 1533881344
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00027035105315947845,
      "loss": 2.6178,
      "theoretical_loss": 3.5091647202506326,
      "tokens_seen": 1533946880
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00027034102306920763,
      "loss": 2.5169,
      "theoretical_loss": 3.5091515585023743,
      "tokens_seen": 1534012416
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002703309929789368,
      "loss": 2.6137,
      "theoretical_loss": 3.5091383974738344,
      "tokens_seen": 1534077952
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.000270320962888666,
      "loss": 2.7768,
      "theoretical_loss": 3.509125237164943,
      "tokens_seen": 1534143488
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00027031093279839517,
      "loss": 2.5983,
      "theoretical_loss": 3.5091120775756304,
      "tokens_seen": 1534209024
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00027030090270812435,
      "loss": 2.6226,
      "theoretical_loss": 3.5090989187058255,
      "tokens_seen": 1534274560
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002702908726178536,
      "loss": 2.7032,
      "theoretical_loss": 3.5090857605554593,
      "tokens_seen": 1534340096
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002702808425275827,
      "loss": 2.5462,
      "theoretical_loss": 3.5090726031244612,
      "tokens_seen": 1534405632
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00027027081243731195,
      "loss": 2.5581,
      "theoretical_loss": 3.509059446412761,
      "tokens_seen": 1534471168
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00027026078234704113,
      "loss": 2.6566,
      "theoretical_loss": 3.5090462904202893,
      "tokens_seen": 1534536704
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002702507522567703,
      "loss": 2.4762,
      "theoretical_loss": 3.5090331351469755,
      "tokens_seen": 1534602240
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00027024072216649955,
      "loss": 2.6938,
      "theoretical_loss": 3.50901998059275,
      "tokens_seen": 1534667776
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002702306920762287,
      "loss": 2.3536,
      "theoretical_loss": 3.5090068267575423,
      "tokens_seen": 1534733312
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002702206619859579,
      "loss": 2.7465,
      "theoretical_loss": 3.5089936736412826,
      "tokens_seen": 1534798848
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002702106318956871,
      "loss": 2.3321,
      "theoretical_loss": 3.508980521243901,
      "tokens_seen": 1534864384
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00027020060180541627,
      "loss": 2.6597,
      "theoretical_loss": 3.508967369565328,
      "tokens_seen": 1534929920
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00027019057171514545,
      "loss": 2.4596,
      "theoretical_loss": 3.5089542186054925,
      "tokens_seen": 1534995456
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00027018054162487463,
      "loss": 2.597,
      "theoretical_loss": 3.508941068364325,
      "tokens_seen": 1535060992
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002701705115346038,
      "loss": 2.7156,
      "theoretical_loss": 3.5089279188417564,
      "tokens_seen": 1535126528
    },
    {
      "epoch": 5.02,
      "objective/train/docs_used": 1751481,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.573998212814331,
      "objective/train/theoretical_loss": 3.5089180571713676,
      "objective/train/tokens_used": 1555635680,
      "theoretical_loss": 3.5089180571713676,
      "tokens_seen": 1535175680
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00027016048144433305,
      "loss": 2.7034,
      "theoretical_loss": 3.5089147700377152,
      "tokens_seen": 1535192064
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002701504513540622,
      "loss": 2.6163,
      "theoretical_loss": 3.508901621952133,
      "tokens_seen": 1535257600
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002701404212637914,
      "loss": 2.6168,
      "theoretical_loss": 3.508888474584939,
      "tokens_seen": 1535323136
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00027013039117352054,
      "loss": 2.6139,
      "theoretical_loss": 3.508875327936063,
      "tokens_seen": 1535388672
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002701203610832498,
      "loss": 2.7076,
      "theoretical_loss": 3.508862182005436,
      "tokens_seen": 1535454208
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00027011033099297896,
      "loss": 2.5746,
      "theoretical_loss": 3.508849036792988,
      "tokens_seen": 1535519744
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00027010030090270814,
      "loss": 2.8796,
      "theoretical_loss": 3.5088358922986482,
      "tokens_seen": 1535585280
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002700902708124373,
      "loss": 2.5963,
      "theoretical_loss": 3.5088227485223475,
      "tokens_seen": 1535650816
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002700802407221665,
      "loss": 2.4503,
      "theoretical_loss": 3.508809605464016,
      "tokens_seen": 1535716352
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002700702106318957,
      "loss": 2.7169,
      "theoretical_loss": 3.5087964631235833,
      "tokens_seen": 1535781888
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002700601805416249,
      "loss": 2.8176,
      "theoretical_loss": 3.5087833215009807,
      "tokens_seen": 1535847424
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00027005015045135404,
      "loss": 2.5217,
      "theoretical_loss": 3.508770180596137,
      "tokens_seen": 1535912960
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002700401203610833,
      "loss": 2.5807,
      "theoretical_loss": 3.508757040408983,
      "tokens_seen": 1535978496
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00027003009027081246,
      "loss": 2.8575,
      "theoretical_loss": 3.508743900939449,
      "tokens_seen": 1536044032
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00027002006018054164,
      "loss": 2.7623,
      "theoretical_loss": 3.508730762187465,
      "tokens_seen": 1536109568
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002700100300902708,
      "loss": 2.7278,
      "theoretical_loss": 3.508717624152961,
      "tokens_seen": 1536175104
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00027,
      "loss": 2.656,
      "theoretical_loss": 3.5087044868358683,
      "tokens_seen": 1536240640
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002699899699097292,
      "loss": 2.6187,
      "theoretical_loss": 3.5086913502361154,
      "tokens_seen": 1536306176
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002699799398194584,
      "loss": 2.4959,
      "theoretical_loss": 3.508678214353634,
      "tokens_seen": 1536371712
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026996990972918755,
      "loss": 2.5832,
      "theoretical_loss": 3.508665079188354,
      "tokens_seen": 1536437248
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002699598796389168,
      "loss": 2.5339,
      "theoretical_loss": 3.5086519447402047,
      "tokens_seen": 1536502784
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002699498495486459,
      "loss": 2.4782,
      "theoretical_loss": 3.508638811009118,
      "tokens_seen": 1536568320
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026993981945837514,
      "loss": 2.7155,
      "theoretical_loss": 3.5086256779950222,
      "tokens_seen": 1536633856
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002699297893681043,
      "loss": 2.7312,
      "theoretical_loss": 3.5086125456978494,
      "tokens_seen": 1536699392
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002699197592778335,
      "loss": 2.7479,
      "theoretical_loss": 3.508599414117529,
      "tokens_seen": 1536764928
    },
    {
      "epoch": 5.02,
      "objective/train/docs_used": 1752901,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.562196969985962,
      "objective/train/theoretical_loss": 3.508589565902681,
      "objective/train/tokens_used": 1557274080,
      "theoretical_loss": 3.508589565902681,
      "tokens_seen": 1536814080
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002699097291875627,
      "loss": 2.6855,
      "theoretical_loss": 3.508586283253991,
      "tokens_seen": 1536830464
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026989969909729187,
      "loss": 2.5581,
      "theoretical_loss": 3.508573153107167,
      "tokens_seen": 1536896000
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026988966900702105,
      "loss": 2.6018,
      "theoretical_loss": 3.508560023676986,
      "tokens_seen": 1536961536
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002698796389167503,
      "loss": 2.5101,
      "theoretical_loss": 3.508546894963379,
      "tokens_seen": 1537027072
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002698696088264794,
      "loss": 2.5575,
      "theoretical_loss": 3.5085337669662753,
      "tokens_seen": 1537092608
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026985957873620865,
      "loss": 2.6392,
      "theoretical_loss": 3.5085206396856075,
      "tokens_seen": 1537158144
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026984954864593783,
      "loss": 2.6987,
      "theoretical_loss": 3.5085075131213035,
      "tokens_seen": 1537223680
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.000269839518555667,
      "loss": 2.6348,
      "theoretical_loss": 3.508494387273295,
      "tokens_seen": 1537289216
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002698294884653962,
      "loss": 2.6593,
      "theoretical_loss": 3.508481262141512,
      "tokens_seen": 1537354752
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026981945837512537,
      "loss": 2.7099,
      "theoretical_loss": 3.508468137725885,
      "tokens_seen": 1537420288
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026980942828485455,
      "loss": 2.5047,
      "theoretical_loss": 3.508455014026345,
      "tokens_seen": 1537485824
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002697993981945838,
      "loss": 2.6675,
      "theoretical_loss": 3.5084418910428212,
      "tokens_seen": 1537551360
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002697893681043129,
      "loss": 2.6193,
      "theoretical_loss": 3.5084287687752447,
      "tokens_seen": 1537616896
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026977933801404215,
      "loss": 2.7267,
      "theoretical_loss": 3.5084156472235457,
      "tokens_seen": 1537682432
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002697693079237713,
      "loss": 2.6067,
      "theoretical_loss": 3.508402526387655,
      "tokens_seen": 1537747968
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002697592778335005,
      "loss": 2.54,
      "theoretical_loss": 3.5083894062675025,
      "tokens_seen": 1537813504
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002697492477432297,
      "loss": 2.7133,
      "theoretical_loss": 3.5083762868630197,
      "tokens_seen": 1537879040
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002697392176529589,
      "loss": 2.3966,
      "theoretical_loss": 3.508363168174136,
      "tokens_seen": 1537944576
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026972918756268806,
      "loss": 2.5644,
      "theoretical_loss": 3.508350050200782,
      "tokens_seen": 1538010112
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002697191574724173,
      "loss": 2.6717,
      "theoretical_loss": 3.508336932942888,
      "tokens_seen": 1538075648
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002697091273821464,
      "loss": 2.7196,
      "theoretical_loss": 3.5083238164003863,
      "tokens_seen": 1538141184
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026969909729187565,
      "loss": 2.482,
      "theoretical_loss": 3.5083107005732046,
      "tokens_seen": 1538206720
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002696890672016048,
      "loss": 2.6231,
      "theoretical_loss": 3.5082975854612757,
      "tokens_seen": 1538272256
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.000269679037111334,
      "loss": 2.6503,
      "theoretical_loss": 3.508284471064529,
      "tokens_seen": 1538337792
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002696690070210632,
      "loss": 2.5515,
      "theoretical_loss": 3.5082713573828954,
      "tokens_seen": 1538403328
    },
    {
      "epoch": 5.02,
      "objective/train/docs_used": 1753471,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.3333654403686523,
      "objective/train/theoretical_loss": 3.508261522590921,
      "objective/train/tokens_used": 1558912480,
      "theoretical_loss": 3.508261522590921,
      "tokens_seen": 1538452480
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002696589769307924,
      "loss": 2.6811,
      "theoretical_loss": 3.508258244416305,
      "tokens_seen": 1538468864
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026964894684052156,
      "loss": 2.8215,
      "theoretical_loss": 3.508245132164689,
      "tokens_seen": 1538534400
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026963891675025074,
      "loss": 2.5912,
      "theoretical_loss": 3.5082320206279776,
      "tokens_seen": 1538599936
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002696288866599799,
      "loss": 2.5781,
      "theoretical_loss": 3.508218909806101,
      "tokens_seen": 1538665472
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026961885656970916,
      "loss": 2.7099,
      "theoretical_loss": 3.508205799698991,
      "tokens_seen": 1538731008
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002696088264794383,
      "loss": 2.8744,
      "theoretical_loss": 3.508192690306577,
      "tokens_seen": 1538796544
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002695987963891675,
      "loss": 2.599,
      "theoretical_loss": 3.5081795816287897,
      "tokens_seen": 1538862080
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026958876629889665,
      "loss": 2.6964,
      "theoretical_loss": 3.5081664736655602,
      "tokens_seen": 1538927616
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002695787362086259,
      "loss": 2.7133,
      "theoretical_loss": 3.5081533664168196,
      "tokens_seen": 1538993152
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026956870611835506,
      "loss": 2.7044,
      "theoretical_loss": 3.508140259882497,
      "tokens_seen": 1539058688
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026955867602808424,
      "loss": 2.4334,
      "theoretical_loss": 3.5081271540625245,
      "tokens_seen": 1539124224
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002695486459378134,
      "loss": 2.7716,
      "theoretical_loss": 3.5081140489568314,
      "tokens_seen": 1539189760
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026953861584754266,
      "loss": 2.6196,
      "theoretical_loss": 3.50810094456535,
      "tokens_seen": 1539255296
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002695285857572718,
      "loss": 2.683,
      "theoretical_loss": 3.5080878408880096,
      "tokens_seen": 1539320832
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.000269518555667001,
      "loss": 2.5391,
      "theoretical_loss": 3.5080747379247414,
      "tokens_seen": 1539386368
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002695085255767302,
      "loss": 2.3732,
      "theoretical_loss": 3.5080616356754764,
      "tokens_seen": 1539451904
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002694984954864594,
      "loss": 2.4932,
      "theoretical_loss": 3.5080485341401446,
      "tokens_seen": 1539517440
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002694884653961886,
      "loss": 2.5774,
      "theoretical_loss": 3.5080354333186774,
      "tokens_seen": 1539582976
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026947843530591775,
      "loss": 2.5516,
      "theoretical_loss": 3.508022333211005,
      "tokens_seen": 1539648512
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.000269468405215647,
      "loss": 2.6029,
      "theoretical_loss": 3.5080092338170585,
      "tokens_seen": 1539714048
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002694583751253761,
      "loss": 2.5132,
      "theoretical_loss": 3.507996135136768,
      "tokens_seen": 1539779584
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026944834503510534,
      "loss": 2.6247,
      "theoretical_loss": 3.507983037170065,
      "tokens_seen": 1539845120
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002694383149448345,
      "loss": 2.4717,
      "theoretical_loss": 3.50796993991688,
      "tokens_seen": 1539910656
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002694282848545637,
      "loss": 2.6097,
      "theoretical_loss": 3.5079568433771438,
      "tokens_seen": 1539976192
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002694182547642929,
      "loss": 2.463,
      "theoretical_loss": 3.5079437475507866,
      "tokens_seen": 1540041728
    },
    {
      "epoch": 5.02,
      "objective/train/docs_used": 1754850,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4801652431488037,
      "objective/train/theoretical_loss": 3.507933926149133,
      "objective/train/tokens_used": 1560550880,
      "theoretical_loss": 3.507933926149133,
      "tokens_seen": 1540090880
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026940822467402207,
      "loss": 2.5182,
      "theoretical_loss": 3.5079306524377403,
      "tokens_seen": 1540107264
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026939819458375125,
      "loss": 2.5174,
      "theoretical_loss": 3.507917558037935,
      "tokens_seen": 1540172800
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002693881644934805,
      "loss": 2.5887,
      "theoretical_loss": 3.5079044643513013,
      "tokens_seen": 1540238336
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002693781344032096,
      "loss": 2.5088,
      "theoretical_loss": 3.507891371377771,
      "tokens_seen": 1540303872
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026936810431293885,
      "loss": 2.6441,
      "theoretical_loss": 3.507878279117273,
      "tokens_seen": 1540369408
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026935807422266803,
      "loss": 2.6603,
      "theoretical_loss": 3.507865187569741,
      "tokens_seen": 1540434944
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002693480441323972,
      "loss": 2.4486,
      "theoretical_loss": 3.507852096735103,
      "tokens_seen": 1540500480
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002693380140421264,
      "loss": 2.4665,
      "theoretical_loss": 3.507839006613291,
      "tokens_seen": 1540566016
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026932798395185557,
      "loss": 2.7573,
      "theoretical_loss": 3.5078259172042365,
      "tokens_seen": 1540631552
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026931795386158475,
      "loss": 2.4961,
      "theoretical_loss": 3.50781282850787,
      "tokens_seen": 1540697088
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.000269307923771314,
      "loss": 2.6341,
      "theoretical_loss": 3.5077997405241215,
      "tokens_seen": 1540762624
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002692978936810431,
      "loss": 2.4809,
      "theoretical_loss": 3.507786653252923,
      "tokens_seen": 1540828160
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026928786359077235,
      "loss": 2.6053,
      "theoretical_loss": 3.5077735666942047,
      "tokens_seen": 1540893696
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002692778335005015,
      "loss": 2.6354,
      "theoretical_loss": 3.507760480847898,
      "tokens_seen": 1540959232
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002692678034102307,
      "loss": 2.5585,
      "theoretical_loss": 3.507747395713934,
      "tokens_seen": 1541024768
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002692577733199599,
      "loss": 2.4733,
      "theoretical_loss": 3.507734311292243,
      "tokens_seen": 1541090304
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002692477432296891,
      "loss": 2.6158,
      "theoretical_loss": 3.507721227582756,
      "tokens_seen": 1541155840
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026923771313941826,
      "loss": 2.6618,
      "theoretical_loss": 3.507708144585404,
      "tokens_seen": 1541221376
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002692276830491475,
      "loss": 2.8516,
      "theoretical_loss": 3.507695062300118,
      "tokens_seen": 1541286912
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002692176529588766,
      "loss": 2.5671,
      "theoretical_loss": 3.5076819807268294,
      "tokens_seen": 1541352448
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026920762286860585,
      "loss": 2.5019,
      "theoretical_loss": 3.507668899865469,
      "tokens_seen": 1541417984
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.000269197592778335,
      "loss": 2.6017,
      "theoretical_loss": 3.5076558197159673,
      "tokens_seen": 1541483520
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002691875626880642,
      "loss": 2.7025,
      "theoretical_loss": 3.507642740278256,
      "tokens_seen": 1541549056
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002691775325977934,
      "loss": 2.6308,
      "theoretical_loss": 3.5076296615522655,
      "tokens_seen": 1541614592
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002691675025075226,
      "loss": 2.4931,
      "theoretical_loss": 3.507616583537927,
      "tokens_seen": 1541680128
    },
    {
      "epoch": 5.02,
      "objective/train/docs_used": 1755645,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.596935987472534,
      "objective/train/theoretical_loss": 3.5076067754941533,
      "objective/train/tokens_used": 1562189280,
      "theoretical_loss": 3.5076067754941533,
      "tokens_seen": 1541729280
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026915747241725176,
      "loss": 2.7508,
      "theoretical_loss": 3.5076035062351716,
      "tokens_seen": 1541745664
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026914744232698094,
      "loss": 2.6642,
      "theoretical_loss": 3.5075904296439306,
      "tokens_seen": 1541811200
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002691374122367101,
      "loss": 2.749,
      "theoretical_loss": 3.507577353764135,
      "tokens_seen": 1541876736
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026912738214643936,
      "loss": 2.6091,
      "theoretical_loss": 3.507564278595715,
      "tokens_seen": 1541942272
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002691173520561685,
      "loss": 2.549,
      "theoretical_loss": 3.5075512041386023,
      "tokens_seen": 1542007808
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002691073219658977,
      "loss": 2.5448,
      "theoretical_loss": 3.507538130392728,
      "tokens_seen": 1542073344
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026909729187562685,
      "loss": 2.4752,
      "theoretical_loss": 3.5075250573580234,
      "tokens_seen": 1542138880
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002690872617853561,
      "loss": 2.6303,
      "theoretical_loss": 3.5075119850344194,
      "tokens_seen": 1542204416
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026907723169508526,
      "loss": 2.5021,
      "theoretical_loss": 3.507498913421847,
      "tokens_seen": 1542269952
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026906720160481444,
      "loss": 2.6581,
      "theoretical_loss": 3.5074858425202375,
      "tokens_seen": 1542335488
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002690571715145436,
      "loss": 2.6053,
      "theoretical_loss": 3.5074727723295216,
      "tokens_seen": 1542401024
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026904714142427286,
      "loss": 2.5677,
      "theoretical_loss": 3.507459702849631,
      "tokens_seen": 1542466560
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.000269037111334002,
      "loss": 2.7589,
      "theoretical_loss": 3.5074466340804964,
      "tokens_seen": 1542532096
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002690270812437312,
      "loss": 2.6401,
      "theoretical_loss": 3.507433566022049,
      "tokens_seen": 1542597632
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026901705115346035,
      "loss": 2.4273,
      "theoretical_loss": 3.5074204986742203,
      "tokens_seen": 1542663168
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002690070210631896,
      "loss": 2.7618,
      "theoretical_loss": 3.507407432036941,
      "tokens_seen": 1542728704
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026899699097291877,
      "loss": 2.7275,
      "theoretical_loss": 3.5073943661101428,
      "tokens_seen": 1542794240
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026898696088264795,
      "loss": 2.6308,
      "theoretical_loss": 3.5073813008937567,
      "tokens_seen": 1542859776
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026897693079237713,
      "loss": 2.3812,
      "theoretical_loss": 3.5073682363877134,
      "tokens_seen": 1542925312
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002689669007021063,
      "loss": 2.7549,
      "theoretical_loss": 3.507355172591945,
      "tokens_seen": 1542990848
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002689568706118355,
      "loss": 2.3385,
      "theoretical_loss": 3.507342109506382,
      "tokens_seen": 1543056384
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002689468405215647,
      "loss": 2.6484,
      "theoretical_loss": 3.5073290471309564,
      "tokens_seen": 1543121920
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026893681043129385,
      "loss": 2.4296,
      "theoretical_loss": 3.507315985465598,
      "tokens_seen": 1543187456
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002689267803410231,
      "loss": 2.732,
      "theoretical_loss": 3.5073029245102396,
      "tokens_seen": 1543252992
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002689167502507522,
      "loss": 2.429,
      "theoretical_loss": 3.507289864264812,
      "tokens_seen": 1543318528
    },
    {
      "epoch": 5.02,
      "objective/train/docs_used": 1757071,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.122141122817993,
      "objective/train/theoretical_loss": 3.5072800695465913,
      "objective/train/tokens_used": 1563827680,
      "theoretical_loss": 3.5072800695465913,
      "tokens_seen": 1543367680
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026890672016048145,
      "loss": 2.5328,
      "theoretical_loss": 3.507276804729246,
      "tokens_seen": 1543384064
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026889669007021063,
      "loss": 2.6986,
      "theoretical_loss": 3.5072637459034732,
      "tokens_seen": 1543449600
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002688866599799398,
      "loss": 2.3406,
      "theoretical_loss": 3.5072506877874243,
      "tokens_seen": 1543515136
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.000268876629889669,
      "loss": 2.4533,
      "theoretical_loss": 3.5072376303810318,
      "tokens_seen": 1543580672
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026886659979939823,
      "loss": 2.6335,
      "theoretical_loss": 3.5072245736842262,
      "tokens_seen": 1543646208
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026885656970912736,
      "loss": 2.7103,
      "theoretical_loss": 3.5072115176969394,
      "tokens_seen": 1543711744
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002688465396188566,
      "loss": 2.771,
      "theoretical_loss": 3.507198462419102,
      "tokens_seen": 1543777280
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002688365095285857,
      "loss": 2.6098,
      "theoretical_loss": 3.5071854078506455,
      "tokens_seen": 1543842816
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026882647943831495,
      "loss": 2.6712,
      "theoretical_loss": 3.507172353991501,
      "tokens_seen": 1543908352
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026881644934804413,
      "loss": 2.5959,
      "theoretical_loss": 3.507159300841601,
      "tokens_seen": 1543973888
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002688064192577733,
      "loss": 2.8207,
      "theoretical_loss": 3.5071462484008755,
      "tokens_seen": 1544039424
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002687963891675025,
      "loss": 2.3538,
      "theoretical_loss": 3.507133196669257,
      "tokens_seen": 1544104960
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002687863590772317,
      "loss": 2.6198,
      "theoretical_loss": 3.507120145646676,
      "tokens_seen": 1544170496
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026877632898696086,
      "loss": 2.6675,
      "theoretical_loss": 3.5071070953330645,
      "tokens_seen": 1544236032
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002687662988966901,
      "loss": 2.5703,
      "theoretical_loss": 3.507094045728353,
      "tokens_seen": 1544301568
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002687562688064193,
      "loss": 2.8102,
      "theoretical_loss": 3.5070809968324745,
      "tokens_seen": 1544367104
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026874623871614846,
      "loss": 2.4871,
      "theoretical_loss": 3.507067948645359,
      "tokens_seen": 1544432640
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002687362086258777,
      "loss": 2.674,
      "theoretical_loss": 3.5070549011669385,
      "tokens_seen": 1544498176
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002687261785356068,
      "loss": 2.6122,
      "theoretical_loss": 3.507041854397144,
      "tokens_seen": 1544563712
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026871614844533605,
      "loss": 2.4205,
      "theoretical_loss": 3.5070288083359076,
      "tokens_seen": 1544629248
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002687061183550652,
      "loss": 2.5357,
      "theoretical_loss": 3.50701576298316,
      "tokens_seen": 1544694784
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002686960882647944,
      "loss": 2.6311,
      "theoretical_loss": 3.507002718338834,
      "tokens_seen": 1544760320
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002686860581745236,
      "loss": 2.7446,
      "theoretical_loss": 3.5069896744028597,
      "tokens_seen": 1544825856
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002686760280842528,
      "loss": 2.7055,
      "theoretical_loss": 3.506976631175169,
      "tokens_seen": 1544891392
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026866599799398196,
      "loss": 2.8652,
      "theoretical_loss": 3.506963588655694,
      "tokens_seen": 1544956928
    },
    {
      "epoch": 5.02,
      "objective/train/docs_used": 1757632,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.365391969680786,
      "objective/train/theoretical_loss": 3.506953807230812,
      "objective/train/tokens_used": 1565466080,
      "theoretical_loss": 3.506953807230812,
      "tokens_seen": 1545006080
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026865596790371114,
      "loss": 2.5397,
      "theoretical_loss": 3.506950546844365,
      "tokens_seen": 1545022464
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002686459378134403,
      "loss": 2.7858,
      "theoretical_loss": 3.5069375057411145,
      "tokens_seen": 1545088000
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026863590772316956,
      "loss": 2.6465,
      "theoretical_loss": 3.5069244653458735,
      "tokens_seen": 1545153536
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002686258776328987,
      "loss": 2.4181,
      "theoretical_loss": 3.506911425658574,
      "tokens_seen": 1545219072
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002686158475426279,
      "loss": 2.6752,
      "theoretical_loss": 3.5068983866791474,
      "tokens_seen": 1545284608
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026860581745235705,
      "loss": 2.8851,
      "theoretical_loss": 3.506885348407525,
      "tokens_seen": 1545350144
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002685957873620863,
      "loss": 2.6663,
      "theoretical_loss": 3.5068723108436384,
      "tokens_seen": 1545415680
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026858575727181546,
      "loss": 2.6977,
      "theoretical_loss": 3.50685927398742,
      "tokens_seen": 1545481216
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026857572718154464,
      "loss": 2.7299,
      "theoretical_loss": 3.5068462378387997,
      "tokens_seen": 1545546752
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002685656970912738,
      "loss": 2.7172,
      "theoretical_loss": 3.5068332023977105,
      "tokens_seen": 1545612288
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026855566700100306,
      "loss": 2.5429,
      "theoretical_loss": 3.5068201676640838,
      "tokens_seen": 1545677824
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002685456369107322,
      "loss": 2.7161,
      "theoretical_loss": 3.506807133637851,
      "tokens_seen": 1545743360
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002685356068204614,
      "loss": 2.6667,
      "theoretical_loss": 3.506794100318943,
      "tokens_seen": 1545808896
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026852557673019055,
      "loss": 2.8401,
      "theoretical_loss": 3.506781067707293,
      "tokens_seen": 1545874432
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002685155466399198,
      "loss": 2.5245,
      "theoretical_loss": 3.506768035802831,
      "tokens_seen": 1545939968
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026850551654964897,
      "loss": 2.7198,
      "theoretical_loss": 3.5067550046054903,
      "tokens_seen": 1546005504
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026849548645937815,
      "loss": 2.5475,
      "theoretical_loss": 3.506741974115201,
      "tokens_seen": 1546071040
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026848545636910733,
      "loss": 2.363,
      "theoretical_loss": 3.5067289443318956,
      "tokens_seen": 1546136576
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002684754262788365,
      "loss": 2.755,
      "theoretical_loss": 3.506715915255506,
      "tokens_seen": 1546202112
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002684653961885657,
      "loss": 2.4602,
      "theoretical_loss": 3.5067028868859627,
      "tokens_seen": 1546267648
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002684553660982949,
      "loss": 2.4851,
      "theoretical_loss": 3.506689859223199,
      "tokens_seen": 1546333184
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026844533600802405,
      "loss": 2.4482,
      "theoretical_loss": 3.5066768322671455,
      "tokens_seen": 1546398720
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002684353059177533,
      "loss": 2.4997,
      "theoretical_loss": 3.506663806017734,
      "tokens_seen": 1546464256
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002684252758274824,
      "loss": 2.5183,
      "theoretical_loss": 3.5066507804748968,
      "tokens_seen": 1546529792
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026841524573721165,
      "loss": 2.7061,
      "theoretical_loss": 3.506637755638565,
      "tokens_seen": 1546595328
    },
    {
      "epoch": 5.02,
      "objective/train/docs_used": 1758939,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.418158769607544,
      "objective/train/theoretical_loss": 3.5066279874749195,
      "objective/train/tokens_used": 1567104480,
      "theoretical_loss": 3.5066279874749195,
      "tokens_seen": 1546644480
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026840521564694083,
      "loss": 2.5213,
      "theoretical_loss": 3.506624731508671,
      "tokens_seen": 1546660864
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026839518555667,
      "loss": 2.4823,
      "theoretical_loss": 3.5066117080851456,
      "tokens_seen": 1546726400
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002683851554663992,
      "loss": 2.4184,
      "theoretical_loss": 3.506598685367922,
      "tokens_seen": 1546791936
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026837512537612843,
      "loss": 2.6234,
      "theoretical_loss": 3.5065856633569306,
      "tokens_seen": 1546857472
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026836509528585756,
      "loss": 2.7849,
      "theoretical_loss": 3.5065726420521033,
      "tokens_seen": 1546923008
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002683550651955868,
      "loss": 2.7529,
      "theoretical_loss": 3.506559621453373,
      "tokens_seen": 1546988544
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002683450351053159,
      "loss": 2.6108,
      "theoretical_loss": 3.5065466015606708,
      "tokens_seen": 1547054080
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026833500501504515,
      "loss": 2.7479,
      "theoretical_loss": 3.506533582373928,
      "tokens_seen": 1547119616
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026832497492477433,
      "loss": 2.6387,
      "theoretical_loss": 3.506520563893077,
      "tokens_seen": 1547185152
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002683149448345035,
      "loss": 2.6088,
      "theoretical_loss": 3.5065075461180495,
      "tokens_seen": 1547250688
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002683049147442327,
      "loss": 2.7851,
      "theoretical_loss": 3.506494529048778,
      "tokens_seen": 1547316224
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002682948846539619,
      "loss": 2.5244,
      "theoretical_loss": 3.506481512685193,
      "tokens_seen": 1547381760
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026828485456369106,
      "loss": 2.6853,
      "theoretical_loss": 3.5064684970272277,
      "tokens_seen": 1547447296
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002682748244734203,
      "loss": 2.7179,
      "theoretical_loss": 3.506455482074813,
      "tokens_seen": 1547512832
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002682647943831494,
      "loss": 2.4258,
      "theoretical_loss": 3.506442467827881,
      "tokens_seen": 1547578368
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026825476429287866,
      "loss": 2.7019,
      "theoretical_loss": 3.5064294542863643,
      "tokens_seen": 1547643904
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002682447342026078,
      "loss": 2.7355,
      "theoretical_loss": 3.506416441450194,
      "tokens_seen": 1547709440
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.000268234704112337,
      "loss": 2.5778,
      "theoretical_loss": 3.506403429319302,
      "tokens_seen": 1547774976
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002682246740220662,
      "loss": 2.4957,
      "theoretical_loss": 3.5063904178936207,
      "tokens_seen": 1547840512
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002682146439317954,
      "loss": 2.7598,
      "theoretical_loss": 3.5063774071730816,
      "tokens_seen": 1547906048
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026820461384152456,
      "loss": 2.859,
      "theoretical_loss": 3.5063643971576166,
      "tokens_seen": 1547971584
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002681945837512538,
      "loss": 2.73,
      "theoretical_loss": 3.5063513878471584,
      "tokens_seen": 1548037120
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002681845536609829,
      "loss": 2.6316,
      "theoretical_loss": 3.506338379241638,
      "tokens_seen": 1548102656
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026817452357071216,
      "loss": 2.4397,
      "theoretical_loss": 3.5063253713409877,
      "tokens_seen": 1548168192
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002681644934804413,
      "loss": 2.6663,
      "theoretical_loss": 3.5063123641451397,
      "tokens_seen": 1548233728
    },
    {
      "epoch": 5.02,
      "objective/train/docs_used": 1759677,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.676605463027954,
      "objective/train/theoretical_loss": 3.5063026092107394,
      "objective/train/tokens_used": 1568742880,
      "theoretical_loss": 3.5063026092107394,
      "tokens_seen": 1548282880
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002681544633901705,
      "loss": 2.6747,
      "theoretical_loss": 3.506299357654026,
      "tokens_seen": 1548299264
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002681444332998997,
      "loss": 2.5841,
      "theoretical_loss": 3.506286351867578,
      "tokens_seen": 1548364800
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002681344032096289,
      "loss": 2.6078,
      "theoretical_loss": 3.5062733467857283,
      "tokens_seen": 1548430336
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026812437311935807,
      "loss": 2.7631,
      "theoretical_loss": 3.5062603424084084,
      "tokens_seen": 1548495872
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026811434302908725,
      "loss": 2.5097,
      "theoretical_loss": 3.506247338735551,
      "tokens_seen": 1548561408
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026810431293881643,
      "loss": 2.6734,
      "theoretical_loss": 3.506234335767088,
      "tokens_seen": 1548626944
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026809428284854566,
      "loss": 2.2408,
      "theoretical_loss": 3.5062213335029506,
      "tokens_seen": 1548692480
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002680842527582748,
      "loss": 2.6834,
      "theoretical_loss": 3.5062083319430717,
      "tokens_seen": 1548758016
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.000268074222668004,
      "loss": 2.9695,
      "theoretical_loss": 3.5061953310873832,
      "tokens_seen": 1548823552
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026806419257773315,
      "loss": 2.544,
      "theoretical_loss": 3.5061823309358164,
      "tokens_seen": 1548889088
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002680541624874624,
      "loss": 2.6153,
      "theoretical_loss": 3.506169331488305,
      "tokens_seen": 1548954624
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026804413239719157,
      "loss": 2.7174,
      "theoretical_loss": 3.506156332744779,
      "tokens_seen": 1549020160
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026803410230692075,
      "loss": 2.7422,
      "theoretical_loss": 3.5061433347051727,
      "tokens_seen": 1549085696
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026802407221664993,
      "loss": 2.7367,
      "theoretical_loss": 3.5061303373694166,
      "tokens_seen": 1549151232
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026801404212637917,
      "loss": 2.5945,
      "theoretical_loss": 3.5061173407374433,
      "tokens_seen": 1549216768
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026800401203610835,
      "loss": 2.4936,
      "theoretical_loss": 3.5061043448091853,
      "tokens_seen": 1549282304
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026799398194583753,
      "loss": 2.7071,
      "theoretical_loss": 3.5060913495845742,
      "tokens_seen": 1549347840
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002679839518555667,
      "loss": 2.6455,
      "theoretical_loss": 3.506078355063542,
      "tokens_seen": 1549413376
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002679739217652959,
      "loss": 2.599,
      "theoretical_loss": 3.506065361246021,
      "tokens_seen": 1549478912
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002679638916750251,
      "loss": 2.7328,
      "theoretical_loss": 3.5060523681319444,
      "tokens_seen": 1549544448
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026795386158475425,
      "loss": 2.6709,
      "theoretical_loss": 3.506039375721243,
      "tokens_seen": 1549609984
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002679438314944835,
      "loss": 2.8018,
      "theoretical_loss": 3.5060263840138495,
      "tokens_seen": 1549675520
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002679338014042126,
      "loss": 2.733,
      "theoretical_loss": 3.506013393009696,
      "tokens_seen": 1549741056
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026792377131394185,
      "loss": 2.5778,
      "theoretical_loss": 3.506000402708715,
      "tokens_seen": 1549806592
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026791374122367103,
      "loss": 2.4503,
      "theoretical_loss": 3.505987413110838,
      "tokens_seen": 1549872128
    },
    {
      "epoch": 5.02,
      "objective/train/docs_used": 1761272,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8500523567199707,
      "objective/train/theoretical_loss": 3.5059776713738025,
      "objective/train/tokens_used": 1570381280,
      "theoretical_loss": 3.5059776713738025,
      "tokens_seen": 1549921280
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002679037111334002,
      "loss": 2.5811,
      "theoretical_loss": 3.505974424215998,
      "tokens_seen": 1549937664
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002678936810431294,
      "loss": 2.7321,
      "theoretical_loss": 3.505961436024127,
      "tokens_seen": 1550003200
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026788365095285863,
      "loss": 2.4762,
      "theoretical_loss": 3.505948448535157,
      "tokens_seen": 1550068736
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026787362086258776,
      "loss": 2.72,
      "theoretical_loss": 3.5059354617490204,
      "tokens_seen": 1550134272
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.000267863590772317,
      "loss": 2.6949,
      "theoretical_loss": 3.5059224756656495,
      "tokens_seen": 1550199808
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002678535606820461,
      "loss": 2.7632,
      "theoretical_loss": 3.505909490284976,
      "tokens_seen": 1550265344
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026784353059177535,
      "loss": 2.6267,
      "theoretical_loss": 3.5058965056069336,
      "tokens_seen": 1550330880
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026783350050150453,
      "loss": 2.6218,
      "theoretical_loss": 3.5058835216314534,
      "tokens_seen": 1550396416
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002678234704112337,
      "loss": 2.7226,
      "theoretical_loss": 3.5058705383584674,
      "tokens_seen": 1550461952
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002678134403209629,
      "loss": 2.5492,
      "theoretical_loss": 3.505857555787909,
      "tokens_seen": 1550527488
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002678034102306921,
      "loss": 2.5225,
      "theoretical_loss": 3.5058445739197097,
      "tokens_seen": 1550593024
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026779338014042126,
      "loss": 2.6679,
      "theoretical_loss": 3.505831592753802,
      "tokens_seen": 1550658560
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026779338014042126,
      "loss": 2.7302,
      "theoretical_loss": 3.5058186122901187,
      "tokens_seen": 1550724096
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002677833500501505,
      "loss": 2.5314,
      "theoretical_loss": 3.5058056325285913,
      "tokens_seen": 1550789632
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002677733199598796,
      "loss": 2.5831,
      "theoretical_loss": 3.5057926534691526,
      "tokens_seen": 1550855168
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026776328986960886,
      "loss": 2.685,
      "theoretical_loss": 3.505779675111735,
      "tokens_seen": 1550920704
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.000267753259779338,
      "loss": 2.6247,
      "theoretical_loss": 3.505766697456271,
      "tokens_seen": 1550986240
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002677432296890672,
      "loss": 2.6779,
      "theoretical_loss": 3.505753720502693,
      "tokens_seen": 1551051776
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002677331995987964,
      "loss": 2.5999,
      "theoretical_loss": 3.5057407442509327,
      "tokens_seen": 1551117312
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002677231695085256,
      "loss": 2.4158,
      "theoretical_loss": 3.5057277687009236,
      "tokens_seen": 1551182848
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026771313941825476,
      "loss": 2.4457,
      "theoretical_loss": 3.5057147938525968,
      "tokens_seen": 1551248384
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.000267703109327984,
      "loss": 2.3523,
      "theoretical_loss": 3.5057018197058856,
      "tokens_seen": 1551313920
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002676930792377131,
      "loss": 2.6527,
      "theoretical_loss": 3.505688846260722,
      "tokens_seen": 1551379456
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026768304914744236,
      "loss": 2.7381,
      "theoretical_loss": 3.5056758735170392,
      "tokens_seen": 1551444992
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002676730190571715,
      "loss": 2.6779,
      "theoretical_loss": 3.5056629014747687,
      "tokens_seen": 1551510528
    },
    {
      "epoch": 5.02,
      "objective/train/docs_used": 1761875,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.585963010787964,
      "objective/train/theoretical_loss": 3.5056531729033273,
      "objective/train/tokens_used": 1572019680,
      "theoretical_loss": 3.5056531729033273,
      "tokens_seen": 1551559680
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002676629889669007,
      "loss": 2.6818,
      "theoretical_loss": 3.5056499301338433,
      "tokens_seen": 1551576064
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002676529588766299,
      "loss": 2.6806,
      "theoretical_loss": 3.5056369594941956,
      "tokens_seen": 1551641600
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002676429287863591,
      "loss": 2.4555,
      "theoretical_loss": 3.505623989555758,
      "tokens_seen": 1551707136
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026763289869608827,
      "loss": 2.4361,
      "theoretical_loss": 3.505611020318463,
      "tokens_seen": 1551772672
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026762286860581745,
      "loss": 2.6085,
      "theoretical_loss": 3.505598051782243,
      "tokens_seen": 1551838208
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026761283851554663,
      "loss": 2.5242,
      "theoretical_loss": 3.5055850839470306,
      "tokens_seen": 1551903744
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026760280842527586,
      "loss": 2.6377,
      "theoretical_loss": 3.5055721168127576,
      "tokens_seen": 1551969280
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.000267592778335005,
      "loss": 2.6599,
      "theoretical_loss": 3.5055591503793577,
      "tokens_seen": 1552034816
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002675827482447342,
      "loss": 2.6179,
      "theoretical_loss": 3.505546184646763,
      "tokens_seen": 1552100352
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026757271815446335,
      "loss": 2.6306,
      "theoretical_loss": 3.5055332196149056,
      "tokens_seen": 1552165888
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002675626880641926,
      "loss": 2.6899,
      "theoretical_loss": 3.505520255283719,
      "tokens_seen": 1552231424
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026755265797392177,
      "loss": 2.6717,
      "theoretical_loss": 3.5055072916531342,
      "tokens_seen": 1552296960
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026754262788365095,
      "loss": 2.8465,
      "theoretical_loss": 3.505494328723085,
      "tokens_seen": 1552362496
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026753259779338013,
      "loss": 2.7475,
      "theoretical_loss": 3.5054813664935036,
      "tokens_seen": 1552428032
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026752256770310937,
      "loss": 2.6695,
      "theoretical_loss": 3.505468404964323,
      "tokens_seen": 1552493568
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002675125376128385,
      "loss": 2.5977,
      "theoretical_loss": 3.5054554441354755,
      "tokens_seen": 1552559104
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026750250752256773,
      "loss": 2.4163,
      "theoretical_loss": 3.5054424840068927,
      "tokens_seen": 1552624640
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026749247743229686,
      "loss": 2.7349,
      "theoretical_loss": 3.505429524578509,
      "tokens_seen": 1552690176
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002674824473420261,
      "loss": 2.5672,
      "theoretical_loss": 3.505416565850256,
      "tokens_seen": 1552755712
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026747241725175527,
      "loss": 2.6153,
      "theoretical_loss": 3.505403607822066,
      "tokens_seen": 1552821248
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026746238716148445,
      "loss": 2.814,
      "theoretical_loss": 3.5053906504938723,
      "tokens_seen": 1552886784
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026745235707121363,
      "loss": 2.6414,
      "theoretical_loss": 3.505377693865608,
      "tokens_seen": 1552952320
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002674423269809428,
      "loss": 2.7226,
      "theoretical_loss": 3.505364737937204,
      "tokens_seen": 1553017856
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.000267432296890672,
      "loss": 2.7976,
      "theoretical_loss": 3.505351782708595,
      "tokens_seen": 1553083392
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026742226680040123,
      "loss": 2.9329,
      "theoretical_loss": 3.5053388281797124,
      "tokens_seen": 1553148928
    },
    {
      "epoch": 5.02,
      "objective/train/docs_used": 1763200,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.717268466949463,
      "objective/train/theoretical_loss": 3.5053291127422055,
      "objective/train/tokens_used": 1573658080,
      "theoretical_loss": 3.5053291127422055,
      "tokens_seen": 1553198080
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026741223671013036,
      "loss": 2.7241,
      "theoretical_loss": 3.505325874350489,
      "tokens_seen": 1553214464
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002674022066198596,
      "loss": 2.6349,
      "theoretical_loss": 3.505312921220858,
      "tokens_seen": 1553280000
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002673921765295888,
      "loss": 2.6627,
      "theoretical_loss": 3.5052999687907516,
      "tokens_seen": 1553345536
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026738214643931796,
      "loss": 2.6706,
      "theoretical_loss": 3.505287017060103,
      "tokens_seen": 1553411072
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026737211634904714,
      "loss": 2.7722,
      "theoretical_loss": 3.5052740660288446,
      "tokens_seen": 1553476608
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002673620862587763,
      "loss": 2.7066,
      "theoretical_loss": 3.5052611156969093,
      "tokens_seen": 1553542144
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002673520561685055,
      "loss": 2.5324,
      "theoretical_loss": 3.50524816606423,
      "tokens_seen": 1553607680
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026734202607823474,
      "loss": 2.6319,
      "theoretical_loss": 3.505235217130738,
      "tokens_seen": 1553673216
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026733199598796386,
      "loss": 2.7366,
      "theoretical_loss": 3.5052222688963686,
      "tokens_seen": 1553738752
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002673219658976931,
      "loss": 2.8268,
      "theoretical_loss": 3.5052093213610522,
      "tokens_seen": 1553804288
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002673119358074222,
      "loss": 2.4775,
      "theoretical_loss": 3.5051963745247234,
      "tokens_seen": 1553869824
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026730190571715146,
      "loss": 2.6987,
      "theoretical_loss": 3.505183428387314,
      "tokens_seen": 1553935360
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026729187562688064,
      "loss": 2.7447,
      "theoretical_loss": 3.5051704829487567,
      "tokens_seen": 1554000896
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002672818455366098,
      "loss": 2.7616,
      "theoretical_loss": 3.5051575382089846,
      "tokens_seen": 1554066432
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.000267271815446339,
      "loss": 2.8065,
      "theoretical_loss": 3.5051445941679304,
      "tokens_seen": 1554131968
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002672617853560682,
      "loss": 2.7473,
      "theoretical_loss": 3.5051316508255272,
      "tokens_seen": 1554197504
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002672517552657974,
      "loss": 2.6068,
      "theoretical_loss": 3.5051187081817075,
      "tokens_seen": 1554263040
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002672417251755266,
      "loss": 2.6967,
      "theoretical_loss": 3.5051057662364045,
      "tokens_seen": 1554328576
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002672316950852558,
      "loss": 2.6442,
      "theoretical_loss": 3.5050928249895508,
      "tokens_seen": 1554394112
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026722166499498496,
      "loss": 2.6032,
      "theoretical_loss": 3.5050798844410793,
      "tokens_seen": 1554459648
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002672116349047142,
      "loss": 2.603,
      "theoretical_loss": 3.5050669445909226,
      "tokens_seen": 1554525184
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002672016048144433,
      "loss": 2.5763,
      "theoretical_loss": 3.505054005439014,
      "tokens_seen": 1554590720
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026719157472417256,
      "loss": 2.8347,
      "theoretical_loss": 3.505041066985286,
      "tokens_seen": 1554656256
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002671815446339017,
      "loss": 2.6721,
      "theoretical_loss": 3.5050281292296726,
      "tokens_seen": 1554721792
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002671715145436309,
      "loss": 2.6613,
      "theoretical_loss": 3.505015192172105,
      "tokens_seen": 1554787328
    },
    {
      "epoch": 5.02,
      "objective/train/docs_used": 1763875,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4189612865448,
      "objective/train/theoretical_loss": 3.5050054898369822,
      "objective/train/tokens_used": 1575296480,
      "theoretical_loss": 3.5050054898369822,
      "tokens_seen": 1554836480
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002671614844533601,
      "loss": 2.6215,
      "theoretical_loss": 3.505002255812517,
      "tokens_seen": 1554852864
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002671514543630893,
      "loss": 2.4481,
      "theoretical_loss": 3.504989320150842,
      "tokens_seen": 1554918400
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026714142427281847,
      "loss": 2.583,
      "theoretical_loss": 3.504976385187012,
      "tokens_seen": 1554983936
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026713139418254765,
      "loss": 2.4658,
      "theoretical_loss": 3.5049634509209606,
      "tokens_seen": 1555049472
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026712136409227683,
      "loss": 2.6812,
      "theoretical_loss": 3.5049505173526203,
      "tokens_seen": 1555115008
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026711133400200606,
      "loss": 2.7396,
      "theoretical_loss": 3.504937584481924,
      "tokens_seen": 1555180544
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002671013039117352,
      "loss": 2.8398,
      "theoretical_loss": 3.5049246523088056,
      "tokens_seen": 1555246080
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002670912738214644,
      "loss": 2.6925,
      "theoretical_loss": 3.5049117208331975,
      "tokens_seen": 1555311616
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026708124373119355,
      "loss": 2.7556,
      "theoretical_loss": 3.504898790055032,
      "tokens_seen": 1555377152
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002670712136409228,
      "loss": 2.7088,
      "theoretical_loss": 3.504885859974243,
      "tokens_seen": 1555442688
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026706118355065197,
      "loss": 2.6861,
      "theoretical_loss": 3.5048729305907633,
      "tokens_seen": 1555508224
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026705115346038115,
      "loss": 2.5036,
      "theoretical_loss": 3.5048600019045257,
      "tokens_seen": 1555573760
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026704112337011033,
      "loss": 2.6393,
      "theoretical_loss": 3.5048470739154634,
      "tokens_seen": 1555639296
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026703109327983957,
      "loss": 2.4422,
      "theoretical_loss": 3.5048341466235096,
      "tokens_seen": 1555704832
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002670210631895687,
      "loss": 2.7981,
      "theoretical_loss": 3.504821220028597,
      "tokens_seen": 1555770368
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026701103309929793,
      "loss": 2.5328,
      "theoretical_loss": 3.504808294130659,
      "tokens_seen": 1555835904
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026700100300902706,
      "loss": 2.7095,
      "theoretical_loss": 3.504795368929628,
      "tokens_seen": 1555901440
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002669909729187563,
      "loss": 2.6313,
      "theoretical_loss": 3.504782444425438,
      "tokens_seen": 1555966976
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026698094282848547,
      "loss": 2.54,
      "theoretical_loss": 3.504769520618021,
      "tokens_seen": 1556032512
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026697091273821465,
      "loss": 2.7396,
      "theoretical_loss": 3.5047565975073116,
      "tokens_seen": 1556098048
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026696088264794383,
      "loss": 2.5316,
      "theoretical_loss": 3.5047436750932417,
      "tokens_seen": 1556163584
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.000266950852557673,
      "loss": 2.624,
      "theoretical_loss": 3.504730753375745,
      "tokens_seen": 1556229120
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002669408224674022,
      "loss": 2.7467,
      "theoretical_loss": 3.504717832354754,
      "tokens_seen": 1556294656
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026693079237713143,
      "loss": 2.5787,
      "theoretical_loss": 3.504704912030202,
      "tokens_seen": 1556360192
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026692076228686056,
      "loss": 2.6299,
      "theoretical_loss": 3.504691992402023,
      "tokens_seen": 1556425728
    },
    {
      "epoch": 5.02,
      "objective/train/docs_used": 1765110,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9499735832214355,
      "objective/train/theoretical_loss": 3.5046823031378422,
      "objective/train/tokens_used": 1576934880,
      "theoretical_loss": 3.5046823031378422,
      "tokens_seen": 1556474880
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002669107321965898,
      "loss": 2.9113,
      "theoretical_loss": 3.5046790734701485,
      "tokens_seen": 1556491264
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.000266900702106319,
      "loss": 2.4546,
      "theoretical_loss": 3.5046661552345135,
      "tokens_seen": 1556556800
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026689067201604816,
      "loss": 2.8173,
      "theoretical_loss": 3.5046532376950497,
      "tokens_seen": 1556622336
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026688064192577734,
      "loss": 2.6191,
      "theoretical_loss": 3.504640320851691,
      "tokens_seen": 1556687872
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002668706118355065,
      "loss": 2.6093,
      "theoretical_loss": 3.504627404704371,
      "tokens_seen": 1556753408
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002668605817452357,
      "loss": 2.7305,
      "theoretical_loss": 3.504614489253022,
      "tokens_seen": 1556818944
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026685055165496494,
      "loss": 2.5957,
      "theoretical_loss": 3.5046015744975776,
      "tokens_seen": 1556884480
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026684052156469406,
      "loss": 2.744,
      "theoretical_loss": 3.504588660437971,
      "tokens_seen": 1556950016
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002668304914744233,
      "loss": 2.8159,
      "theoretical_loss": 3.504575747074135,
      "tokens_seen": 1557015552
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002668204613841524,
      "loss": 2.5389,
      "theoretical_loss": 3.504562834406004,
      "tokens_seen": 1557081088
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026681043129388166,
      "loss": 2.4884,
      "theoretical_loss": 3.50454992243351,
      "tokens_seen": 1557146624
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026680040120361084,
      "loss": 2.7264,
      "theoretical_loss": 3.5045370111565863,
      "tokens_seen": 1557212160
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026679037111334,
      "loss": 2.6224,
      "theoretical_loss": 3.504524100575167,
      "tokens_seen": 1557277696
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002667803410230692,
      "loss": 2.5575,
      "theoretical_loss": 3.504511190689185,
      "tokens_seen": 1557343232
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002667703109327984,
      "loss": 2.5781,
      "theoretical_loss": 3.5044982814985737,
      "tokens_seen": 1557408768
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026676028084252757,
      "loss": 2.6302,
      "theoretical_loss": 3.5044853730032655,
      "tokens_seen": 1557474304
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002667502507522568,
      "loss": 2.5893,
      "theoretical_loss": 3.504472465203195,
      "tokens_seen": 1557539840
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026674022066198593,
      "loss": 2.5045,
      "theoretical_loss": 3.5044595580982945,
      "tokens_seen": 1557605376
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026673019057171516,
      "loss": 2.556,
      "theoretical_loss": 3.504446651688498,
      "tokens_seen": 1557670912
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026672016048144434,
      "loss": 2.8052,
      "theoretical_loss": 3.5044337459737385,
      "tokens_seen": 1557736448
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002667101303911735,
      "loss": 2.6902,
      "theoretical_loss": 3.5044208409539492,
      "tokens_seen": 1557801984
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002667001003009027,
      "loss": 2.7238,
      "theoretical_loss": 3.5044079366290637,
      "tokens_seen": 1557867520
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002666900702106319,
      "loss": 2.6548,
      "theoretical_loss": 3.5043950329990152,
      "tokens_seen": 1557933056
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026668004012036107,
      "loss": 2.6523,
      "theoretical_loss": 3.504382130063737,
      "tokens_seen": 1557998592
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002666700100300903,
      "loss": 2.5946,
      "theoretical_loss": 3.504369227823163,
      "tokens_seen": 1558064128
    },
    {
      "epoch": 5.02,
      "objective/train/docs_used": 1766105,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.555666446685791,
      "objective/train/theoretical_loss": 3.5043595515985917,
      "objective/train/tokens_used": 1578573280,
      "theoretical_loss": 3.5043595515985917,
      "tokens_seen": 1558113280
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026665997993981943,
      "loss": 2.6693,
      "theoretical_loss": 3.504356326277226,
      "tokens_seen": 1558129664
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026664994984954867,
      "loss": 2.6442,
      "theoretical_loss": 3.5043434254258594,
      "tokens_seen": 1558195200
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002666399197592778,
      "loss": 2.6077,
      "theoretical_loss": 3.504330525268997,
      "tokens_seen": 1558260736
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026662988966900703,
      "loss": 2.8557,
      "theoretical_loss": 3.504317625806572,
      "tokens_seen": 1558326272
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002666198595787362,
      "loss": 2.6912,
      "theoretical_loss": 3.5043047270385177,
      "tokens_seen": 1558391808
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002666098294884654,
      "loss": 2.5286,
      "theoretical_loss": 3.5042918289647673,
      "tokens_seen": 1558457344
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026659979939819457,
      "loss": 2.6179,
      "theoretical_loss": 3.5042789315852545,
      "tokens_seen": 1558522880
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026658976930792375,
      "loss": 2.7925,
      "theoretical_loss": 3.504266034899913,
      "tokens_seen": 1558588416
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026657973921765293,
      "loss": 2.6147,
      "theoretical_loss": 3.504253138908676,
      "tokens_seen": 1558653952
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026656970912738217,
      "loss": 2.64,
      "theoretical_loss": 3.504240243611477,
      "tokens_seen": 1558719488
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002665596790371113,
      "loss": 2.7022,
      "theoretical_loss": 3.504227349008249,
      "tokens_seen": 1558785024
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026654964894684053,
      "loss": 2.422,
      "theoretical_loss": 3.5042144550989267,
      "tokens_seen": 1558850560
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002665396188565697,
      "loss": 2.7155,
      "theoretical_loss": 3.5042015618834426,
      "tokens_seen": 1558916096
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002665295887662989,
      "loss": 2.6079,
      "theoretical_loss": 3.50418866936173,
      "tokens_seen": 1558981632
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002665195586760281,
      "loss": 2.6836,
      "theoretical_loss": 3.5041757775337237,
      "tokens_seen": 1559047168
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026650952858575726,
      "loss": 2.4562,
      "theoretical_loss": 3.5041628863993557,
      "tokens_seen": 1559112704
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002664994984954865,
      "loss": 2.7825,
      "theoretical_loss": 3.50414999595856,
      "tokens_seen": 1559178240
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026648946840521567,
      "loss": 2.8305,
      "theoretical_loss": 3.5041371062112705,
      "tokens_seen": 1559243776
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026647943831494485,
      "loss": 2.5331,
      "theoretical_loss": 3.504124217157421,
      "tokens_seen": 1559309312
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026646940822467403,
      "loss": 2.619,
      "theoretical_loss": 3.5041113287969434,
      "tokens_seen": 1559374848
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002664593781344032,
      "loss": 2.7494,
      "theoretical_loss": 3.5040984411297735,
      "tokens_seen": 1559440384
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002664493480441324,
      "loss": 2.7947,
      "theoretical_loss": 3.5040855541558438,
      "tokens_seen": 1559505920
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026643931795386163,
      "loss": 2.5748,
      "theoretical_loss": 3.5040726678750875,
      "tokens_seen": 1559571456
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026642928786359076,
      "loss": 2.9177,
      "theoretical_loss": 3.5040597822874386,
      "tokens_seen": 1559636992
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026641925777332,
      "loss": 2.75,
      "theoretical_loss": 3.504046897392831,
      "tokens_seen": 1559702528
    },
    {
      "epoch": 5.02,
      "objective/train/docs_used": 1766788,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9355313777923584,
      "objective/train/theoretical_loss": 3.5040372341766433,
      "objective/train/tokens_used": 1580211680,
      "theoretical_loss": 3.5040372341766433,
      "tokens_seen": 1559751680
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002664092276830492,
      "loss": 2.9388,
      "theoretical_loss": 3.5040340131911973,
      "tokens_seen": 1559768064
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026639919759277836,
      "loss": 2.8814,
      "theoretical_loss": 3.5040211296824726,
      "tokens_seen": 1559833600
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026638916750250754,
      "loss": 2.5502,
      "theoretical_loss": 3.5040082468665896,
      "tokens_seen": 1559899136
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002663791374122367,
      "loss": 2.7294,
      "theoretical_loss": 3.503995364743482,
      "tokens_seen": 1559964672
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002663691073219659,
      "loss": 2.781,
      "theoretical_loss": 3.5039824833130835,
      "tokens_seen": 1560030208
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026635907723169514,
      "loss": 2.7887,
      "theoretical_loss": 3.503969602575328,
      "tokens_seen": 1560095744
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026634904714142426,
      "loss": 2.626,
      "theoretical_loss": 3.5039567225301482,
      "tokens_seen": 1560161280
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002663390170511535,
      "loss": 2.7191,
      "theoretical_loss": 3.5039438431774794,
      "tokens_seen": 1560226816
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002663289869608826,
      "loss": 2.5363,
      "theoretical_loss": 3.5039309645172536,
      "tokens_seen": 1560292352
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026631895687061186,
      "loss": 2.8227,
      "theoretical_loss": 3.5039180865494055,
      "tokens_seen": 1560357888
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026630892678034104,
      "loss": 2.8357,
      "theoretical_loss": 3.503905209273869,
      "tokens_seen": 1560423424
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002662988966900702,
      "loss": 2.599,
      "theoretical_loss": 3.503892332690577,
      "tokens_seen": 1560488960
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002662888665997994,
      "loss": 2.6265,
      "theoretical_loss": 3.5038794567994636,
      "tokens_seen": 1560554496
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002662788365095286,
      "loss": 2.6774,
      "theoretical_loss": 3.503866581600463,
      "tokens_seen": 1560620032
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026626880641925777,
      "loss": 2.6937,
      "theoretical_loss": 3.503853707093508,
      "tokens_seen": 1560685568
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.000266258776328987,
      "loss": 2.4141,
      "theoretical_loss": 3.5038408332785327,
      "tokens_seen": 1560751104
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026624874623871613,
      "loss": 2.4629,
      "theoretical_loss": 3.5038279601554714,
      "tokens_seen": 1560816640
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026623871614844536,
      "loss": 2.7788,
      "theoretical_loss": 3.503815087724257,
      "tokens_seen": 1560882176
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026622868605817454,
      "loss": 2.4715,
      "theoretical_loss": 3.503802215984824,
      "tokens_seen": 1560947712
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002662186559679037,
      "loss": 2.6418,
      "theoretical_loss": 3.503789344937106,
      "tokens_seen": 1561013248
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002662086258776329,
      "loss": 2.796,
      "theoretical_loss": 3.503776474581036,
      "tokens_seen": 1561078784
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002661985957873621,
      "loss": 2.5447,
      "theoretical_loss": 3.503763604916549,
      "tokens_seen": 1561144320
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026618856569709127,
      "loss": 2.6333,
      "theoretical_loss": 3.503750735943578,
      "tokens_seen": 1561209856
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002661785356068205,
      "loss": 2.657,
      "theoretical_loss": 3.5037378676620574,
      "tokens_seen": 1561275392
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026616850551654963,
      "loss": 2.8328,
      "theoretical_loss": 3.5037250000719204,
      "tokens_seen": 1561340928
    },
    {
      "epoch": 5.02,
      "objective/train/docs_used": 1767472,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.634784460067749,
      "objective/train/theoretical_loss": 3.5037153498329987,
      "objective/train/tokens_used": 1581850080,
      "theoretical_loss": 3.5037153498329987,
      "tokens_seen": 1561390080
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026615847542627887,
      "loss": 2.4677,
      "theoretical_loss": 3.5037121331731016,
      "tokens_seen": 1561406464
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.000266148445336008,
      "loss": 2.7243,
      "theoretical_loss": 3.5036992669655342,
      "tokens_seen": 1561472000
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026613841524573723,
      "loss": 2.814,
      "theoretical_loss": 3.503686401449152,
      "tokens_seen": 1561537536
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002661283851554664,
      "loss": 2.7473,
      "theoretical_loss": 3.5036735366238894,
      "tokens_seen": 1561603072
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002661183550651956,
      "loss": 2.5054,
      "theoretical_loss": 3.50366067248968,
      "tokens_seen": 1561668608
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026610832497492477,
      "loss": 2.3481,
      "theoretical_loss": 3.5036478090464573,
      "tokens_seen": 1561734144
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026609829488465395,
      "loss": 2.5554,
      "theoretical_loss": 3.503634946294156,
      "tokens_seen": 1561799680
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026608826479438313,
      "loss": 2.5268,
      "theoretical_loss": 3.503622084232709,
      "tokens_seen": 1561865216
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026607823470411237,
      "loss": 2.5213,
      "theoretical_loss": 3.5036092228620515,
      "tokens_seen": 1561930752
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002660682046138415,
      "loss": 2.7919,
      "theoretical_loss": 3.5035963621821162,
      "tokens_seen": 1561996288
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026605817452357073,
      "loss": 2.657,
      "theoretical_loss": 3.5035835021928374,
      "tokens_seen": 1562061824
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002660481444332999,
      "loss": 2.4765,
      "theoretical_loss": 3.5035706428941493,
      "tokens_seen": 1562127360
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002660381143430291,
      "loss": 2.5715,
      "theoretical_loss": 3.5035577842859857,
      "tokens_seen": 1562192896
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002660280842527583,
      "loss": 2.3602,
      "theoretical_loss": 3.5035449263682805,
      "tokens_seen": 1562258432
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026601805416248746,
      "loss": 2.741,
      "theoretical_loss": 3.503532069140968,
      "tokens_seen": 1562323968
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026600802407221664,
      "loss": 2.6975,
      "theoretical_loss": 3.5035192126039814,
      "tokens_seen": 1562389504
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026599799398194587,
      "loss": 2.6357,
      "theoretical_loss": 3.5035063567572555,
      "tokens_seen": 1562455040
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.000265987963891675,
      "loss": 2.7784,
      "theoretical_loss": 3.5034935016007243,
      "tokens_seen": 1562520576
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026597793380140423,
      "loss": 2.7628,
      "theoretical_loss": 3.503480647134321,
      "tokens_seen": 1562586112
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026596790371113336,
      "loss": 2.7268,
      "theoretical_loss": 3.50346779335798,
      "tokens_seen": 1562651648
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002659578736208626,
      "loss": 2.5596,
      "theoretical_loss": 3.5034549402716353,
      "tokens_seen": 1562717184
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002659478435305918,
      "loss": 2.7433,
      "theoretical_loss": 3.5034420878752206,
      "tokens_seen": 1562782720
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026593781344032096,
      "loss": 2.7509,
      "theoretical_loss": 3.503429236168671,
      "tokens_seen": 1562848256
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026592778335005014,
      "loss": 2.7708,
      "theoretical_loss": 3.5034163851519198,
      "tokens_seen": 1562913792
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002659177532597794,
      "loss": 2.6534,
      "theoretical_loss": 3.5034035348249013,
      "tokens_seen": 1562979328
    },
    {
      "epoch": 5.02,
      "objective/train/docs_used": 1768239,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8608767986297607,
      "objective/train/theoretical_loss": 3.5033938975322343,
      "objective/train/tokens_used": 1583488480,
      "theoretical_loss": 3.5033938975322343,
      "tokens_seen": 1563028480
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002659077231695085,
      "loss": 2.7638,
      "theoretical_loss": 3.503390685187549,
      "tokens_seen": 1563044864
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026589769307923774,
      "loss": 2.6592,
      "theoretical_loss": 3.503377836239798,
      "tokens_seen": 1563110400
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026588766298896686,
      "loss": 2.7249,
      "theoretical_loss": 3.5033649879815814,
      "tokens_seen": 1563175936
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002658776328986961,
      "loss": 2.6386,
      "theoretical_loss": 3.503352140412833,
      "tokens_seen": 1563241472
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002658676028084253,
      "loss": 2.5453,
      "theoretical_loss": 3.503339293533488,
      "tokens_seen": 1563307008
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026585757271815446,
      "loss": 2.7011,
      "theoretical_loss": 3.5033264473434804,
      "tokens_seen": 1563372544
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026584754262788364,
      "loss": 2.805,
      "theoretical_loss": 3.503313601842744,
      "tokens_seen": 1563438080
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002658375125376128,
      "loss": 2.708,
      "theoretical_loss": 3.5033007570312122,
      "tokens_seen": 1563503616
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.000265827482447342,
      "loss": 2.8063,
      "theoretical_loss": 3.50328791290882,
      "tokens_seen": 1563569152
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026581745235707124,
      "loss": 2.6561,
      "theoretical_loss": 3.5032750694755017,
      "tokens_seen": 1563634688
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026580742226680037,
      "loss": 2.6712,
      "theoretical_loss": 3.5032622267311915,
      "tokens_seen": 1563700224
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002657973921765296,
      "loss": 2.639,
      "theoretical_loss": 3.503249384675822,
      "tokens_seen": 1563765760
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026578736208625873,
      "loss": 2.7569,
      "theoretical_loss": 3.5032365433093293,
      "tokens_seen": 1563831296
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026577733199598797,
      "loss": 2.5356,
      "theoretical_loss": 3.5032237026316473,
      "tokens_seen": 1563896832
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026576730190571715,
      "loss": 2.5125,
      "theoretical_loss": 3.503210862642709,
      "tokens_seen": 1563962368
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026575727181544633,
      "loss": 2.7183,
      "theoretical_loss": 3.503198023342449,
      "tokens_seen": 1564027904
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026574724172517556,
      "loss": 2.4693,
      "theoretical_loss": 3.5031851847308024,
      "tokens_seen": 1564093440
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026573721163490474,
      "loss": 2.8912,
      "theoretical_loss": 3.503172346807703,
      "tokens_seen": 1564158976
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002657271815446339,
      "loss": 2.3473,
      "theoretical_loss": 3.503159509573085,
      "tokens_seen": 1564224512
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002657171514543631,
      "loss": 2.6239,
      "theoretical_loss": 3.5031466730268823,
      "tokens_seen": 1564290048
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002657071213640923,
      "loss": 2.5645,
      "theoretical_loss": 3.5031338371690293,
      "tokens_seen": 1564355584
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026569709127382147,
      "loss": 2.6451,
      "theoretical_loss": 3.5031210019994603,
      "tokens_seen": 1564421120
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002656870611835507,
      "loss": 2.6027,
      "theoretical_loss": 3.5031081675181093,
      "tokens_seen": 1564486656
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026567703109327983,
      "loss": 2.6622,
      "theoretical_loss": 3.503095333724911,
      "tokens_seen": 1564552192
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026566700100300907,
      "loss": 2.5073,
      "theoretical_loss": 3.5030825006198,
      "tokens_seen": 1564617728
    },
    {
      "epoch": 5.02,
      "objective/train/docs_used": 1769768,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.596804141998291,
      "objective/train/theoretical_loss": 3.5030728762424834,
      "objective/train/tokens_used": 1585126880,
      "theoretical_loss": 3.5030728762424834,
      "tokens_seen": 1564666880
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002656569709127382,
      "loss": 2.5982,
      "theoretical_loss": 3.503069668202709,
      "tokens_seen": 1564683264
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026564694082246743,
      "loss": 2.6338,
      "theoretical_loss": 3.5030568364735744,
      "tokens_seen": 1564748800
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002656369107321966,
      "loss": 2.7805,
      "theoretical_loss": 3.5030440054323293,
      "tokens_seen": 1564814336
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002656268806419258,
      "loss": 2.7623,
      "theoretical_loss": 3.503031175078908,
      "tokens_seen": 1564879872
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026561685055165497,
      "loss": 2.4328,
      "theoretical_loss": 3.503018345413245,
      "tokens_seen": 1564945408
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026560682046138415,
      "loss": 2.615,
      "theoretical_loss": 3.503005516435275,
      "tokens_seen": 1565010944
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026559679037111333,
      "loss": 2.6627,
      "theoretical_loss": 3.502992688144932,
      "tokens_seen": 1565076480
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026558676028084257,
      "loss": 2.5018,
      "theoretical_loss": 3.5029798605421503,
      "tokens_seen": 1565142016
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002655767301905717,
      "loss": 2.3945,
      "theoretical_loss": 3.5029670336268643,
      "tokens_seen": 1565207552
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026556670010030093,
      "loss": 2.6775,
      "theoretical_loss": 3.5029542073990085,
      "tokens_seen": 1565273088
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002655566700100301,
      "loss": 2.7513,
      "theoretical_loss": 3.502941381858517,
      "tokens_seen": 1565338624
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002655466399197593,
      "loss": 2.4477,
      "theoretical_loss": 3.502928557005325,
      "tokens_seen": 1565404160
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002655366098294885,
      "loss": 2.8699,
      "theoretical_loss": 3.502915732839366,
      "tokens_seen": 1565469696
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026552657973921766,
      "loss": 2.6521,
      "theoretical_loss": 3.502902909360574,
      "tokens_seen": 1565535232
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026551654964894684,
      "loss": 2.6694,
      "theoretical_loss": 3.502890086568885,
      "tokens_seen": 1565600768
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026550651955867607,
      "loss": 2.5778,
      "theoretical_loss": 3.5028772644642316,
      "tokens_seen": 1565666304
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002654964894684052,
      "loss": 2.7916,
      "theoretical_loss": 3.5028644430465503,
      "tokens_seen": 1565731840
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026548645937813443,
      "loss": 2.5472,
      "theoretical_loss": 3.5028516223157737,
      "tokens_seen": 1565797376
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026547642928786356,
      "loss": 2.6669,
      "theoretical_loss": 3.502838802271837,
      "tokens_seen": 1565862912
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002654663991975928,
      "loss": 2.6802,
      "theoretical_loss": 3.5028259829146746,
      "tokens_seen": 1565928448
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.000265456369107322,
      "loss": 2.548,
      "theoretical_loss": 3.502813164244221,
      "tokens_seen": 1565993984
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026544633901705116,
      "loss": 2.7317,
      "theoretical_loss": 3.5028003462604107,
      "tokens_seen": 1566059520
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.00026543630892678034,
      "loss": 2.7592,
      "theoretical_loss": 3.5027875289631782,
      "tokens_seen": 1566125056
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002654262788365096,
      "loss": 2.2943,
      "theoretical_loss": 3.502774712352458,
      "tokens_seen": 1566190592
    },
    {
      "epoch": 5.02,
      "learning_rate": 0.0002654162487462387,
      "loss": 2.5666,
      "theoretical_loss": 3.5027618964281846,
      "tokens_seen": 1566256128
    },
    {
      "epoch": 5.02,
      "objective/train/docs_used": 1770273,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4401655197143555,
      "objective/train/theoretical_loss": 3.5027522849354202,
      "objective/train/tokens_used": 1586765280,
      "theoretical_loss": 3.5027522849354202,
      "tokens_seen": 1566305280
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026540621865596794,
      "loss": 2.5743,
      "theoretical_loss": 3.5027490811902924,
      "tokens_seen": 1566321664
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026539618856569707,
      "loss": 2.7278,
      "theoretical_loss": 3.5027362666387156,
      "tokens_seen": 1566387200
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002653861584754263,
      "loss": 2.7307,
      "theoretical_loss": 3.5027234527733895,
      "tokens_seen": 1566452736
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002653761283851555,
      "loss": 2.7276,
      "theoretical_loss": 3.502710639594248,
      "tokens_seen": 1566518272
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026536609829488466,
      "loss": 2.4546,
      "theoretical_loss": 3.502697827101226,
      "tokens_seen": 1566583808
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026535606820461384,
      "loss": 2.4774,
      "theoretical_loss": 3.5026850152942584,
      "tokens_seen": 1566649344
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.000265346038114343,
      "loss": 2.6686,
      "theoretical_loss": 3.5026722041732787,
      "tokens_seen": 1566714880
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002653360080240722,
      "loss": 2.5979,
      "theoretical_loss": 3.502659393738223,
      "tokens_seen": 1566780416
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026532597793380144,
      "loss": 2.6913,
      "theoretical_loss": 3.5026465839890237,
      "tokens_seen": 1566845952
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026531594784353057,
      "loss": 2.5873,
      "theoretical_loss": 3.5026337749256182,
      "tokens_seen": 1566911488
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002653059177532598,
      "loss": 2.4014,
      "theoretical_loss": 3.5026209665479384,
      "tokens_seen": 1566977024
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026529588766298893,
      "loss": 2.7904,
      "theoretical_loss": 3.502608158855921,
      "tokens_seen": 1567042560
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026528585757271817,
      "loss": 2.5611,
      "theoretical_loss": 3.5025953518494988,
      "tokens_seen": 1567108096
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026527582748244735,
      "loss": 2.5719,
      "theoretical_loss": 3.502582545528608,
      "tokens_seen": 1567173632
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026526579739217653,
      "loss": 2.5116,
      "theoretical_loss": 3.5025697398931825,
      "tokens_seen": 1567239168
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002652557673019057,
      "loss": 2.4293,
      "theoretical_loss": 3.502556934943157,
      "tokens_seen": 1567304704
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026524573721163494,
      "loss": 2.7416,
      "theoretical_loss": 3.502544130678466,
      "tokens_seen": 1567370240
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026523570712136407,
      "loss": 2.5996,
      "theoretical_loss": 3.5025313270990446,
      "tokens_seen": 1567435776
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002652256770310933,
      "loss": 2.5542,
      "theoretical_loss": 3.5025185242048273,
      "tokens_seen": 1567501312
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026521564694082243,
      "loss": 2.6636,
      "theoretical_loss": 3.5025057219957487,
      "tokens_seen": 1567566848
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026520561685055167,
      "loss": 2.7647,
      "theoretical_loss": 3.5024929204717434,
      "tokens_seen": 1567632384
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026519558676028085,
      "loss": 2.7057,
      "theoretical_loss": 3.502480119632746,
      "tokens_seen": 1567697920
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026518555667001003,
      "loss": 2.5071,
      "theoretical_loss": 3.5024673194786917,
      "tokens_seen": 1567763456
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002651755265797392,
      "loss": 2.8567,
      "theoretical_loss": 3.5024545200095156,
      "tokens_seen": 1567828992
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002651654964894684,
      "loss": 2.7552,
      "theoretical_loss": 3.502441721225151,
      "tokens_seen": 1567894528
    },
    {
      "epoch": 5.03,
      "objective/train/docs_used": 1771462,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.3107831478118896,
      "objective/train/theoretical_loss": 3.5024321225862467,
      "objective/train/tokens_used": 1588403680,
      "theoretical_loss": 3.5024321225862467,
      "tokens_seen": 1567943680
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002651554663991976,
      "loss": 2.8068,
      "theoretical_loss": 3.502428923125534,
      "tokens_seen": 1567960064
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002651454363089268,
      "loss": 2.6769,
      "theoretical_loss": 3.5024161257105986,
      "tokens_seen": 1568025600
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026513540621865594,
      "loss": 2.4889,
      "theoretical_loss": 3.5024033289802796,
      "tokens_seen": 1568091136
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026512537612838517,
      "loss": 2.7356,
      "theoretical_loss": 3.5023905329345117,
      "tokens_seen": 1568156672
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002651153460381143,
      "loss": 2.5249,
      "theoretical_loss": 3.50237773757323,
      "tokens_seen": 1568222208
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026510531594784353,
      "loss": 2.4725,
      "theoretical_loss": 3.502364942896369,
      "tokens_seen": 1568287744
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002650952858575727,
      "loss": 2.6477,
      "theoretical_loss": 3.5023521489038645,
      "tokens_seen": 1568353280
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002650852557673019,
      "loss": 2.5498,
      "theoretical_loss": 3.5023393555956495,
      "tokens_seen": 1568418816
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002650752256770311,
      "loss": 2.7988,
      "theoretical_loss": 3.50232656297166,
      "tokens_seen": 1568484352
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002650651955867603,
      "loss": 2.7056,
      "theoretical_loss": 3.5023137710318304,
      "tokens_seen": 1568549888
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026505516549648944,
      "loss": 2.723,
      "theoretical_loss": 3.5023009797760962,
      "tokens_seen": 1568615424
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002650451354062187,
      "loss": 2.8434,
      "theoretical_loss": 3.5022881892043918,
      "tokens_seen": 1568680960
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002650351053159478,
      "loss": 2.6988,
      "theoretical_loss": 3.5022753993166518,
      "tokens_seen": 1568746496
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026502507522567704,
      "loss": 2.5269,
      "theoretical_loss": 3.502262610112811,
      "tokens_seen": 1568812032
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002650150451354062,
      "loss": 2.7634,
      "theoretical_loss": 3.502249821592805,
      "tokens_seen": 1568877568
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002650050150451354,
      "loss": 2.457,
      "theoretical_loss": 3.502237033756568,
      "tokens_seen": 1568943104
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026499498495486464,
      "loss": 2.5449,
      "theoretical_loss": 3.5022242466040345,
      "tokens_seen": 1569008640
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026498495486459376,
      "loss": 2.5428,
      "theoretical_loss": 3.5022114601351406,
      "tokens_seen": 1569074176
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.000264974924774323,
      "loss": 2.5365,
      "theoretical_loss": 3.5021986743498204,
      "tokens_seen": 1569139712
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002649648946840522,
      "loss": 2.7373,
      "theoretical_loss": 3.502185889248009,
      "tokens_seen": 1569205248
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026495486459378136,
      "loss": 2.6758,
      "theoretical_loss": 3.502173104829641,
      "tokens_seen": 1569270784
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026494483450351054,
      "loss": 2.9497,
      "theoretical_loss": 3.5021603210946517,
      "tokens_seen": 1569336320
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002649348044132398,
      "loss": 2.7386,
      "theoretical_loss": 3.5021475380429763,
      "tokens_seen": 1569401856
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002649247743229689,
      "loss": 2.3083,
      "theoretical_loss": 3.502134755674549,
      "tokens_seen": 1569467392
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026491474423269814,
      "loss": 2.6491,
      "theoretical_loss": 3.502121973989305,
      "tokens_seen": 1569532928
    },
    {
      "epoch": 5.03,
      "objective/train/docs_used": 1772261,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.672778844833374,
      "objective/train/theoretical_loss": 3.5021123881736727,
      "objective/train/tokens_used": 1590042080,
      "theoretical_loss": 3.5021123881736727,
      "tokens_seen": 1569582080
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026490471414242727,
      "loss": 2.5048,
      "theoretical_loss": 3.50210919298718,
      "tokens_seen": 1569598464
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002648946840521565,
      "loss": 2.5785,
      "theoretical_loss": 3.502096412668108,
      "tokens_seen": 1569664000
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002648846539618857,
      "loss": 2.6309,
      "theoretical_loss": 3.5020836330320244,
      "tokens_seen": 1569729536
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026487462387161486,
      "loss": 2.6183,
      "theoretical_loss": 3.5020708540788643,
      "tokens_seen": 1569795072
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026486459378134404,
      "loss": 2.7797,
      "theoretical_loss": 3.502058075808562,
      "tokens_seen": 1569860608
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002648545636910732,
      "loss": 2.5406,
      "theoretical_loss": 3.5020452982210535,
      "tokens_seen": 1569926144
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002648445336008024,
      "loss": 2.641,
      "theoretical_loss": 3.5020325213162735,
      "tokens_seen": 1569991680
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026483450351053164,
      "loss": 2.7582,
      "theoretical_loss": 3.5020197450941564,
      "tokens_seen": 1570057216
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026482447342026077,
      "loss": 2.8391,
      "theoretical_loss": 3.502006969554638,
      "tokens_seen": 1570122752
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026481444332999,
      "loss": 2.5771,
      "theoretical_loss": 3.501994194697653,
      "tokens_seen": 1570188288
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026480441323971913,
      "loss": 2.791,
      "theoretical_loss": 3.5019814205231365,
      "tokens_seen": 1570253824
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026479438314944837,
      "loss": 2.7583,
      "theoretical_loss": 3.501968647031023,
      "tokens_seen": 1570319360
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026478435305917755,
      "loss": 2.6849,
      "theoretical_loss": 3.5019558742212493,
      "tokens_seen": 1570384896
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026477432296890673,
      "loss": 2.4884,
      "theoretical_loss": 3.5019431020937484,
      "tokens_seen": 1570450432
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002647642928786359,
      "loss": 2.7081,
      "theoretical_loss": 3.501930330648457,
      "tokens_seen": 1570515968
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026475426278836514,
      "loss": 2.5406,
      "theoretical_loss": 3.501917559885309,
      "tokens_seen": 1570581504
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026474423269809427,
      "loss": 2.7217,
      "theoretical_loss": 3.50190478980424,
      "tokens_seen": 1570647040
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002647342026078235,
      "loss": 2.8213,
      "theoretical_loss": 3.501892020405185,
      "tokens_seen": 1570712576
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026472417251755263,
      "loss": 2.8029,
      "theoretical_loss": 3.5018792516880795,
      "tokens_seen": 1570778112
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026471414242728187,
      "loss": 2.7009,
      "theoretical_loss": 3.501866483652858,
      "tokens_seen": 1570843648
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026470411233701105,
      "loss": 2.5567,
      "theoretical_loss": 3.5018537162994567,
      "tokens_seen": 1570909184
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026469408224674023,
      "loss": 2.6985,
      "theoretical_loss": 3.5018409496278093,
      "tokens_seen": 1570974720
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002646840521564694,
      "loss": 2.6412,
      "theoretical_loss": 3.501828183637852,
      "tokens_seen": 1571040256
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002646740220661986,
      "loss": 2.6996,
      "theoretical_loss": 3.50181541832952,
      "tokens_seen": 1571105792
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002646639919759278,
      "loss": 2.4996,
      "theoretical_loss": 3.5018026537027476,
      "tokens_seen": 1571171328
    },
    {
      "epoch": 5.03,
      "objective/train/docs_used": 1773327,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5414953231811523,
      "objective/train/theoretical_loss": 3.5017930806799034,
      "objective/train/tokens_used": 1591680480,
      "theoretical_loss": 3.5017930806799034,
      "tokens_seen": 1571220480
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.000264653961885657,
      "loss": 2.4725,
      "theoretical_loss": 3.501789889757471,
      "tokens_seen": 1571236864
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026464393179538614,
      "loss": 2.4687,
      "theoretical_loss": 3.5017771264936246,
      "tokens_seen": 1571302400
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026463390170511537,
      "loss": 2.6159,
      "theoretical_loss": 3.5017643639111435,
      "tokens_seen": 1571367936
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002646238716148445,
      "loss": 2.6949,
      "theoretical_loss": 3.501751602009964,
      "tokens_seen": 1571433472
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026461384152457373,
      "loss": 2.6275,
      "theoretical_loss": 3.5017388407900207,
      "tokens_seen": 1571499008
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002646038114343029,
      "loss": 2.6047,
      "theoretical_loss": 3.5017260802512484,
      "tokens_seen": 1571564544
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002645937813440321,
      "loss": 2.6284,
      "theoretical_loss": 3.501713320393583,
      "tokens_seen": 1571630080
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002645837512537613,
      "loss": 2.8202,
      "theoretical_loss": 3.5017005612169596,
      "tokens_seen": 1571695616
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002645737211634905,
      "loss": 2.5874,
      "theoretical_loss": 3.501687802721313,
      "tokens_seen": 1571761152
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026456369107321964,
      "loss": 2.7119,
      "theoretical_loss": 3.5016750449065785,
      "tokens_seen": 1571826688
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002645536609829489,
      "loss": 2.5645,
      "theoretical_loss": 3.501662287772692,
      "tokens_seen": 1571892224
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.000264543630892678,
      "loss": 2.6312,
      "theoretical_loss": 3.5016495313195883,
      "tokens_seen": 1571957760
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026453360080240724,
      "loss": 2.5627,
      "theoretical_loss": 3.501636775547203,
      "tokens_seen": 1572023296
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002645235707121364,
      "loss": 2.8078,
      "theoretical_loss": 3.5016240204554716,
      "tokens_seen": 1572088832
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002645135406218656,
      "loss": 2.4745,
      "theoretical_loss": 3.5016112660443284,
      "tokens_seen": 1572154368
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002645035105315948,
      "loss": 2.4327,
      "theoretical_loss": 3.501598512313709,
      "tokens_seen": 1572219904
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026449348044132396,
      "loss": 2.6913,
      "theoretical_loss": 3.50158575926355,
      "tokens_seen": 1572285440
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026448345035105314,
      "loss": 2.6209,
      "theoretical_loss": 3.5015730068937856,
      "tokens_seen": 1572350976
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002644734202607824,
      "loss": 2.6321,
      "theoretical_loss": 3.501560255204351,
      "tokens_seen": 1572416512
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002644633901705115,
      "loss": 2.8599,
      "theoretical_loss": 3.501547504195182,
      "tokens_seen": 1572482048
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026445336008024074,
      "loss": 2.852,
      "theoretical_loss": 3.5015347538662143,
      "tokens_seen": 1572547584
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026444332998996987,
      "loss": 2.7477,
      "theoretical_loss": 3.501522004217382,
      "tokens_seen": 1572613120
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002644332998996991,
      "loss": 2.7597,
      "theoretical_loss": 3.5015092552486218,
      "tokens_seen": 1572678656
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002644232698094283,
      "loss": 2.8664,
      "theoretical_loss": 3.5014965069598687,
      "tokens_seen": 1572744192
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026441323971915747,
      "loss": 2.6851,
      "theoretical_loss": 3.501483759351058,
      "tokens_seen": 1572809728
    },
    {
      "epoch": 5.03,
      "objective/train/docs_used": 1773816,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 1.9573558568954468,
      "objective/train/theoretical_loss": 3.501474199090623,
      "objective/train/tokens_used": 1593318880,
      "theoretical_loss": 3.501474199090623,
      "tokens_seen": 1572858880
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026440320962888665,
      "loss": 2.6914,
      "theoretical_loss": 3.501471012422125,
      "tokens_seen": 1572875264
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002643931795386159,
      "loss": 2.5659,
      "theoretical_loss": 3.501458266173005,
      "tokens_seen": 1572940800
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.000264383149448345,
      "loss": 2.8077,
      "theoretical_loss": 3.501445520603634,
      "tokens_seen": 1573006336
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026437311935807424,
      "loss": 2.7236,
      "theoretical_loss": 3.5014327757139467,
      "tokens_seen": 1573071872
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026436308926780337,
      "loss": 2.6091,
      "theoretical_loss": 3.501420031503879,
      "tokens_seen": 1573137408
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002643530591775326,
      "loss": 2.5839,
      "theoretical_loss": 3.5014072879733664,
      "tokens_seen": 1573202944
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002643430290872618,
      "loss": 2.5391,
      "theoretical_loss": 3.501394545122344,
      "tokens_seen": 1573268480
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026433299899699097,
      "loss": 2.8725,
      "theoretical_loss": 3.501381802950748,
      "tokens_seen": 1573334016
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026432296890672015,
      "loss": 2.803,
      "theoretical_loss": 3.501369061458513,
      "tokens_seen": 1573399552
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026431293881644933,
      "loss": 2.682,
      "theoretical_loss": 3.5013563206455744,
      "tokens_seen": 1573465088
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002643029087261785,
      "loss": 2.5005,
      "theoretical_loss": 3.501343580511869,
      "tokens_seen": 1573530624
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026429287863590775,
      "loss": 2.9047,
      "theoretical_loss": 3.5013308410573307,
      "tokens_seen": 1573596160
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002642828485456369,
      "loss": 2.7758,
      "theoretical_loss": 3.501318102281896,
      "tokens_seen": 1573661696
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002642728184553661,
      "loss": 2.6322,
      "theoretical_loss": 3.5013053641855008,
      "tokens_seen": 1573727232
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026426278836509534,
      "loss": 2.8375,
      "theoretical_loss": 3.501292626768079,
      "tokens_seen": 1573792768
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026425275827482447,
      "loss": 2.714,
      "theoretical_loss": 3.5012798900295676,
      "tokens_seen": 1573858304
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002642427281845537,
      "loss": 2.743,
      "theoretical_loss": 3.501267153969902,
      "tokens_seen": 1573923840
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026423269809428283,
      "loss": 2.6854,
      "theoretical_loss": 3.5012544185890166,
      "tokens_seen": 1573989376
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026422266800401207,
      "loss": 2.5074,
      "theoretical_loss": 3.5012416838868483,
      "tokens_seen": 1574054912
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026421263791374125,
      "loss": 2.6326,
      "theoretical_loss": 3.501228949863332,
      "tokens_seen": 1574120448
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026420260782347043,
      "loss": 2.7372,
      "theoretical_loss": 3.5012162165184035,
      "tokens_seen": 1574185984
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002641925777331996,
      "loss": 2.724,
      "theoretical_loss": 3.5012034838519988,
      "tokens_seen": 1574251520
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002641825476429288,
      "loss": 2.444,
      "theoretical_loss": 3.5011907518640526,
      "tokens_seen": 1574317056
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.000264172517552658,
      "loss": 2.8405,
      "theoretical_loss": 3.5011780205545007,
      "tokens_seen": 1574382592
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002641624874623872,
      "loss": 2.5441,
      "theoretical_loss": 3.5011652899232795,
      "tokens_seen": 1574448128
    },
    {
      "epoch": 5.03,
      "objective/train/docs_used": 1774770,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.815545082092285,
      "objective/train/theoretical_loss": 3.501155742394979,
      "objective/train/tokens_used": 1594957280,
      "theoretical_loss": 3.501155742394979,
      "tokens_seen": 1574497280
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026415245737211634,
      "loss": 2.7969,
      "theoretical_loss": 3.5011525599703237,
      "tokens_seen": 1574513664
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026414242728184557,
      "loss": 2.6354,
      "theoretical_loss": 3.5011398306955694,
      "tokens_seen": 1574579200
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002641323971915747,
      "loss": 2.7739,
      "theoretical_loss": 3.5011271020989527,
      "tokens_seen": 1574644736
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026412236710130393,
      "loss": 2.8288,
      "theoretical_loss": 3.501114374180408,
      "tokens_seen": 1574710272
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002641123370110331,
      "loss": 2.5933,
      "theoretical_loss": 3.5011016469398717,
      "tokens_seen": 1574775808
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002641023069207623,
      "loss": 2.6361,
      "theoretical_loss": 3.5010889203772804,
      "tokens_seen": 1574841344
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002640922768304915,
      "loss": 2.5667,
      "theoretical_loss": 3.5010761944925677,
      "tokens_seen": 1574906880
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002640822467402207,
      "loss": 2.7329,
      "theoretical_loss": 3.501063469285671,
      "tokens_seen": 1574972416
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026407221664994984,
      "loss": 2.7016,
      "theoretical_loss": 3.501050744756525,
      "tokens_seen": 1575037952
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002640621865596791,
      "loss": 2.7043,
      "theoretical_loss": 3.501038020905066,
      "tokens_seen": 1575103488
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002640521564694082,
      "loss": 2.8215,
      "theoretical_loss": 3.5010252977312297,
      "tokens_seen": 1575169024
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026404212637913744,
      "loss": 2.5045,
      "theoretical_loss": 3.5010125752349515,
      "tokens_seen": 1575234560
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002640320962888666,
      "loss": 2.7202,
      "theoretical_loss": 3.5009998534161673,
      "tokens_seen": 1575300096
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002640220661985958,
      "loss": 2.7227,
      "theoretical_loss": 3.5009871322748127,
      "tokens_seen": 1575365632
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.000264012036108325,
      "loss": 2.6589,
      "theoretical_loss": 3.500974411810824,
      "tokens_seen": 1575431168
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026400200601805416,
      "loss": 2.7051,
      "theoretical_loss": 3.500961692024136,
      "tokens_seen": 1575496704
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026399197592778334,
      "loss": 2.717,
      "theoretical_loss": 3.500948972914685,
      "tokens_seen": 1575562240
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002639819458375126,
      "loss": 2.7733,
      "theoretical_loss": 3.500936254482407,
      "tokens_seen": 1575627776
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002639719157472417,
      "loss": 2.8105,
      "theoretical_loss": 3.500923536727237,
      "tokens_seen": 1575693312
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026396188565697094,
      "loss": 2.8266,
      "theoretical_loss": 3.5009108196491123,
      "tokens_seen": 1575758848
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026395185556670007,
      "loss": 2.8239,
      "theoretical_loss": 3.500898103247967,
      "tokens_seen": 1575824384
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002639418254764293,
      "loss": 2.6971,
      "theoretical_loss": 3.500885387523738,
      "tokens_seen": 1575889920
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002639317953861585,
      "loss": 2.7925,
      "theoretical_loss": 3.500872672476361,
      "tokens_seen": 1575955456
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026392176529588767,
      "loss": 2.7641,
      "theoretical_loss": 3.500859958105771,
      "tokens_seen": 1576020992
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026391173520561685,
      "loss": 2.9513,
      "theoretical_loss": 3.5008472444119048,
      "tokens_seen": 1576086528
    },
    {
      "epoch": 5.03,
      "objective/train/docs_used": 1774770,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.41874098777771,
      "objective/train/theoretical_loss": 3.500837709585566,
      "objective/train/tokens_used": 1596595680,
      "theoretical_loss": 3.500837709585566,
      "tokens_seen": 1576135680
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002639017051153461,
      "loss": 2.8032,
      "theoretical_loss": 3.5008345313946974,
      "tokens_seen": 1576152064
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002638916750250752,
      "loss": 2.5515,
      "theoretical_loss": 3.5008218190540856,
      "tokens_seen": 1576217600
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026388164493480444,
      "loss": 2.8856,
      "theoretical_loss": 3.5008091073900047,
      "tokens_seen": 1576283136
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026387161484453357,
      "loss": 2.8323,
      "theoretical_loss": 3.5007963964023903,
      "tokens_seen": 1576348672
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002638615847542628,
      "loss": 2.6877,
      "theoretical_loss": 3.500783686091179,
      "tokens_seen": 1576414208
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.000263851554663992,
      "loss": 2.5179,
      "theoretical_loss": 3.500770976456306,
      "tokens_seen": 1576479744
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026384152457372117,
      "loss": 2.9558,
      "theoretical_loss": 3.5007582674977082,
      "tokens_seen": 1576545280
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026383149448345035,
      "loss": 2.8958,
      "theoretical_loss": 3.5007455592153205,
      "tokens_seen": 1576610816
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026382146439317953,
      "loss": 2.7204,
      "theoretical_loss": 3.500732851609079,
      "tokens_seen": 1576676352
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002638114343029087,
      "loss": 2.9708,
      "theoretical_loss": 3.50072014467892,
      "tokens_seen": 1576741888
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026380140421263795,
      "loss": 2.8209,
      "theoretical_loss": 3.5007074384247785,
      "tokens_seen": 1576807424
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002637913741223671,
      "loss": 3.039,
      "theoretical_loss": 3.5006947328465916,
      "tokens_seen": 1576872960
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002637813440320963,
      "loss": 2.7014,
      "theoretical_loss": 3.500682027944295,
      "tokens_seen": 1576938496
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002637713139418255,
      "loss": 2.9119,
      "theoretical_loss": 3.5006693237178244,
      "tokens_seen": 1577004032
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026376128385155467,
      "loss": 2.7541,
      "theoretical_loss": 3.5006566201671157,
      "tokens_seen": 1577069568
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026375125376128385,
      "loss": 3.0195,
      "theoretical_loss": 3.5006439172921047,
      "tokens_seen": 1577135104
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026374122367101303,
      "loss": 2.8242,
      "theoretical_loss": 3.500631215092728,
      "tokens_seen": 1577200640
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002637311935807422,
      "loss": 2.9552,
      "theoretical_loss": 3.5006185135689214,
      "tokens_seen": 1577266176
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026372116349047145,
      "loss": 2.8054,
      "theoretical_loss": 3.5006058127206208,
      "tokens_seen": 1577331712
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002637111334002006,
      "loss": 2.9058,
      "theoretical_loss": 3.5005931125477616,
      "tokens_seen": 1577397248
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002637011033099298,
      "loss": 2.7677,
      "theoretical_loss": 3.5005804130502804,
      "tokens_seen": 1577462784
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026369107321965894,
      "loss": 2.8287,
      "theoretical_loss": 3.500567714228114,
      "tokens_seen": 1577528320
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002636810431293882,
      "loss": 2.7696,
      "theoretical_loss": 3.500555016081197,
      "tokens_seen": 1577593856
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026367101303911736,
      "loss": 2.9985,
      "theoretical_loss": 3.5005423186094666,
      "tokens_seen": 1577659392
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026366098294884654,
      "loss": 2.96,
      "theoretical_loss": 3.5005296218128583,
      "tokens_seen": 1577724928
    },
    {
      "epoch": 5.03,
      "objective/train/docs_used": 1775526,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7164270877838135,
      "objective/train/theoretical_loss": 3.500520099658412,
      "objective/train/tokens_used": 1598234080,
      "theoretical_loss": 3.500520099658412,
      "tokens_seen": 1577774080
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002636509528585757,
      "loss": 2.664,
      "theoretical_loss": 3.500516925691308,
      "tokens_seen": 1577790464
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002636409227683049,
      "loss": 2.6874,
      "theoretical_loss": 3.5005042302447515,
      "tokens_seen": 1577856000
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002636308926780341,
      "loss": 2.6584,
      "theoretical_loss": 3.500491535473126,
      "tokens_seen": 1577921536
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002636208625877633,
      "loss": 2.9151,
      "theoretical_loss": 3.500478841376367,
      "tokens_seen": 1577987072
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026361083249749244,
      "loss": 3.0229,
      "theoretical_loss": 3.50046614795441,
      "tokens_seen": 1578052608
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002636008024072217,
      "loss": 2.8641,
      "theoretical_loss": 3.500453455207192,
      "tokens_seen": 1578118144
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026359077231695086,
      "loss": 2.692,
      "theoretical_loss": 3.500440763134649,
      "tokens_seen": 1578183680
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026358074222668004,
      "loss": 2.9396,
      "theoretical_loss": 3.5004280717367164,
      "tokens_seen": 1578249216
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002635707121364092,
      "loss": 2.8203,
      "theoretical_loss": 3.5004153810133314,
      "tokens_seen": 1578314752
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002635606820461384,
      "loss": 2.5741,
      "theoretical_loss": 3.5004026909644295,
      "tokens_seen": 1578380288
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002635506519558676,
      "loss": 2.9397,
      "theoretical_loss": 3.5003900015899467,
      "tokens_seen": 1578445824
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002635406218655968,
      "loss": 2.8295,
      "theoretical_loss": 3.5003773128898192,
      "tokens_seen": 1578511360
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026353059177532595,
      "loss": 2.7902,
      "theoretical_loss": 3.500364624863984,
      "tokens_seen": 1578576896
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002635205616850552,
      "loss": 2.9844,
      "theoretical_loss": 3.5003519375123764,
      "tokens_seen": 1578642432
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026351053159478436,
      "loss": 2.8751,
      "theoretical_loss": 3.5003392508349327,
      "tokens_seen": 1578707968
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026350050150451354,
      "loss": 2.7126,
      "theoretical_loss": 3.500326564831589,
      "tokens_seen": 1578773504
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002634904714142428,
      "loss": 2.8797,
      "theoretical_loss": 3.500313879502282,
      "tokens_seen": 1578839040
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002634804413239719,
      "loss": 2.9553,
      "theoretical_loss": 3.500301194846947,
      "tokens_seen": 1578904576
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026347041123370114,
      "loss": 2.9701,
      "theoretical_loss": 3.5002885108655217,
      "tokens_seen": 1578970112
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026346038114343027,
      "loss": 2.9054,
      "theoretical_loss": 3.5002758275579406,
      "tokens_seen": 1579035648
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002634503510531595,
      "loss": 2.67,
      "theoretical_loss": 3.5002631449241415,
      "tokens_seen": 1579101184
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002634403209628887,
      "loss": 2.7871,
      "theoretical_loss": 3.5002504629640594,
      "tokens_seen": 1579166720
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026343029087261787,
      "loss": 2.5263,
      "theoretical_loss": 3.5002377816776318,
      "tokens_seen": 1579232256
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026342026078234705,
      "loss": 2.8631,
      "theoretical_loss": 3.500225101064794,
      "tokens_seen": 1579297792
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002634102306920763,
      "loss": 2.9627,
      "theoretical_loss": 3.5002124211254824,
      "tokens_seen": 1579363328
    },
    {
      "epoch": 5.03,
      "objective/train/docs_used": 1776340,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.49752140045166,
      "objective/train/theoretical_loss": 3.500202911612962,
      "objective/train/tokens_used": 1599872480,
      "theoretical_loss": 3.500202911612962,
      "tokens_seen": 1579412480
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002634002006018054,
      "loss": 2.5422,
      "theoretical_loss": 3.500199741859633,
      "tokens_seen": 1579428864
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026339017051153464,
      "loss": 2.7841,
      "theoretical_loss": 3.500187063267183,
      "tokens_seen": 1579494400
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026338014042126377,
      "loss": 2.7422,
      "theoretical_loss": 3.5001743853480685,
      "tokens_seen": 1579559936
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.000263370110330993,
      "loss": 2.7585,
      "theoretical_loss": 3.500161708102225,
      "tokens_seen": 1579625472
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002633600802407222,
      "loss": 2.6565,
      "theoretical_loss": 3.5001490315295896,
      "tokens_seen": 1579691008
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026335005015045137,
      "loss": 3.0777,
      "theoretical_loss": 3.500136355630098,
      "tokens_seen": 1579756544
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026334002006018055,
      "loss": 2.8242,
      "theoretical_loss": 3.5001236804036875,
      "tokens_seen": 1579822080
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026332998996990973,
      "loss": 2.898,
      "theoretical_loss": 3.500111005850293,
      "tokens_seen": 1579887616
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002633199598796389,
      "loss": 2.5527,
      "theoretical_loss": 3.500098331969852,
      "tokens_seen": 1579953152
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026330992978936815,
      "loss": 2.9106,
      "theoretical_loss": 3.5000856587623006,
      "tokens_seen": 1580018688
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002632998996990973,
      "loss": 2.6539,
      "theoretical_loss": 3.5000729862275755,
      "tokens_seen": 1580084224
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002632898696088265,
      "loss": 2.8476,
      "theoretical_loss": 3.500060314365612,
      "tokens_seen": 1580149760
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002632798395185557,
      "loss": 2.7165,
      "theoretical_loss": 3.500047643176347,
      "tokens_seen": 1580215296
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026326980942828487,
      "loss": 2.874,
      "theoretical_loss": 3.5000349726597175,
      "tokens_seen": 1580280832
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026325977933801405,
      "loss": 2.8665,
      "theoretical_loss": 3.5000223028156596,
      "tokens_seen": 1580346368
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026324974924774323,
      "loss": 3.0691,
      "theoretical_loss": 3.5000096336441096,
      "tokens_seen": 1580411904
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002632397191574724,
      "loss": 2.854,
      "theoretical_loss": 3.4999969651450034,
      "tokens_seen": 1580477440
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026322968906720165,
      "loss": 2.6651,
      "theoretical_loss": 3.499984297318278,
      "tokens_seen": 1580542976
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002632196589769308,
      "loss": 2.9626,
      "theoretical_loss": 3.49997163016387,
      "tokens_seen": 1580608512
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026320962888666,
      "loss": 2.6678,
      "theoretical_loss": 3.499958963681715,
      "tokens_seen": 1580674048
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026319959879638914,
      "loss": 2.7437,
      "theoretical_loss": 3.499946297871751,
      "tokens_seen": 1580739584
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002631895687061184,
      "loss": 2.8165,
      "theoretical_loss": 3.499933632733913,
      "tokens_seen": 1580805120
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026317953861584756,
      "loss": 2.8247,
      "theoretical_loss": 3.499920968268138,
      "tokens_seen": 1580870656
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026316950852557674,
      "loss": 2.7621,
      "theoretical_loss": 3.499908304474362,
      "tokens_seen": 1580936192
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002631594784353059,
      "loss": 2.7327,
      "theoretical_loss": 3.4998956413525226,
      "tokens_seen": 1581001728
    },
    {
      "epoch": 5.03,
      "objective/train/docs_used": 1777599,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8258018493652344,
      "objective/train/theoretical_loss": 3.499886144452063,
      "objective/train/tokens_used": 1601510880,
      "theoretical_loss": 3.499886144452063,
      "tokens_seen": 1581050880
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002631494483450351,
      "loss": 2.9461,
      "theoretical_loss": 3.499882978902555,
      "tokens_seen": 1581067264
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002631394182547643,
      "loss": 2.7952,
      "theoretical_loss": 3.499870317124397,
      "tokens_seen": 1581132800
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002631293881644935,
      "loss": 2.8345,
      "theoretical_loss": 3.499857656017984,
      "tokens_seen": 1581198336
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026311935807422264,
      "loss": 2.544,
      "theoretical_loss": 3.499844995583253,
      "tokens_seen": 1581263872
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002631093279839519,
      "loss": 2.7987,
      "theoretical_loss": 3.499832335820141,
      "tokens_seen": 1581329408
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026309929789368106,
      "loss": 2.6151,
      "theoretical_loss": 3.4998196767285834,
      "tokens_seen": 1581394944
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026308926780341024,
      "loss": 2.4125,
      "theoretical_loss": 3.499807018308518,
      "tokens_seen": 1581460480
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002630792377131394,
      "loss": 2.914,
      "theoretical_loss": 3.4997943605598802,
      "tokens_seen": 1581526016
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002630692076228686,
      "loss": 2.8157,
      "theoretical_loss": 3.499781703482607,
      "tokens_seen": 1581591552
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002630591775325978,
      "loss": 3.023,
      "theoretical_loss": 3.4997690470766356,
      "tokens_seen": 1581657088
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.000263049147442327,
      "loss": 2.708,
      "theoretical_loss": 3.4997563913419016,
      "tokens_seen": 1581722624
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026303911735205615,
      "loss": 2.8942,
      "theoretical_loss": 3.4997437362783423,
      "tokens_seen": 1581788160
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002630290872617854,
      "loss": 2.7689,
      "theoretical_loss": 3.499731081885894,
      "tokens_seen": 1581853696
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002630190571715145,
      "loss": 2.7761,
      "theoretical_loss": 3.4997184281644937,
      "tokens_seen": 1581919232
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026300902708124374,
      "loss": 3.1327,
      "theoretical_loss": 3.4997057751140774,
      "tokens_seen": 1581984768
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002629989969909729,
      "loss": 2.7557,
      "theoretical_loss": 3.499693122734582,
      "tokens_seen": 1582050304
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002629889669007021,
      "loss": 2.7901,
      "theoretical_loss": 3.499680471025944,
      "tokens_seen": 1582115840
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002629789368104313,
      "loss": 2.7257,
      "theoretical_loss": 3.4996678199881,
      "tokens_seen": 1582181376
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026296890672016047,
      "loss": 2.7572,
      "theoretical_loss": 3.499655169620987,
      "tokens_seen": 1582246912
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026295887662988965,
      "loss": 2.7508,
      "theoretical_loss": 3.4996425199245422,
      "tokens_seen": 1582312448
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002629488465396189,
      "loss": 2.8917,
      "theoretical_loss": 3.4996298708987013,
      "tokens_seen": 1582377984
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.000262938816449348,
      "loss": 3.0054,
      "theoretical_loss": 3.499617222543401,
      "tokens_seen": 1582443520
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026292878635907725,
      "loss": 2.4528,
      "theoretical_loss": 3.499604574858578,
      "tokens_seen": 1582509056
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026291875626880643,
      "loss": 2.8008,
      "theoretical_loss": 3.499591927844169,
      "tokens_seen": 1582574592
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002629087261785356,
      "loss": 2.8668,
      "theoretical_loss": 3.499579281500112,
      "tokens_seen": 1582640128
    },
    {
      "epoch": 5.03,
      "objective/train/docs_used": 1778167,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9653608798980713,
      "objective/train/theoretical_loss": 3.499569797181948,
      "objective/train/tokens_used": 1603149280,
      "theoretical_loss": 3.499569797181948,
      "tokens_seen": 1582689280
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002628986960882648,
      "loss": 2.8982,
      "theoretical_loss": 3.4995666358263415,
      "tokens_seen": 1582705664
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026288866599799397,
      "loss": 2.8606,
      "theoretical_loss": 3.499553990822796,
      "tokens_seen": 1582771200
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026287863590772315,
      "loss": 2.8867,
      "theoretical_loss": 3.499541346489411,
      "tokens_seen": 1582836736
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002628686058174524,
      "loss": 2.7481,
      "theoretical_loss": 3.499528702826125,
      "tokens_seen": 1582902272
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002628585757271815,
      "loss": 2.69,
      "theoretical_loss": 3.4995160598328727,
      "tokens_seen": 1582967808
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026284854563691075,
      "loss": 2.939,
      "theoretical_loss": 3.499503417509592,
      "tokens_seen": 1583033344
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002628385155466399,
      "loss": 2.9496,
      "theoretical_loss": 3.499490775856219,
      "tokens_seen": 1583098880
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002628284854563691,
      "loss": 3.0367,
      "theoretical_loss": 3.4994781348726915,
      "tokens_seen": 1583164416
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002628184553660983,
      "loss": 2.8122,
      "theoretical_loss": 3.4994654945589447,
      "tokens_seen": 1583229952
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002628084252758275,
      "loss": 2.8815,
      "theoretical_loss": 3.4994528549149173,
      "tokens_seen": 1583295488
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026279839518555666,
      "loss": 2.6773,
      "theoretical_loss": 3.499440215940545,
      "tokens_seen": 1583361024
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002627883650952859,
      "loss": 2.7158,
      "theoretical_loss": 3.499427577635764,
      "tokens_seen": 1583426560
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.000262778335005015,
      "loss": 2.6024,
      "theoretical_loss": 3.499414940000513,
      "tokens_seen": 1583492096
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026276830491474425,
      "loss": 2.7819,
      "theoretical_loss": 3.4994023030347265,
      "tokens_seen": 1583557632
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026275827482447343,
      "loss": 2.7682,
      "theoretical_loss": 3.499389666738343,
      "tokens_seen": 1583623168
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002627482447342026,
      "loss": 2.922,
      "theoretical_loss": 3.499377031111299,
      "tokens_seen": 1583688704
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026273821464393185,
      "loss": 2.8512,
      "theoretical_loss": 3.4993643961535312,
      "tokens_seen": 1583754240
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.000262728184553661,
      "loss": 2.9017,
      "theoretical_loss": 3.4993517618649763,
      "tokens_seen": 1583819776
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002627181544633902,
      "loss": 2.6698,
      "theoretical_loss": 3.499339128245571,
      "tokens_seen": 1583885312
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026270812437311934,
      "loss": 2.8969,
      "theoretical_loss": 3.4993264952952527,
      "tokens_seen": 1583950848
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002626980942828486,
      "loss": 2.7105,
      "theoretical_loss": 3.4993138630139584,
      "tokens_seen": 1584016384
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026268806419257776,
      "loss": 2.7529,
      "theoretical_loss": 3.4993012314016245,
      "tokens_seen": 1584081920
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026267803410230694,
      "loss": 2.9912,
      "theoretical_loss": 3.499288600458188,
      "tokens_seen": 1584147456
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002626680040120361,
      "loss": 2.6269,
      "theoretical_loss": 3.499275970183586,
      "tokens_seen": 1584212992
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002626579739217653,
      "loss": 2.8153,
      "theoretical_loss": 3.499263340577755,
      "tokens_seen": 1584278528
    },
    {
      "epoch": 5.03,
      "objective/train/docs_used": 1779618,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.2659285068511963,
      "objective/train/theoretical_loss": 3.499253868812225,
      "objective/train/tokens_used": 1604787680,
      "theoretical_loss": 3.499253868812225,
      "tokens_seen": 1584327680
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002626479438314945,
      "loss": 2.7086,
      "theoretical_loss": 3.4992507116406326,
      "tokens_seen": 1584344064
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002626379137412237,
      "loss": 2.8713,
      "theoretical_loss": 3.499238083372155,
      "tokens_seen": 1584409600
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026262788365095284,
      "loss": 2.7341,
      "theoretical_loss": 3.49922545577226,
      "tokens_seen": 1584475136
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002626178535606821,
      "loss": 2.5213,
      "theoretical_loss": 3.4992128288408835,
      "tokens_seen": 1584540672
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026260782347041126,
      "loss": 3.0435,
      "theoretical_loss": 3.4992002025779634,
      "tokens_seen": 1584606208
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026259779338014044,
      "loss": 2.5818,
      "theoretical_loss": 3.499187576983436,
      "tokens_seen": 1584671744
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002625877632898696,
      "loss": 2.8307,
      "theoretical_loss": 3.4991749520572384,
      "tokens_seen": 1584737280
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002625777331995988,
      "loss": 2.7998,
      "theoretical_loss": 3.499162327799308,
      "tokens_seen": 1584802816
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.000262567703109328,
      "loss": 2.754,
      "theoretical_loss": 3.4991497042095814,
      "tokens_seen": 1584868352
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002625576730190572,
      "loss": 2.781,
      "theoretical_loss": 3.499137081287996,
      "tokens_seen": 1584933888
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026254764292878635,
      "loss": 2.876,
      "theoretical_loss": 3.4991244590344888,
      "tokens_seen": 1584999424
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002625376128385156,
      "loss": 2.8324,
      "theoretical_loss": 3.4991118374489965,
      "tokens_seen": 1585064960
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002625275827482447,
      "loss": 2.6642,
      "theoretical_loss": 3.4990992165314556,
      "tokens_seen": 1585130496
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026251755265797394,
      "loss": 2.7775,
      "theoretical_loss": 3.499086596281804,
      "tokens_seen": 1585196032
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002625075225677031,
      "loss": 2.9426,
      "theoretical_loss": 3.4990739766999788,
      "tokens_seen": 1585261568
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002624974924774323,
      "loss": 2.7238,
      "theoretical_loss": 3.4990613577859166,
      "tokens_seen": 1585327104
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002624874623871615,
      "loss": 2.7153,
      "theoretical_loss": 3.499048739539554,
      "tokens_seen": 1585392640
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026247743229689067,
      "loss": 2.6911,
      "theoretical_loss": 3.4990361219608292,
      "tokens_seen": 1585458176
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026246740220661985,
      "loss": 2.6232,
      "theoretical_loss": 3.4990235050496787,
      "tokens_seen": 1585523712
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002624573721163491,
      "loss": 2.8488,
      "theoretical_loss": 3.49901088880604,
      "tokens_seen": 1585589248
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002624473420260782,
      "loss": 2.6884,
      "theoretical_loss": 3.4989982732298497,
      "tokens_seen": 1585654784
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026243731193580745,
      "loss": 2.7215,
      "theoretical_loss": 3.4989856583210446,
      "tokens_seen": 1585720320
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026242728184553663,
      "loss": 2.5312,
      "theoretical_loss": 3.4989730440795626,
      "tokens_seen": 1585785856
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002624172517552658,
      "loss": 2.8339,
      "theoretical_loss": 3.49896043050534,
      "tokens_seen": 1585851392
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.000262407221664995,
      "loss": 2.8494,
      "theoretical_loss": 3.498947817598315,
      "tokens_seen": 1585916928
    },
    {
      "epoch": 5.03,
      "objective/train/docs_used": 1782479,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7037224769592285,
      "objective/train/theoretical_loss": 3.4989383583558564,
      "objective/train/tokens_used": 1606426080,
      "theoretical_loss": 3.4989383583558564,
      "tokens_seen": 1585966080
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026239719157472417,
      "loss": 2.5762,
      "theoretical_loss": 3.498935205358424,
      "tokens_seen": 1585982464
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026238716148445335,
      "loss": 2.5875,
      "theoretical_loss": 3.4989225937856046,
      "tokens_seen": 1586048000
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002623771313941826,
      "loss": 2.8368,
      "theoretical_loss": 3.498909982879793,
      "tokens_seen": 1586113536
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002623671013039117,
      "loss": 2.8362,
      "theoretical_loss": 3.4988973726409274,
      "tokens_seen": 1586179072
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026235707121364095,
      "loss": 2.8293,
      "theoretical_loss": 3.4988847630689444,
      "tokens_seen": 1586244608
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002623470411233701,
      "loss": 2.7497,
      "theoretical_loss": 3.4988721541637817,
      "tokens_seen": 1586310144
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002623370110330993,
      "loss": 2.6262,
      "theoretical_loss": 3.498859545925376,
      "tokens_seen": 1586375680
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002623269809428285,
      "loss": 2.7631,
      "theoretical_loss": 3.4988469383536644,
      "tokens_seen": 1586441216
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002623169508525577,
      "loss": 2.7859,
      "theoretical_loss": 3.498834331448585,
      "tokens_seen": 1586506752
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026230692076228686,
      "loss": 2.6273,
      "theoretical_loss": 3.4988217252100737,
      "tokens_seen": 1586572288
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002622968906720161,
      "loss": 2.9191,
      "theoretical_loss": 3.498809119638069,
      "tokens_seen": 1586637824
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002622868605817452,
      "loss": 2.5803,
      "theoretical_loss": 3.4987965147325073,
      "tokens_seen": 1586703360
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026227683049147445,
      "loss": 2.6403,
      "theoretical_loss": 3.4987839104933265,
      "tokens_seen": 1586768896
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002622668004012036,
      "loss": 2.7257,
      "theoretical_loss": 3.498771306920463,
      "tokens_seen": 1586834432
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002622567703109328,
      "loss": 2.9422,
      "theoretical_loss": 3.4987587040138544,
      "tokens_seen": 1586899968
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.000262246740220662,
      "loss": 2.6503,
      "theoretical_loss": 3.4987461017734383,
      "tokens_seen": 1586965504
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002622367101303912,
      "loss": 2.7249,
      "theoretical_loss": 3.4987335001991515,
      "tokens_seen": 1587031040
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026222668004012036,
      "loss": 2.7806,
      "theoretical_loss": 3.4987208992909316,
      "tokens_seen": 1587096576
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026221664994984954,
      "loss": 2.6487,
      "theoretical_loss": 3.498708299048716,
      "tokens_seen": 1587162112
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002622066198595787,
      "loss": 2.6574,
      "theoretical_loss": 3.4986956994724414,
      "tokens_seen": 1587227648
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026219658976930796,
      "loss": 2.7044,
      "theoretical_loss": 3.498683100562046,
      "tokens_seen": 1587293184
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002621865596790371,
      "loss": 2.8528,
      "theoretical_loss": 3.4986705023174665,
      "tokens_seen": 1587358720
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002621765295887663,
      "loss": 2.8764,
      "theoretical_loss": 3.4986579047386406,
      "tokens_seen": 1587424256
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026216649949849545,
      "loss": 2.6997,
      "theoretical_loss": 3.4986453078255053,
      "tokens_seen": 1587489792
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002621564694082247,
      "loss": 2.639,
      "theoretical_loss": 3.498632711577998,
      "tokens_seen": 1587555328
    },
    {
      "epoch": 5.03,
      "objective/train/docs_used": 1787756,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.56807279586792,
      "objective/train/theoretical_loss": 3.498623264829148,
      "objective/train/tokens_used": 1608064480,
      "theoretical_loss": 3.498623264829148,
      "tokens_seen": 1587604480
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026214643931795386,
      "loss": 2.8254,
      "theoretical_loss": 3.498620115996056,
      "tokens_seen": 1587620864
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026213640922768304,
      "loss": 2.7024,
      "theoretical_loss": 3.4986075210796166,
      "tokens_seen": 1587686400
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002621263791374122,
      "loss": 2.7606,
      "theoretical_loss": 3.498594926828618,
      "tokens_seen": 1587751936
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026211634904714146,
      "loss": 2.6963,
      "theoretical_loss": 3.498582333242996,
      "tokens_seen": 1587817472
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002621063189568706,
      "loss": 2.6815,
      "theoretical_loss": 3.498569740322689,
      "tokens_seen": 1587883008
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002620962888665998,
      "loss": 2.7735,
      "theoretical_loss": 3.4985571480676354,
      "tokens_seen": 1587948544
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026208625877632895,
      "loss": 2.8927,
      "theoretical_loss": 3.4985445564777704,
      "tokens_seen": 1588014080
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002620762286860582,
      "loss": 2.7365,
      "theoretical_loss": 3.498531965553033,
      "tokens_seen": 1588079616
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026206619859578737,
      "loss": 2.8605,
      "theoretical_loss": 3.49851937529336,
      "tokens_seen": 1588145152
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026205616850551655,
      "loss": 2.768,
      "theoretical_loss": 3.498506785698689,
      "tokens_seen": 1588210688
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026204613841524573,
      "loss": 2.7571,
      "theoretical_loss": 3.4984941967689576,
      "tokens_seen": 1588276224
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002620361083249749,
      "loss": 2.9175,
      "theoretical_loss": 3.498481608504103,
      "tokens_seen": 1588341760
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002620260782347041,
      "loss": 2.8245,
      "theoretical_loss": 3.4984690209040625,
      "tokens_seen": 1588407296
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002620160481444333,
      "loss": 2.7586,
      "theoretical_loss": 3.498456433968774,
      "tokens_seen": 1588472832
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002620060180541625,
      "loss": 2.6923,
      "theoretical_loss": 3.498443847698174,
      "tokens_seen": 1588538368
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002619959879638917,
      "loss": 2.4443,
      "theoretical_loss": 3.4984312620922013,
      "tokens_seen": 1588603904
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026198595787362087,
      "loss": 2.8155,
      "theoretical_loss": 3.498418677150793,
      "tokens_seen": 1588669440
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026197592778335005,
      "loss": 2.9109,
      "theoretical_loss": 3.4984060928738856,
      "tokens_seen": 1588734976
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002619658976930793,
      "loss": 2.6826,
      "theoretical_loss": 3.498393509261418,
      "tokens_seen": 1588800512
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002619558676028084,
      "loss": 2.8529,
      "theoretical_loss": 3.4983809263133274,
      "tokens_seen": 1588866048
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026194583751253765,
      "loss": 2.6624,
      "theoretical_loss": 3.4983683440295503,
      "tokens_seen": 1588931584
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026193580742226683,
      "loss": 2.7856,
      "theoretical_loss": 3.4983557624100254,
      "tokens_seen": 1588997120
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.000261925777331996,
      "loss": 2.7946,
      "theoretical_loss": 3.4983431814546897,
      "tokens_seen": 1589062656
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002619157472417252,
      "loss": 2.487,
      "theoretical_loss": 3.498330601163481,
      "tokens_seen": 1589128192
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026190571715145437,
      "loss": 2.6616,
      "theoretical_loss": 3.4983180215363365,
      "tokens_seen": 1589193728
    },
    {
      "epoch": 5.03,
      "objective/train/docs_used": 1792626,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8070321083068848,
      "objective/train/theoretical_loss": 3.4983085872517328,
      "objective/train/tokens_used": 1609702880,
      "theoretical_loss": 3.4983085872517328,
      "tokens_seen": 1589242880
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026189568706118355,
      "loss": 2.7286,
      "theoretical_loss": 3.498305442573194,
      "tokens_seen": 1589259264
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002618856569709128,
      "loss": 2.8236,
      "theoretical_loss": 3.498292864273991,
      "tokens_seen": 1589324800
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002618756268806419,
      "loss": 2.7265,
      "theoretical_loss": 3.498280286638665,
      "tokens_seen": 1589390336
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026186559679037115,
      "loss": 2.6342,
      "theoretical_loss": 3.498267709667154,
      "tokens_seen": 1589455872
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002618555667001003,
      "loss": 2.5883,
      "theoretical_loss": 3.498255133359395,
      "tokens_seen": 1589521408
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002618455366098295,
      "loss": 2.8205,
      "theoretical_loss": 3.4982425577153267,
      "tokens_seen": 1589586944
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002618355065195587,
      "loss": 2.4463,
      "theoretical_loss": 3.498229982734885,
      "tokens_seen": 1589652480
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002618254764292879,
      "loss": 2.73,
      "theoretical_loss": 3.498217408418009,
      "tokens_seen": 1589718016
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026181544633901706,
      "loss": 2.6927,
      "theoretical_loss": 3.4982048347646355,
      "tokens_seen": 1589783552
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002618054162487463,
      "loss": 2.8462,
      "theoretical_loss": 3.4981922617747023,
      "tokens_seen": 1589849088
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002617953861584754,
      "loss": 2.648,
      "theoretical_loss": 3.4981796894481474,
      "tokens_seen": 1589914624
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026178535606820465,
      "loss": 2.6182,
      "theoretical_loss": 3.498167117784908,
      "tokens_seen": 1589980160
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002617753259779338,
      "loss": 2.7333,
      "theoretical_loss": 3.498154546784922,
      "tokens_seen": 1590045696
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.000261765295887663,
      "loss": 2.7277,
      "theoretical_loss": 3.498141976448127,
      "tokens_seen": 1590111232
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002617552657973922,
      "loss": 2.7771,
      "theoretical_loss": 3.4981294067744617,
      "tokens_seen": 1590176768
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002617452357071214,
      "loss": 2.7176,
      "theoretical_loss": 3.498116837763862,
      "tokens_seen": 1590242304
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026173520561685056,
      "loss": 2.5995,
      "theoretical_loss": 3.4981042694162663,
      "tokens_seen": 1590307840
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026172517552657974,
      "loss": 2.8918,
      "theoretical_loss": 3.498091701731613,
      "tokens_seen": 1590373376
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002617151454363089,
      "loss": 2.9067,
      "theoretical_loss": 3.4980791347098386,
      "tokens_seen": 1590438912
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026170511534603816,
      "loss": 2.637,
      "theoretical_loss": 3.498066568350882,
      "tokens_seen": 1590504448
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002616950852557673,
      "loss": 2.973,
      "theoretical_loss": 3.49805400265468,
      "tokens_seen": 1590569984
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002616850551654965,
      "loss": 2.7506,
      "theoretical_loss": 3.498041437621171,
      "tokens_seen": 1590635520
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026167502507522565,
      "loss": 2.6337,
      "theoretical_loss": 3.498028873250292,
      "tokens_seen": 1590701056
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002616649949849549,
      "loss": 2.7196,
      "theoretical_loss": 3.498016309541982,
      "tokens_seen": 1590766592
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026165496489468406,
      "loss": 2.7316,
      "theoretical_loss": 3.498003746496178,
      "tokens_seen": 1590832128
    },
    {
      "epoch": 5.03,
      "objective/train/docs_used": 1797804,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7346248626708984,
      "objective/train/theoretical_loss": 3.497994324646557,
      "objective/train/tokens_used": 1611341280,
      "theoretical_loss": 3.497994324646557,
      "tokens_seen": 1590881280
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026164493480441324,
      "loss": 2.7022,
      "theoretical_loss": 3.4979911841128173,
      "tokens_seen": 1590897664
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002616349047141424,
      "loss": 2.4981,
      "theoretical_loss": 3.4979786223918388,
      "tokens_seen": 1590963200
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026162487462387166,
      "loss": 2.7815,
      "theoretical_loss": 3.4979660613331793,
      "tokens_seen": 1591028736
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002616148445336008,
      "loss": 2.7909,
      "theoretical_loss": 3.4979535009367773,
      "tokens_seen": 1591094272
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026160481444333,
      "loss": 2.8896,
      "theoretical_loss": 3.4979409412025695,
      "tokens_seen": 1591159808
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026159478435305915,
      "loss": 2.8332,
      "theoretical_loss": 3.497928382130495,
      "tokens_seen": 1591225344
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002615847542627884,
      "loss": 2.6551,
      "theoretical_loss": 3.4979158237204913,
      "tokens_seen": 1591290880
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026157472417251757,
      "loss": 2.8388,
      "theoretical_loss": 3.4979032659724956,
      "tokens_seen": 1591356416
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026156469408224675,
      "loss": 2.8684,
      "theoretical_loss": 3.4978907088864464,
      "tokens_seen": 1591421952
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026155466399197593,
      "loss": 2.8727,
      "theoretical_loss": 3.4978781524622815,
      "tokens_seen": 1591487488
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002615446339017051,
      "loss": 2.7349,
      "theoretical_loss": 3.4978655966999384,
      "tokens_seen": 1591553024
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002615346038114343,
      "loss": 2.8585,
      "theoretical_loss": 3.497853041599355,
      "tokens_seen": 1591618560
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002615245737211635,
      "loss": 2.9453,
      "theoretical_loss": 3.4978404871604702,
      "tokens_seen": 1591684096
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026151454363089265,
      "loss": 2.53,
      "theoretical_loss": 3.4978279333832205,
      "tokens_seen": 1591749632
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002615045135406219,
      "loss": 2.663,
      "theoretical_loss": 3.497815380267544,
      "tokens_seen": 1591815168
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.000261494483450351,
      "loss": 2.6538,
      "theoretical_loss": 3.4978028278133797,
      "tokens_seen": 1591880704
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026148445336008025,
      "loss": 2.9559,
      "theoretical_loss": 3.497790276020664,
      "tokens_seen": 1591946240
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026147442326980943,
      "loss": 2.7134,
      "theoretical_loss": 3.4977777248893354,
      "tokens_seen": 1592011776
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002614643931795386,
      "loss": 2.4606,
      "theoretical_loss": 3.4977651744193325,
      "tokens_seen": 1592077312
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002614543630892678,
      "loss": 2.692,
      "theoretical_loss": 3.4977526246105928,
      "tokens_seen": 1592142848
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026144433299899703,
      "loss": 2.8035,
      "theoretical_loss": 3.497740075463054,
      "tokens_seen": 1592208384
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026143430290872616,
      "loss": 2.7924,
      "theoretical_loss": 3.4977275269766537,
      "tokens_seen": 1592273920
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002614242728184554,
      "loss": 2.8615,
      "theoretical_loss": 3.497714979151331,
      "tokens_seen": 1592339456
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002614142427281845,
      "loss": 2.8544,
      "theoretical_loss": 3.4977024319870234,
      "tokens_seen": 1592404992
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026140421263791375,
      "loss": 2.6875,
      "theoretical_loss": 3.497689885483668,
      "tokens_seen": 1592470528
    },
    {
      "epoch": 5.03,
      "objective/train/docs_used": 1800873,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.813413143157959,
      "objective/train/theoretical_loss": 3.4976804760398648,
      "objective/train/tokens_used": 1612979680,
      "theoretical_loss": 3.4976804760398648,
      "tokens_seen": 1592519680
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026139418254764293,
      "loss": 2.5509,
      "theoretical_loss": 3.497677339641204,
      "tokens_seen": 1592536064
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002613841524573721,
      "loss": 2.6167,
      "theoretical_loss": 3.4976647944595687,
      "tokens_seen": 1592601600
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002613741223671013,
      "loss": 2.8014,
      "theoretical_loss": 3.4976522499387004,
      "tokens_seen": 1592667136
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002613640922768305,
      "loss": 2.6458,
      "theoretical_loss": 3.4976397060785365,
      "tokens_seen": 1592732672
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026135406218655966,
      "loss": 2.7171,
      "theoretical_loss": 3.497627162879016,
      "tokens_seen": 1592798208
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002613440320962889,
      "loss": 2.8442,
      "theoretical_loss": 3.4976146203400766,
      "tokens_seen": 1592863744
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.000261334002006018,
      "loss": 2.9746,
      "theoretical_loss": 3.4976020784616555,
      "tokens_seen": 1592929280
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026132397191574726,
      "loss": 2.8905,
      "theoretical_loss": 3.497589537243692,
      "tokens_seen": 1592994816
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002613139418254764,
      "loss": 2.7806,
      "theoretical_loss": 3.4975769966861234,
      "tokens_seen": 1593060352
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002613039117352056,
      "loss": 2.7803,
      "theoretical_loss": 3.497564456788888,
      "tokens_seen": 1593125888
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002612938816449348,
      "loss": 2.7665,
      "theoretical_loss": 3.497551917551924,
      "tokens_seen": 1593191424
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.000261283851554664,
      "loss": 2.6711,
      "theoretical_loss": 3.4975393789751688,
      "tokens_seen": 1593256960
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026127382146439316,
      "loss": 2.6177,
      "theoretical_loss": 3.497526841058561,
      "tokens_seen": 1593322496
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002612637913741224,
      "loss": 2.5415,
      "theoretical_loss": 3.4975143038020384,
      "tokens_seen": 1593388032
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002612537612838516,
      "loss": 2.8254,
      "theoretical_loss": 3.4975017672055397,
      "tokens_seen": 1593453568
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026124373119358076,
      "loss": 2.6499,
      "theoretical_loss": 3.497489231269003,
      "tokens_seen": 1593519104
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026123370110330994,
      "loss": 2.8149,
      "theoretical_loss": 3.4974766959923658,
      "tokens_seen": 1593584640
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002612236710130391,
      "loss": 2.6921,
      "theoretical_loss": 3.497464161375566,
      "tokens_seen": 1593650176
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026121364092276836,
      "loss": 2.878,
      "theoretical_loss": 3.4974516274185428,
      "tokens_seen": 1593715712
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002612036108324975,
      "loss": 2.6749,
      "theoretical_loss": 3.497439094121234,
      "tokens_seen": 1593781248
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002611935807422267,
      "loss": 2.6689,
      "theoretical_loss": 3.497426561483577,
      "tokens_seen": 1593846784
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026118355065195585,
      "loss": 2.6218,
      "theoretical_loss": 3.497414029505511,
      "tokens_seen": 1593912320
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002611735205616851,
      "loss": 2.685,
      "theoretical_loss": 3.4974014981869734,
      "tokens_seen": 1593977856
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026116349047141426,
      "loss": 2.7086,
      "theoretical_loss": 3.497388967527902,
      "tokens_seen": 1594043392
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026115346038114344,
      "loss": 2.5245,
      "theoretical_loss": 3.4973764375282363,
      "tokens_seen": 1594108928
    },
    {
      "epoch": 5.03,
      "objective/train/docs_used": 1801349,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.83113694190979,
      "objective/train/theoretical_loss": 3.4973670404611843,
      "objective/train/tokens_used": 1614618080,
      "theoretical_loss": 3.4973670404611843,
      "tokens_seen": 1594158080
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002611434302908726,
      "loss": 2.5249,
      "theoretical_loss": 3.4973639081879138,
      "tokens_seen": 1594174464
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026113340020060186,
      "loss": 2.8712,
      "theoretical_loss": 3.4973513795068727,
      "tokens_seen": 1594240000
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.000261123370110331,
      "loss": 2.8316,
      "theoretical_loss": 3.4973388514850514,
      "tokens_seen": 1594305536
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002611133400200602,
      "loss": 2.7948,
      "theoretical_loss": 3.4973263241223873,
      "tokens_seen": 1594371072
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026110330992978935,
      "loss": 2.8875,
      "theoretical_loss": 3.49731379741882,
      "tokens_seen": 1594436608
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002610932798395186,
      "loss": 2.7071,
      "theoretical_loss": 3.4973012713742864,
      "tokens_seen": 1594502144
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026108324974924777,
      "loss": 2.6609,
      "theoretical_loss": 3.4972887459887256,
      "tokens_seen": 1594567680
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026107321965897695,
      "loss": 2.9401,
      "theoretical_loss": 3.4972762212620756,
      "tokens_seen": 1594633216
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026106318956870613,
      "loss": 2.6306,
      "theoretical_loss": 3.4972636971942745,
      "tokens_seen": 1594698752
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002610531594784353,
      "loss": 2.9314,
      "theoretical_loss": 3.497251173785261,
      "tokens_seen": 1594764288
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002610431293881645,
      "loss": 2.7628,
      "theoretical_loss": 3.497238651034973,
      "tokens_seen": 1594829824
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002610330992978937,
      "loss": 2.6542,
      "theoretical_loss": 3.497226128943349,
      "tokens_seen": 1594895360
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026102306920762285,
      "loss": 2.7527,
      "theoretical_loss": 3.497213607510327,
      "tokens_seen": 1594960896
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002610130391173521,
      "loss": 2.9548,
      "theoretical_loss": 3.4972010867358456,
      "tokens_seen": 1595026432
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002610030090270812,
      "loss": 2.7296,
      "theoretical_loss": 3.4971885666198426,
      "tokens_seen": 1595091968
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026099297893681045,
      "loss": 2.8855,
      "theoretical_loss": 3.497176047162257,
      "tokens_seen": 1595157504
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026098294884653963,
      "loss": 2.5206,
      "theoretical_loss": 3.4971635283630267,
      "tokens_seen": 1595223040
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002609729187562688,
      "loss": 2.8073,
      "theoretical_loss": 3.4971510102220904,
      "tokens_seen": 1595288576
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.000260962888665998,
      "loss": 2.5685,
      "theoretical_loss": 3.497138492739386,
      "tokens_seen": 1595354112
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026095285857572723,
      "loss": 2.7139,
      "theoretical_loss": 3.4971259759148525,
      "tokens_seen": 1595419648
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026094282848545636,
      "loss": 2.6919,
      "theoretical_loss": 3.4971134597484275,
      "tokens_seen": 1595485184
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002609327983951856,
      "loss": 2.7325,
      "theoretical_loss": 3.4971009442400494,
      "tokens_seen": 1595550720
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002609227683049147,
      "loss": 2.6534,
      "theoretical_loss": 3.497088429389657,
      "tokens_seen": 1595616256
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026091273821464395,
      "loss": 2.6995,
      "theoretical_loss": 3.497075915197189,
      "tokens_seen": 1595681792
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026090270812437313,
      "loss": 2.8369,
      "theoretical_loss": 3.497063401662583,
      "tokens_seen": 1595747328
    },
    {
      "epoch": 5.03,
      "objective/train/docs_used": 1802636,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.037900686264038,
      "objective/train/theoretical_loss": 3.4970540169433133,
      "objective/train/tokens_used": 1616256480,
      "theoretical_loss": 3.4970540169433133,
      "tokens_seen": 1595796480
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002608926780341023,
      "loss": 2.5637,
      "theoretical_loss": 3.4970508887857776,
      "tokens_seen": 1595812864
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002608826479438315,
      "loss": 2.7449,
      "theoretical_loss": 3.4970383765667115,
      "tokens_seen": 1595878400
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002608726178535607,
      "loss": 2.7603,
      "theoretical_loss": 3.497025865005323,
      "tokens_seen": 1595943936
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026086258776328986,
      "loss": 2.7163,
      "theoretical_loss": 3.4970133541015507,
      "tokens_seen": 1596009472
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002608525576730191,
      "loss": 2.7163,
      "theoretical_loss": 3.497000843855332,
      "tokens_seen": 1596075008
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002608425275827482,
      "loss": 2.876,
      "theoretical_loss": 3.4969883342666073,
      "tokens_seen": 1596140544
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026083249749247746,
      "loss": 2.9052,
      "theoretical_loss": 3.496975825335313,
      "tokens_seen": 1596206080
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002608224674022066,
      "loss": 2.751,
      "theoretical_loss": 3.496963317061389,
      "tokens_seen": 1596271616
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002608124373119358,
      "loss": 2.7262,
      "theoretical_loss": 3.496950809444773,
      "tokens_seen": 1596337152
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.000260802407221665,
      "loss": 2.8612,
      "theoretical_loss": 3.496938302485404,
      "tokens_seen": 1596402688
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002607923771313942,
      "loss": 2.6211,
      "theoretical_loss": 3.49692579618322,
      "tokens_seen": 1596468224
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026078234704112336,
      "loss": 2.7479,
      "theoretical_loss": 3.49691329053816,
      "tokens_seen": 1596533760
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002607723169508526,
      "loss": 2.7243,
      "theoretical_loss": 3.496900785550162,
      "tokens_seen": 1596599296
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002607622868605817,
      "loss": 2.7206,
      "theoretical_loss": 3.4968882812191646,
      "tokens_seen": 1596664832
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026075225677031096,
      "loss": 2.637,
      "theoretical_loss": 3.4968757775451063,
      "tokens_seen": 1596730368
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002607422266800401,
      "loss": 2.6786,
      "theoretical_loss": 3.496863274527926,
      "tokens_seen": 1596795904
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002607321965897693,
      "loss": 2.7166,
      "theoretical_loss": 3.4968507721675617,
      "tokens_seen": 1596861440
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002607221664994985,
      "loss": 2.801,
      "theoretical_loss": 3.4968382704639525,
      "tokens_seen": 1596926976
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002607121364092277,
      "loss": 2.6867,
      "theoretical_loss": 3.4968257694170366,
      "tokens_seen": 1596992512
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026070210631895687,
      "loss": 2.7219,
      "theoretical_loss": 3.4968132690267524,
      "tokens_seen": 1597058048
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026069207622868605,
      "loss": 2.7463,
      "theoretical_loss": 3.4968007692930394,
      "tokens_seen": 1597123584
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026068204613841523,
      "loss": 2.6852,
      "theoretical_loss": 3.496788270215835,
      "tokens_seen": 1597189120
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026067201604814446,
      "loss": 2.6713,
      "theoretical_loss": 3.496775771795078,
      "tokens_seen": 1597254656
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002606619859578736,
      "loss": 2.618,
      "theoretical_loss": 3.496763274030708,
      "tokens_seen": 1597320192
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002606519558676028,
      "loss": 2.7221,
      "theoretical_loss": 3.496750776922662,
      "tokens_seen": 1597385728
    },
    {
      "epoch": 5.03,
      "objective/train/docs_used": 1803304,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7198164463043213,
      "objective/train/theoretical_loss": 3.4967414045223038,
      "objective/train/tokens_used": 1617894880,
      "theoretical_loss": 3.4967414045223038,
      "tokens_seen": 1597434880
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026064192577733195,
      "loss": 2.8111,
      "theoretical_loss": 3.49673828047088,
      "tokens_seen": 1597451264
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002606318956870612,
      "loss": 2.7815,
      "theoretical_loss": 3.4967257846752995,
      "tokens_seen": 1597516800
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026062186559679037,
      "loss": 2.6373,
      "theoretical_loss": 3.4967132895358604,
      "tokens_seen": 1597582336
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026061183550651955,
      "loss": 2.8427,
      "theoretical_loss": 3.4967007950525,
      "tokens_seen": 1597647872
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026060180541624873,
      "loss": 2.7914,
      "theoretical_loss": 3.4966883012251584,
      "tokens_seen": 1597713408
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026059177532597797,
      "loss": 2.6316,
      "theoretical_loss": 3.496675808053773,
      "tokens_seen": 1597778944
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002605817452357071,
      "loss": 2.8277,
      "theoretical_loss": 3.4966633155382825,
      "tokens_seen": 1597844480
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026057171514543633,
      "loss": 2.7669,
      "theoretical_loss": 3.4966508236786265,
      "tokens_seen": 1597910016
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026056168505516546,
      "loss": 2.4671,
      "theoretical_loss": 3.4966383324747428,
      "tokens_seen": 1597975552
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002605516549648947,
      "loss": 2.7547,
      "theoretical_loss": 3.49662584192657,
      "tokens_seen": 1598041088
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026054162487462387,
      "loss": 2.4883,
      "theoretical_loss": 3.496613352034048,
      "tokens_seen": 1598106624
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026053159478435305,
      "loss": 2.676,
      "theoretical_loss": 3.496600862797114,
      "tokens_seen": 1598172160
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026052156469408223,
      "loss": 2.6167,
      "theoretical_loss": 3.4965883742157082,
      "tokens_seen": 1598237696
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002605115346038114,
      "loss": 2.9133,
      "theoretical_loss": 3.496575886289768,
      "tokens_seen": 1598303232
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026050150451354065,
      "loss": 2.7645,
      "theoretical_loss": 3.4965633990192324,
      "tokens_seen": 1598368768
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026049147442326983,
      "loss": 2.6988,
      "theoretical_loss": 3.496550912404041,
      "tokens_seen": 1598434304
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.000260481444332999,
      "loss": 2.688,
      "theoretical_loss": 3.4965384264441313,
      "tokens_seen": 1598499840
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002604714142427282,
      "loss": 2.9453,
      "theoretical_loss": 3.496525941139443,
      "tokens_seen": 1598565376
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026046138415245743,
      "loss": 2.7754,
      "theoretical_loss": 3.4965134564899145,
      "tokens_seen": 1598630912
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026045135406218656,
      "loss": 2.8114,
      "theoretical_loss": 3.4965009724954843,
      "tokens_seen": 1598696448
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002604413239719158,
      "loss": 2.7888,
      "theoretical_loss": 3.496488489156092,
      "tokens_seen": 1598761984
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002604312938816449,
      "loss": 3.0041,
      "theoretical_loss": 3.496476006471675,
      "tokens_seen": 1598827520
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026042126379137415,
      "loss": 2.6466,
      "theoretical_loss": 3.4964635244421736,
      "tokens_seen": 1598893056
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026041123370110333,
      "loss": 2.5899,
      "theoretical_loss": 3.496451043067526,
      "tokens_seen": 1598958592
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002604012036108325,
      "loss": 2.84,
      "theoretical_loss": 3.49643856234767,
      "tokens_seen": 1599024128
    },
    {
      "epoch": 5.03,
      "objective/train/docs_used": 1804585,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.1171164512634277,
      "objective/train/theoretical_loss": 3.4964292022374495,
      "objective/train/tokens_used": 1619533280,
      "theoretical_loss": 3.4964292022374495,
      "tokens_seen": 1599073280
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002603911735205617,
      "loss": 2.8474,
      "theoretical_loss": 3.496426082282546,
      "tokens_seen": 1599089664
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002603811434302909,
      "loss": 2.5638,
      "theoretical_loss": 3.4964136028720922,
      "tokens_seen": 1599155200
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.00026037111334002006,
      "loss": 2.6904,
      "theoretical_loss": 3.4964011241162476,
      "tokens_seen": 1599220736
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0002603610832497493,
      "loss": 2.6742,
      "theoretical_loss": 3.4963886460149505,
      "tokens_seen": 1599286272
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002603510531594784,
      "loss": 2.7654,
      "theoretical_loss": 3.49637616856814,
      "tokens_seen": 1599351808
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00026034102306920766,
      "loss": 2.6412,
      "theoretical_loss": 3.4963636917757555,
      "tokens_seen": 1599417344
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002603309929789368,
      "loss": 2.8438,
      "theoretical_loss": 3.4963512156377345,
      "tokens_seen": 1599482880
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.000260320962888666,
      "loss": 2.7242,
      "theoretical_loss": 3.4963387401540174,
      "tokens_seen": 1599548416
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002603109327983952,
      "loss": 2.8938,
      "theoretical_loss": 3.4963262653245426,
      "tokens_seen": 1599613952
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002603009027081244,
      "loss": 2.7322,
      "theoretical_loss": 3.496313791149248,
      "tokens_seen": 1599679488
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00026029087261785356,
      "loss": 2.489,
      "theoretical_loss": 3.4963013176280744,
      "tokens_seen": 1599745024
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002602808425275828,
      "loss": 2.7086,
      "theoretical_loss": 3.496288844760959,
      "tokens_seen": 1599810560
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002602708124373119,
      "loss": 2.6735,
      "theoretical_loss": 3.4962763725478414,
      "tokens_seen": 1599876096
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00026026078234704116,
      "loss": 2.527,
      "theoretical_loss": 3.4962639009886605,
      "tokens_seen": 1599941632
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002602507522567703,
      "loss": 2.7704,
      "theoretical_loss": 3.4962514300833556,
      "tokens_seen": 1600007168
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002602407221664995,
      "loss": 2.7558,
      "theoretical_loss": 3.4962389598318646,
      "tokens_seen": 1600072704
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002602306920762287,
      "loss": 2.4681,
      "theoretical_loss": 3.496226490234127,
      "tokens_seen": 1600138240
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002602206619859579,
      "loss": 2.437,
      "theoretical_loss": 3.4962140212900827,
      "tokens_seen": 1600203776
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00026021063189568707,
      "loss": 2.6458,
      "theoretical_loss": 3.4962015529996693,
      "tokens_seen": 1600269312
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00026020060180541625,
      "loss": 2.7163,
      "theoretical_loss": 3.4961890853628264,
      "tokens_seen": 1600334848
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00026019057171514543,
      "loss": 2.6421,
      "theoretical_loss": 3.4961766183794922,
      "tokens_seen": 1600400384
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00026018054162487466,
      "loss": 2.5906,
      "theoretical_loss": 3.496164152049607,
      "tokens_seen": 1600465920
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002601705115346038,
      "loss": 2.7296,
      "theoretical_loss": 3.496151686373109,
      "tokens_seen": 1600531456
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.000260160481444333,
      "loss": 2.9529,
      "theoretical_loss": 3.496139221349937,
      "tokens_seen": 1600596992
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00026015045135406215,
      "loss": 2.7484,
      "theoretical_loss": 3.4961267569800305,
      "tokens_seen": 1600662528
    },
    {
      "epoch": 5.04,
      "objective/train/docs_used": 1805307,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9074008464813232,
      "objective/train/theoretical_loss": 3.4961174091312692,
      "objective/train/tokens_used": 1621171680,
      "theoretical_loss": 3.4961174091312692,
      "tokens_seen": 1600711680
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002601404212637914,
      "loss": 2.8461,
      "theoretical_loss": 3.4961142932633287,
      "tokens_seen": 1600728064
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00026013039117352057,
      "loss": 2.701,
      "theoretical_loss": 3.4961018301997697,
      "tokens_seen": 1600793600
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00026012036108324975,
      "loss": 2.7185,
      "theoretical_loss": 3.4960893677892937,
      "tokens_seen": 1600859136
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00026011033099297893,
      "loss": 2.4665,
      "theoretical_loss": 3.496076906031839,
      "tokens_seen": 1600924672
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00026010030090270817,
      "loss": 2.6753,
      "theoretical_loss": 3.4960644449273444,
      "tokens_seen": 1600990208
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002600902708124373,
      "loss": 2.8114,
      "theoretical_loss": 3.49605198447575,
      "tokens_seen": 1601055744
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00026008024072216653,
      "loss": 2.7058,
      "theoretical_loss": 3.496039524676994,
      "tokens_seen": 1601121280
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00026007021063189566,
      "loss": 2.7671,
      "theoretical_loss": 3.496027065531015,
      "tokens_seen": 1601186816
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002600601805416249,
      "loss": 2.688,
      "theoretical_loss": 3.496014607037754,
      "tokens_seen": 1601252352
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00026005015045135407,
      "loss": 2.7581,
      "theoretical_loss": 3.496002149197148,
      "tokens_seen": 1601317888
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00026004012036108325,
      "loss": 2.8447,
      "theoretical_loss": 3.4959896920091373,
      "tokens_seen": 1601383424
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00026003009027081243,
      "loss": 2.6624,
      "theoretical_loss": 3.495977235473661,
      "tokens_seen": 1601448960
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002600200601805416,
      "loss": 2.6325,
      "theoretical_loss": 3.4959647795906577,
      "tokens_seen": 1601514496
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002600100300902708,
      "loss": 2.5836,
      "theoretical_loss": 3.4959523243600668,
      "tokens_seen": 1601580032
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00026000000000000003,
      "loss": 2.7603,
      "theoretical_loss": 3.495939869781827,
      "tokens_seen": 1601645568
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025998996990972916,
      "loss": 2.6747,
      "theoretical_loss": 3.495927415855878,
      "tokens_seen": 1601711104
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002599799398194584,
      "loss": 2.5385,
      "theoretical_loss": 3.495914962582159,
      "tokens_seen": 1601776640
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002599699097291876,
      "loss": 2.6931,
      "theoretical_loss": 3.495902509960609,
      "tokens_seen": 1601842176
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025995987963891676,
      "loss": 2.5442,
      "theoretical_loss": 3.4958900579911667,
      "tokens_seen": 1601907712
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025994984954864594,
      "loss": 2.7225,
      "theoretical_loss": 3.495877606673772,
      "tokens_seen": 1601973248
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002599398194583751,
      "loss": 2.8537,
      "theoretical_loss": 3.495865156008364,
      "tokens_seen": 1602038784
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002599297893681043,
      "loss": 2.4171,
      "theoretical_loss": 3.495852705994881,
      "tokens_seen": 1602104320
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025991975927783353,
      "loss": 2.6795,
      "theoretical_loss": 3.4958402566332634,
      "tokens_seen": 1602169856
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025990972918756266,
      "loss": 2.7819,
      "theoretical_loss": 3.4958278079234493,
      "tokens_seen": 1602235392
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002598996990972919,
      "loss": 2.6898,
      "theoretical_loss": 3.4958153598653787,
      "tokens_seen": 1602300928
    },
    {
      "epoch": 5.04,
      "objective/train/docs_used": 1806486,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.2253284454345703,
      "objective/train/theoretical_loss": 3.4958060242494957,
      "objective/train/tokens_used": 1622810080,
      "theoretical_loss": 3.4958060242494957,
      "tokens_seen": 1602350080
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.000259889669007021,
      "loss": 2.7106,
      "theoretical_loss": 3.495802912458991,
      "tokens_seen": 1602366464
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025987963891675026,
      "loss": 2.9012,
      "theoretical_loss": 3.4957904657042245,
      "tokens_seen": 1602432000
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025986960882647944,
      "loss": 2.6426,
      "theoretical_loss": 3.495778019601019,
      "tokens_seen": 1602497536
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002598595787362086,
      "loss": 2.8334,
      "theoretical_loss": 3.495765574149314,
      "tokens_seen": 1602563072
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002598495486459378,
      "loss": 2.8165,
      "theoretical_loss": 3.4957531293490485,
      "tokens_seen": 1602628608
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.000259839518555667,
      "loss": 2.6539,
      "theoretical_loss": 3.495740685200161,
      "tokens_seen": 1602694144
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025982948846539616,
      "loss": 2.5846,
      "theoretical_loss": 3.4957282417025928,
      "tokens_seen": 1602759680
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002598194583751254,
      "loss": 2.8119,
      "theoretical_loss": 3.495715798856281,
      "tokens_seen": 1602825216
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002598094282848545,
      "loss": 2.7783,
      "theoretical_loss": 3.4957033566611657,
      "tokens_seen": 1602890752
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025979939819458376,
      "loss": 2.4276,
      "theoretical_loss": 3.4956909151171867,
      "tokens_seen": 1602956288
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025978936810431294,
      "loss": 2.6884,
      "theoretical_loss": 3.495678474224283,
      "tokens_seen": 1603021824
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002597793380140421,
      "loss": 2.5727,
      "theoretical_loss": 3.495666033982393,
      "tokens_seen": 1603087360
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002597693079237713,
      "loss": 2.6541,
      "theoretical_loss": 3.4956535943914573,
      "tokens_seen": 1603152896
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002597592778335005,
      "loss": 2.813,
      "theoretical_loss": 3.4956411554514144,
      "tokens_seen": 1603218432
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002597492477432297,
      "loss": 2.6882,
      "theoretical_loss": 3.4956287171622047,
      "tokens_seen": 1603283968
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002597392176529589,
      "loss": 2.6274,
      "theoretical_loss": 3.495616279523766,
      "tokens_seen": 1603349504
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002597291875626881,
      "loss": 2.5218,
      "theoretical_loss": 3.495603842536039,
      "tokens_seen": 1603415040
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025971915747241727,
      "loss": 2.8809,
      "theoretical_loss": 3.4955914061989626,
      "tokens_seen": 1603480576
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025970912738214645,
      "loss": 2.6464,
      "theoretical_loss": 3.495578970512476,
      "tokens_seen": 1603546112
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025969909729187563,
      "loss": 2.7924,
      "theoretical_loss": 3.4955665354765184,
      "tokens_seen": 1603611648
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025968906720160486,
      "loss": 2.6967,
      "theoretical_loss": 3.49555410109103,
      "tokens_seen": 1603677184
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.000259679037111334,
      "loss": 2.6436,
      "theoretical_loss": 3.4955416673559494,
      "tokens_seen": 1603742720
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002596690070210632,
      "loss": 2.6897,
      "theoretical_loss": 3.495529234271216,
      "tokens_seen": 1603808256
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025965897693079235,
      "loss": 2.7027,
      "theoretical_loss": 3.4955168018367697,
      "tokens_seen": 1603873792
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002596489468405216,
      "loss": 2.8437,
      "theoretical_loss": 3.49550437005255,
      "tokens_seen": 1603939328
    },
    {
      "epoch": 5.04,
      "objective/train/docs_used": 1807150,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.3969147205352783,
      "objective/train/theoretical_loss": 3.495495046641059,
      "objective/train/tokens_used": 1624448480,
      "theoretical_loss": 3.495495046641059,
      "tokens_seen": 1603988480
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025963891675025077,
      "loss": 2.6088,
      "theoretical_loss": 3.4954919389184953,
      "tokens_seen": 1604004864
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025962888665997995,
      "loss": 2.5807,
      "theoretical_loss": 3.4954795084345465,
      "tokens_seen": 1604070400
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025961885656970913,
      "loss": 2.5547,
      "theoretical_loss": 3.4954670786006417,
      "tokens_seen": 1604135936
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025960882647943837,
      "loss": 2.7026,
      "theoretical_loss": 3.495454649416721,
      "tokens_seen": 1604201472
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002595987963891675,
      "loss": 2.704,
      "theoretical_loss": 3.4954422208827243,
      "tokens_seen": 1604267008
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025958876629889673,
      "loss": 2.7917,
      "theoretical_loss": 3.4954297929985905,
      "tokens_seen": 1604332544
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025957873620862586,
      "loss": 2.8454,
      "theoretical_loss": 3.4954173657642587,
      "tokens_seen": 1604398080
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002595687061183551,
      "loss": 2.7291,
      "theoretical_loss": 3.4954049391796693,
      "tokens_seen": 1604463616
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025955867602808427,
      "loss": 2.582,
      "theoretical_loss": 3.495392513244761,
      "tokens_seen": 1604529152
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025954864593781345,
      "loss": 2.6793,
      "theoretical_loss": 3.4953800879594743,
      "tokens_seen": 1604594688
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025953861584754263,
      "loss": 2.5043,
      "theoretical_loss": 3.495367663323747,
      "tokens_seen": 1604660224
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002595285857572718,
      "loss": 2.5954,
      "theoretical_loss": 3.4953552393375205,
      "tokens_seen": 1604725760
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.000259518555667001,
      "loss": 2.6704,
      "theoretical_loss": 3.495342816000733,
      "tokens_seen": 1604791296
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025950852557673023,
      "loss": 2.5959,
      "theoretical_loss": 3.4953303933133246,
      "tokens_seen": 1604856832
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025949849548645936,
      "loss": 2.6485,
      "theoretical_loss": 3.495317971275235,
      "tokens_seen": 1604922368
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002594884653961886,
      "loss": 2.6488,
      "theoretical_loss": 3.495305549886403,
      "tokens_seen": 1604987904
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002594784353059178,
      "loss": 2.664,
      "theoretical_loss": 3.495293129146769,
      "tokens_seen": 1605053440
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025946840521564696,
      "loss": 2.5702,
      "theoretical_loss": 3.4952807090562725,
      "tokens_seen": 1605118976
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025945837512537614,
      "loss": 2.8248,
      "theoretical_loss": 3.4952682896148524,
      "tokens_seen": 1605184512
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002594483450351053,
      "loss": 2.5666,
      "theoretical_loss": 3.495255870822449,
      "tokens_seen": 1605250048
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002594383149448345,
      "loss": 2.5176,
      "theoretical_loss": 3.4952434526790013,
      "tokens_seen": 1605315584
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025942828485456373,
      "loss": 2.7384,
      "theoretical_loss": 3.495231035184449,
      "tokens_seen": 1605381120
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025941825476429286,
      "loss": 2.4549,
      "theoretical_loss": 3.4952186183387326,
      "tokens_seen": 1605446656
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002594082246740221,
      "loss": 2.6885,
      "theoretical_loss": 3.49520620214179,
      "tokens_seen": 1605512192
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002593981945837512,
      "loss": 2.7188,
      "theoretical_loss": 3.4951937865935623,
      "tokens_seen": 1605577728
    },
    {
      "epoch": 5.04,
      "objective/train/docs_used": 1808698,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.928004503250122,
      "objective/train/theoretical_loss": 3.495184475358074,
      "objective/train/tokens_used": 1626086880,
      "theoretical_loss": 3.495184475358074,
      "tokens_seen": 1605626880
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025938816449348046,
      "loss": 2.8332,
      "theoretical_loss": 3.495181371693989,
      "tokens_seen": 1605643264
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025937813440320964,
      "loss": 2.6397,
      "theoretical_loss": 3.4951689574430085,
      "tokens_seen": 1605708800
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002593681043129388,
      "loss": 2.628,
      "theoretical_loss": 3.4951565438405616,
      "tokens_seen": 1605774336
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.000259358074222668,
      "loss": 2.8962,
      "theoretical_loss": 3.4951441308865885,
      "tokens_seen": 1605839872
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002593480441323972,
      "loss": 2.7003,
      "theoretical_loss": 3.495131718581027,
      "tokens_seen": 1605905408
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025933801404212636,
      "loss": 2.779,
      "theoretical_loss": 3.4951193069238182,
      "tokens_seen": 1605970944
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002593279839518556,
      "loss": 2.77,
      "theoretical_loss": 3.495106895914901,
      "tokens_seen": 1606036480
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002593179538615847,
      "loss": 2.6084,
      "theoretical_loss": 3.4950944855542163,
      "tokens_seen": 1606102016
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025930792377131396,
      "loss": 2.7552,
      "theoretical_loss": 3.4950820758417023,
      "tokens_seen": 1606167552
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025929789368104314,
      "loss": 2.5836,
      "theoretical_loss": 3.4950696667772996,
      "tokens_seen": 1606233088
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002592878635907723,
      "loss": 2.63,
      "theoretical_loss": 3.495057258360948,
      "tokens_seen": 1606298624
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002592778335005015,
      "loss": 2.6871,
      "theoretical_loss": 3.495044850592586,
      "tokens_seen": 1606364160
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002592678034102307,
      "loss": 2.5204,
      "theoretical_loss": 3.4950324434721547,
      "tokens_seen": 1606429696
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025925777331995987,
      "loss": 2.8965,
      "theoretical_loss": 3.4950200369995934,
      "tokens_seen": 1606495232
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002592477432296891,
      "loss": 2.6848,
      "theoretical_loss": 3.4950076311748415,
      "tokens_seen": 1606560768
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025923771313941823,
      "loss": 2.5228,
      "theoretical_loss": 3.4949952259978394,
      "tokens_seen": 1606626304
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025922768304914747,
      "loss": 2.6486,
      "theoretical_loss": 3.4949828214685263,
      "tokens_seen": 1606691840
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002592176529588766,
      "loss": 2.7272,
      "theoretical_loss": 3.4949704175868423,
      "tokens_seen": 1606757376
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025920762286860583,
      "loss": 2.7516,
      "theoretical_loss": 3.494958014352727,
      "tokens_seen": 1606822912
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.000259197592778335,
      "loss": 2.7038,
      "theoretical_loss": 3.4949456117661204,
      "tokens_seen": 1606888448
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002591875626880642,
      "loss": 2.6016,
      "theoretical_loss": 3.4949332098269617,
      "tokens_seen": 1606953984
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025917753259779337,
      "loss": 2.7287,
      "theoretical_loss": 3.4949208085351913,
      "tokens_seen": 1607019520
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025916750250752255,
      "loss": 2.6791,
      "theoretical_loss": 3.4949084078907493,
      "tokens_seen": 1607085056
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025915747241725173,
      "loss": 2.9492,
      "theoretical_loss": 3.4948960078935745,
      "tokens_seen": 1607150592
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025914744232698097,
      "loss": 2.7316,
      "theoretical_loss": 3.4948836085436072,
      "tokens_seen": 1607216128
    },
    {
      "epoch": 5.04,
      "objective/train/docs_used": 1809362,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.568077564239502,
      "objective/train/theoretical_loss": 3.4948743094558257,
      "objective/train/tokens_used": 1627725280,
      "theoretical_loss": 3.4948743094558257,
      "tokens_seen": 1607265280
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002591374122367101,
      "loss": 2.7754,
      "theoretical_loss": 3.4948712098407873,
      "tokens_seen": 1607281664
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025912738214643933,
      "loss": 2.8811,
      "theoretical_loss": 3.494858811785055,
      "tokens_seen": 1607347200
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002591173520561685,
      "loss": 2.8683,
      "theoretical_loss": 3.494846414376349,
      "tokens_seen": 1607412736
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002591073219658977,
      "loss": 2.8006,
      "theoretical_loss": 3.494834017614611,
      "tokens_seen": 1607478272
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002590972918756269,
      "loss": 2.6925,
      "theoretical_loss": 3.494821621499779,
      "tokens_seen": 1607543808
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025908726178535606,
      "loss": 2.5741,
      "theoretical_loss": 3.4948092260317942,
      "tokens_seen": 1607609344
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025907723169508524,
      "loss": 2.6539,
      "theoretical_loss": 3.4947968312105955,
      "tokens_seen": 1607674880
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025906720160481447,
      "loss": 2.5719,
      "theoretical_loss": 3.4947844370361234,
      "tokens_seen": 1607740416
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002590571715145436,
      "loss": 2.818,
      "theoretical_loss": 3.4947720435083176,
      "tokens_seen": 1607805952
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025904714142427283,
      "loss": 2.7345,
      "theoretical_loss": 3.494759650627118,
      "tokens_seen": 1607871488
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025903711133400196,
      "loss": 2.6123,
      "theoretical_loss": 3.494747258392465,
      "tokens_seen": 1607937024
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002590270812437312,
      "loss": 2.5474,
      "theoretical_loss": 3.4947348668042975,
      "tokens_seen": 1608002560
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002590170511534604,
      "loss": 2.691,
      "theoretical_loss": 3.494722475862556,
      "tokens_seen": 1608068096
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025900702106318956,
      "loss": 2.5017,
      "theoretical_loss": 3.4947100855671804,
      "tokens_seen": 1608133632
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002589969909729188,
      "loss": 2.5006,
      "theoretical_loss": 3.494697695918111,
      "tokens_seen": 1608199168
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.000258986960882648,
      "loss": 2.4911,
      "theoretical_loss": 3.4946853069152874,
      "tokens_seen": 1608264704
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025897693079237716,
      "loss": 2.7563,
      "theoretical_loss": 3.494672918558649,
      "tokens_seen": 1608330240
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025896690070210634,
      "loss": 2.8289,
      "theoretical_loss": 3.494660530848137,
      "tokens_seen": 1608395776
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002589568706118355,
      "loss": 2.7974,
      "theoretical_loss": 3.4946481437836905,
      "tokens_seen": 1608461312
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002589468405215647,
      "loss": 2.6105,
      "theoretical_loss": 3.49463575736525,
      "tokens_seen": 1608526848
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025893681043129393,
      "loss": 2.6397,
      "theoretical_loss": 3.4946233715927546,
      "tokens_seen": 1608592384
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025892678034102306,
      "loss": 2.7687,
      "theoretical_loss": 3.494610986466145,
      "tokens_seen": 1608657920
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002589167502507523,
      "loss": 2.7845,
      "theoretical_loss": 3.494598601985362,
      "tokens_seen": 1608723456
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002589067201604814,
      "loss": 2.6956,
      "theoretical_loss": 3.4945862181503435,
      "tokens_seen": 1608788992
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025889669007021066,
      "loss": 2.8939,
      "theoretical_loss": 3.4945738349610314,
      "tokens_seen": 1608854528
    },
    {
      "epoch": 5.04,
      "objective/train/docs_used": 1810830,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7473294734954834,
      "objective/train/theoretical_loss": 3.4945645479927556,
      "objective/train/tokens_used": 1629363680,
      "theoretical_loss": 3.4945645479927556,
      "tokens_seen": 1608903680
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025888665997993984,
      "loss": 2.6125,
      "theoretical_loss": 3.494561452417365,
      "tokens_seen": 1608920064
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.000258876629889669,
      "loss": 2.8013,
      "theoretical_loss": 3.4945490705192843,
      "tokens_seen": 1608985600
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002588665997993982,
      "loss": 2.5571,
      "theoretical_loss": 3.4945366892667296,
      "tokens_seen": 1609051136
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002588565697091274,
      "loss": 2.6408,
      "theoretical_loss": 3.4945243086596407,
      "tokens_seen": 1609116672
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025884653961885657,
      "loss": 2.5487,
      "theoretical_loss": 3.494511928697958,
      "tokens_seen": 1609182208
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002588365095285858,
      "loss": 2.8844,
      "theoretical_loss": 3.4944995493816213,
      "tokens_seen": 1609247744
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025882647943831493,
      "loss": 2.5929,
      "theoretical_loss": 3.4944871707105705,
      "tokens_seen": 1609313280
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025881644934804416,
      "loss": 2.7201,
      "theoretical_loss": 3.4944747926847457,
      "tokens_seen": 1609378816
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025880641925777334,
      "loss": 2.591,
      "theoretical_loss": 3.4944624153040875,
      "tokens_seen": 1609444352
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002587963891675025,
      "loss": 2.6937,
      "theoretical_loss": 3.494450038568536,
      "tokens_seen": 1609509888
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002587863590772317,
      "loss": 2.7347,
      "theoretical_loss": 3.4944376624780307,
      "tokens_seen": 1609575424
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002587763289869609,
      "loss": 2.5359,
      "theoretical_loss": 3.494425287032512,
      "tokens_seen": 1609640960
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025876629889669007,
      "loss": 2.771,
      "theoretical_loss": 3.4944129122319207,
      "tokens_seen": 1609706496
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002587562688064193,
      "loss": 2.6761,
      "theoretical_loss": 3.4944005380761958,
      "tokens_seen": 1609772032
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025874623871614843,
      "loss": 2.6478,
      "theoretical_loss": 3.4943881645652777,
      "tokens_seen": 1609837568
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025873620862587767,
      "loss": 2.5769,
      "theoretical_loss": 3.494375791699107,
      "tokens_seen": 1609903104
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002587261785356068,
      "loss": 2.5977,
      "theoretical_loss": 3.494363419477624,
      "tokens_seen": 1609968640
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025871614844533603,
      "loss": 2.528,
      "theoretical_loss": 3.4943510479007687,
      "tokens_seen": 1610034176
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002587061183550652,
      "loss": 2.6623,
      "theoretical_loss": 3.4943386769684803,
      "tokens_seen": 1610099712
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002586960882647944,
      "loss": 2.8623,
      "theoretical_loss": 3.4943263066807004,
      "tokens_seen": 1610165248
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025868605817452357,
      "loss": 2.5631,
      "theoretical_loss": 3.4943139370373686,
      "tokens_seen": 1610230784
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025867602808425275,
      "loss": 2.7814,
      "theoretical_loss": 3.494301568038425,
      "tokens_seen": 1610296320
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025866599799398193,
      "loss": 2.2406,
      "theoretical_loss": 3.4942891996838097,
      "tokens_seen": 1610361856
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025865596790371117,
      "loss": 2.6263,
      "theoretical_loss": 3.494276831973463,
      "tokens_seen": 1610427392
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002586459378134403,
      "loss": 2.7764,
      "theoretical_loss": 3.494264464907326,
      "tokens_seen": 1610492928
    },
    {
      "epoch": 5.04,
      "objective/train/docs_used": 1811590,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.983952760696411,
      "objective/train/theoretical_loss": 3.494255190030449,
      "objective/train/tokens_used": 1631002080,
      "theoretical_loss": 3.494255190030449,
      "tokens_seen": 1610542080
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025863590772316953,
      "loss": 2.7173,
      "theoretical_loss": 3.4942520984853376,
      "tokens_seen": 1610558464
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002586258776328987,
      "loss": 2.7323,
      "theoretical_loss": 3.494239732707438,
      "tokens_seen": 1610624000
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002586158475426279,
      "loss": 2.8016,
      "theoretical_loss": 3.4942273675735693,
      "tokens_seen": 1610689536
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002586058174523571,
      "loss": 2.7218,
      "theoretical_loss": 3.49421500308367,
      "tokens_seen": 1610755072
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025859578736208626,
      "loss": 2.6106,
      "theoretical_loss": 3.4942026392376806,
      "tokens_seen": 1610820608
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025858575727181544,
      "loss": 2.6007,
      "theoretical_loss": 3.4941902760355417,
      "tokens_seen": 1610886144
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025857572718154467,
      "loss": 2.7377,
      "theoretical_loss": 3.494177913477194,
      "tokens_seen": 1610951680
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002585656970912738,
      "loss": 2.5021,
      "theoretical_loss": 3.4941655515625767,
      "tokens_seen": 1611017216
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025855566700100303,
      "loss": 2.7546,
      "theoretical_loss": 3.494153190291631,
      "tokens_seen": 1611082752
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025854563691073216,
      "loss": 2.6762,
      "theoretical_loss": 3.494140829664297,
      "tokens_seen": 1611148288
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002585356068204614,
      "loss": 2.6792,
      "theoretical_loss": 3.494128469680515,
      "tokens_seen": 1611213824
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002585255767301906,
      "loss": 2.6197,
      "theoretical_loss": 3.494116110340225,
      "tokens_seen": 1611279360
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025851554663991976,
      "loss": 2.4191,
      "theoretical_loss": 3.4941037516433675,
      "tokens_seen": 1611344896
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025850551654964894,
      "loss": 2.7003,
      "theoretical_loss": 3.4940913935898834,
      "tokens_seen": 1611410432
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002584954864593782,
      "loss": 2.8108,
      "theoretical_loss": 3.494079036179712,
      "tokens_seen": 1611475968
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002584854563691073,
      "loss": 2.676,
      "theoretical_loss": 3.4940666794127946,
      "tokens_seen": 1611541504
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025847542627883654,
      "loss": 2.7139,
      "theoretical_loss": 3.494054323289071,
      "tokens_seen": 1611607040
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025846539618856566,
      "loss": 2.7514,
      "theoretical_loss": 3.4940419678084815,
      "tokens_seen": 1611672576
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002584553660982949,
      "loss": 2.6642,
      "theoretical_loss": 3.4940296129709667,
      "tokens_seen": 1611738112
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002584453360080241,
      "loss": 2.6856,
      "theoretical_loss": 3.4940172587764673,
      "tokens_seen": 1611803648
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025843530591775326,
      "loss": 2.4908,
      "theoretical_loss": 3.4940049052249234,
      "tokens_seen": 1611869184
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025842527582748244,
      "loss": 2.5908,
      "theoretical_loss": 3.4939925523162754,
      "tokens_seen": 1611934720
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002584152457372116,
      "loss": 2.4254,
      "theoretical_loss": 3.4939802000504634,
      "tokens_seen": 1612000256
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002584052156469408,
      "loss": 2.5605,
      "theoretical_loss": 3.4939678484274284,
      "tokens_seen": 1612065792
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025839518555667004,
      "loss": 2.5772,
      "theoretical_loss": 3.49395549744711,
      "tokens_seen": 1612131328
    },
    {
      "epoch": 5.04,
      "objective/train/docs_used": 1813183,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.669978380203247,
      "objective/train/theoretical_loss": 3.4939462346336185,
      "objective/train/tokens_used": 1632640480,
      "theoretical_loss": 3.4939462346336185,
      "tokens_seen": 1612180480
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025838515546639917,
      "loss": 2.6425,
      "theoretical_loss": 3.4939431471094498,
      "tokens_seen": 1612196864
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002583751253761284,
      "loss": 2.664,
      "theoretical_loss": 3.4939307974143867,
      "tokens_seen": 1612262400
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025836509528585753,
      "loss": 2.7776,
      "theoretical_loss": 3.493918448361863,
      "tokens_seen": 1612327936
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025835506519558677,
      "loss": 2.6187,
      "theoretical_loss": 3.4939060999518174,
      "tokens_seen": 1612393472
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025834503510531595,
      "loss": 2.6924,
      "theoretical_loss": 3.493893752184192,
      "tokens_seen": 1612459008
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025833500501504513,
      "loss": 2.572,
      "theoretical_loss": 3.493881405058926,
      "tokens_seen": 1612524544
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002583249749247743,
      "loss": 2.5499,
      "theoretical_loss": 3.49386905857596,
      "tokens_seen": 1612590080
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025831494483450354,
      "loss": 2.7041,
      "theoretical_loss": 3.4938567127352353,
      "tokens_seen": 1612655616
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025830491474423267,
      "loss": 2.5322,
      "theoretical_loss": 3.4938443675366915,
      "tokens_seen": 1612721152
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002582948846539619,
      "loss": 2.6045,
      "theoretical_loss": 3.4938320229802695,
      "tokens_seen": 1612786688
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025828485456369103,
      "loss": 2.9335,
      "theoretical_loss": 3.49381967906591,
      "tokens_seen": 1612852224
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025827482447342027,
      "loss": 2.556,
      "theoretical_loss": 3.4938073357935533,
      "tokens_seen": 1612917760
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025826479438314945,
      "loss": 2.5735,
      "theoretical_loss": 3.4937949931631396,
      "tokens_seen": 1612983296
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025825476429287863,
      "loss": 2.7174,
      "theoretical_loss": 3.49378265117461,
      "tokens_seen": 1613048832
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025824473420260787,
      "loss": 2.7566,
      "theoretical_loss": 3.493770309827905,
      "tokens_seen": 1613114368
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.000258234704112337,
      "loss": 2.668,
      "theoretical_loss": 3.493757969122965,
      "tokens_seen": 1613179904
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025822467402206623,
      "loss": 2.6568,
      "theoretical_loss": 3.49374562905973,
      "tokens_seen": 1613245440
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002582146439317954,
      "loss": 2.5037,
      "theoretical_loss": 3.4937332896381417,
      "tokens_seen": 1613310976
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002582046138415246,
      "loss": 2.7663,
      "theoretical_loss": 3.4937209508581395,
      "tokens_seen": 1613376512
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025819458375125377,
      "loss": 2.7212,
      "theoretical_loss": 3.493708612719665,
      "tokens_seen": 1613442048
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025818455366098295,
      "loss": 2.6807,
      "theoretical_loss": 3.493696275222658,
      "tokens_seen": 1613507584
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025817452357071213,
      "loss": 2.5853,
      "theoretical_loss": 3.4936839383670595,
      "tokens_seen": 1613573120
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025816449348044137,
      "loss": 2.5835,
      "theoretical_loss": 3.4936716021528103,
      "tokens_seen": 1613638656
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002581544633901705,
      "loss": 2.6914,
      "theoretical_loss": 3.4936592665798507,
      "tokens_seen": 1613704192
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025814443329989973,
      "loss": 2.6472,
      "theoretical_loss": 3.493646931648121,
      "tokens_seen": 1613769728
    },
    {
      "epoch": 5.04,
      "objective/train/docs_used": 1814061,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6768484115600586,
      "objective/train/theoretical_loss": 3.493637680870095,
      "objective/train/tokens_used": 1634278880,
      "theoretical_loss": 3.493637680870095,
      "tokens_seen": 1613818880
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002581344032096289,
      "loss": 2.5848,
      "theoretical_loss": 3.493634597357562,
      "tokens_seen": 1613835264
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002581243731193581,
      "loss": 2.6934,
      "theoretical_loss": 3.493622263708115,
      "tokens_seen": 1613900800
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002581143430290873,
      "loss": 2.5939,
      "theoretical_loss": 3.49360993069972,
      "tokens_seen": 1613966336
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025810431293881646,
      "loss": 2.3703,
      "theoretical_loss": 3.493597598332318,
      "tokens_seen": 1614031872
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025809428284854564,
      "loss": 2.7793,
      "theoretical_loss": 3.49358526660585,
      "tokens_seen": 1614097408
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025808425275827487,
      "loss": 2.7513,
      "theoretical_loss": 3.4935729355202554,
      "tokens_seen": 1614162944
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.000258074222668004,
      "loss": 2.7046,
      "theoretical_loss": 3.4935606050754755,
      "tokens_seen": 1614228480
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025806419257773323,
      "loss": 2.6072,
      "theoretical_loss": 3.4935482752714515,
      "tokens_seen": 1614294016
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025805416248746236,
      "loss": 2.6282,
      "theoretical_loss": 3.4935359461081235,
      "tokens_seen": 1614359552
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002580441323971916,
      "loss": 2.5873,
      "theoretical_loss": 3.4935236175854323,
      "tokens_seen": 1614425088
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002580341023069208,
      "loss": 2.7102,
      "theoretical_loss": 3.493511289703319,
      "tokens_seen": 1614490624
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025802407221664996,
      "loss": 2.6728,
      "theoretical_loss": 3.4934989624617243,
      "tokens_seen": 1614556160
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025801404212637914,
      "loss": 2.7576,
      "theoretical_loss": 3.493486635860588,
      "tokens_seen": 1614621696
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002580040120361084,
      "loss": 2.7063,
      "theoretical_loss": 3.493474309899852,
      "tokens_seen": 1614687232
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002579939819458375,
      "loss": 2.6767,
      "theoretical_loss": 3.493461984579456,
      "tokens_seen": 1614752768
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025798395185556674,
      "loss": 2.5628,
      "theoretical_loss": 3.493449659899342,
      "tokens_seen": 1614818304
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025797392176529586,
      "loss": 2.701,
      "theoretical_loss": 3.4934373358594497,
      "tokens_seen": 1614883840
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002579638916750251,
      "loss": 2.6278,
      "theoretical_loss": 3.4934250124597197,
      "tokens_seen": 1614949376
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002579538615847543,
      "loss": 2.565,
      "theoretical_loss": 3.4934126897000937,
      "tokens_seen": 1615014912
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025794383149448346,
      "loss": 2.6736,
      "theoretical_loss": 3.493400367580512,
      "tokens_seen": 1615080448
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025793380140421264,
      "loss": 2.7948,
      "theoretical_loss": 3.4933880461009155,
      "tokens_seen": 1615145984
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002579237713139418,
      "loss": 2.6424,
      "theoretical_loss": 3.493375725261245,
      "tokens_seen": 1615211520
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.000257913741223671,
      "loss": 2.597,
      "theoretical_loss": 3.4933634050614413,
      "tokens_seen": 1615277056
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025790371113340024,
      "loss": 2.5277,
      "theoretical_loss": 3.4933510855014447,
      "tokens_seen": 1615342592
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025789368104312937,
      "loss": 2.7266,
      "theoretical_loss": 3.493338766581197,
      "tokens_seen": 1615408128
    },
    {
      "epoch": 5.04,
      "objective/train/docs_used": 1814677,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.616262197494507,
      "objective/train/theoretical_loss": 3.4933295278108103,
      "objective/train/tokens_used": 1635917280,
      "theoretical_loss": 3.4933295278108103,
      "tokens_seen": 1615457280
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002578836509528586,
      "loss": 2.6068,
      "theoretical_loss": 3.493326448300638,
      "tokens_seen": 1615473664
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025787362086258773,
      "loss": 2.5708,
      "theoretical_loss": 3.493314130659709,
      "tokens_seen": 1615539200
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025786359077231697,
      "loss": 2.6691,
      "theoretical_loss": 3.493301813658351,
      "tokens_seen": 1615604736
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025785356068204615,
      "loss": 2.6395,
      "theoretical_loss": 3.493289497296505,
      "tokens_seen": 1615670272
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025784353059177533,
      "loss": 2.6006,
      "theoretical_loss": 3.4932771815741113,
      "tokens_seen": 1615735808
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002578335005015045,
      "loss": 2.5032,
      "theoretical_loss": 3.493264866491111,
      "tokens_seen": 1615801344
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025782347041123374,
      "loss": 2.67,
      "theoretical_loss": 3.493252552047445,
      "tokens_seen": 1615866880
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025781344032096287,
      "loss": 2.7613,
      "theoretical_loss": 3.493240238243054,
      "tokens_seen": 1615932416
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002578034102306921,
      "loss": 2.6414,
      "theoretical_loss": 3.4932279250778793,
      "tokens_seen": 1615997952
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025779338014042123,
      "loss": 2.7801,
      "theoretical_loss": 3.493215612551862,
      "tokens_seen": 1616063488
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025778335005015047,
      "loss": 2.6191,
      "theoretical_loss": 3.4932033006649417,
      "tokens_seen": 1616129024
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025777331995987965,
      "loss": 2.4602,
      "theoretical_loss": 3.4931909894170605,
      "tokens_seen": 1616194560
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025776328986960883,
      "loss": 2.7383,
      "theoretical_loss": 3.4931786788081594,
      "tokens_seen": 1616260096
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.000257753259779338,
      "loss": 2.7904,
      "theoretical_loss": 3.493166368838178,
      "tokens_seen": 1616325632
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002577432296890672,
      "loss": 2.7796,
      "theoretical_loss": 3.4931540595070594,
      "tokens_seen": 1616391168
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002577331995987964,
      "loss": 2.615,
      "theoretical_loss": 3.4931417508147424,
      "tokens_seen": 1616456704
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002577231695085256,
      "loss": 2.3553,
      "theoretical_loss": 3.4931294427611697,
      "tokens_seen": 1616522240
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025771313941825474,
      "loss": 2.7692,
      "theoretical_loss": 3.4931171353462807,
      "tokens_seen": 1616587776
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025770310932798397,
      "loss": 2.5657,
      "theoretical_loss": 3.4931048285700177,
      "tokens_seen": 1616653312
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002576930792377131,
      "loss": 2.8594,
      "theoretical_loss": 3.4930925224323204,
      "tokens_seen": 1616718848
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025768304914744233,
      "loss": 2.7142,
      "theoretical_loss": 3.493080216933131,
      "tokens_seen": 1616784384
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002576730190571715,
      "loss": 2.5003,
      "theoretical_loss": 3.493067912072389,
      "tokens_seen": 1616849920
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002576629889669007,
      "loss": 2.7273,
      "theoretical_loss": 3.493055607850037,
      "tokens_seen": 1616915456
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002576529588766299,
      "loss": 2.7189,
      "theoretical_loss": 3.4930433042660156,
      "tokens_seen": 1616980992
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002576429287863591,
      "loss": 2.5459,
      "theoretical_loss": 3.4930310013202654,
      "tokens_seen": 1617046528
    },
    {
      "epoch": 5.04,
      "objective/train/docs_used": 1816051,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6610658168792725,
      "objective/train/theoretical_loss": 3.493021774529783,
      "objective/train/tokens_used": 1637555680,
      "theoretical_loss": 3.493021774529783,
      "tokens_seen": 1617095680
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025763289869608824,
      "loss": 2.6226,
      "theoretical_loss": 3.493018699012728,
      "tokens_seen": 1617112064
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002576228686058175,
      "loss": 2.4937,
      "theoretical_loss": 3.493006397343343,
      "tokens_seen": 1617177600
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002576128385155466,
      "loss": 2.5907,
      "theoretical_loss": 3.492994096312053,
      "tokens_seen": 1617243136
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025760280842527584,
      "loss": 2.4163,
      "theoretical_loss": 3.4929817959187988,
      "tokens_seen": 1617308672
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.000257592778335005,
      "loss": 2.4671,
      "theoretical_loss": 3.4929694961635205,
      "tokens_seen": 1617374208
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002575827482447342,
      "loss": 2.5927,
      "theoretical_loss": 3.4929571970461604,
      "tokens_seen": 1617439744
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002575727181544634,
      "loss": 2.5462,
      "theoretical_loss": 3.492944898566659,
      "tokens_seen": 1617505280
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025756268806419256,
      "loss": 2.4217,
      "theoretical_loss": 3.4929326007249566,
      "tokens_seen": 1617570816
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025755265797392174,
      "loss": 2.6032,
      "theoretical_loss": 3.4929203035209957,
      "tokens_seen": 1617636352
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.000257542627883651,
      "loss": 2.7347,
      "theoretical_loss": 3.4929080069547167,
      "tokens_seen": 1617701888
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002575325977933801,
      "loss": 2.5122,
      "theoretical_loss": 3.4928957110260606,
      "tokens_seen": 1617767424
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025752256770310934,
      "loss": 2.7135,
      "theoretical_loss": 3.4928834157349686,
      "tokens_seen": 1617832960
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025751253761283847,
      "loss": 2.6551,
      "theoretical_loss": 3.4928711210813823,
      "tokens_seen": 1617898496
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002575025075225677,
      "loss": 2.9374,
      "theoretical_loss": 3.4928588270652416,
      "tokens_seen": 1617964032
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025749247743229694,
      "loss": 2.5136,
      "theoretical_loss": 3.492846533686489,
      "tokens_seen": 1618029568
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025748244734202606,
      "loss": 2.5835,
      "theoretical_loss": 3.492834240945065,
      "tokens_seen": 1618095104
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002574724172517553,
      "loss": 2.5203,
      "theoretical_loss": 3.492821948840911,
      "tokens_seen": 1618160640
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002574623871614845,
      "loss": 2.6176,
      "theoretical_loss": 3.4928096573739675,
      "tokens_seen": 1618226176
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025745235707121366,
      "loss": 2.6967,
      "theoretical_loss": 3.4927973665441763,
      "tokens_seen": 1618291712
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025744232698094284,
      "loss": 2.3947,
      "theoretical_loss": 3.4927850763514785,
      "tokens_seen": 1618357248
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.000257432296890672,
      "loss": 2.8494,
      "theoretical_loss": 3.4927727867958147,
      "tokens_seen": 1618422784
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002574222668004012,
      "loss": 2.5314,
      "theoretical_loss": 3.4927604978771276,
      "tokens_seen": 1618488320
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025741223671013044,
      "loss": 2.4884,
      "theoretical_loss": 3.492748209595357,
      "tokens_seen": 1618553856
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025740220661985957,
      "loss": 2.6668,
      "theoretical_loss": 3.4927359219504437,
      "tokens_seen": 1618619392
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002573921765295888,
      "loss": 2.5161,
      "theoretical_loss": 3.4927236349423305,
      "tokens_seen": 1618684928
    },
    {
      "epoch": 5.04,
      "objective/train/docs_used": 1816712,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6617588996887207,
      "objective/train/theoretical_loss": 3.492714420104109,
      "objective/train/tokens_used": 1639194080,
      "theoretical_loss": 3.492714420104109,
      "tokens_seen": 1618734080
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025738214643931793,
      "loss": 2.4688,
      "theoretical_loss": 3.492711348570957,
      "tokens_seen": 1618750464
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025737211634904717,
      "loss": 2.6443,
      "theoretical_loss": 3.492699062836266,
      "tokens_seen": 1618816000
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025736208625877635,
      "loss": 2.8617,
      "theoretical_loss": 3.4926867777381974,
      "tokens_seen": 1618881536
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025735205616850553,
      "loss": 2.5455,
      "theoretical_loss": 3.4926744932766933,
      "tokens_seen": 1618947072
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002573420260782347,
      "loss": 2.6395,
      "theoretical_loss": 3.4926622094516944,
      "tokens_seen": 1619012608
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025733199598796394,
      "loss": 2.5923,
      "theoretical_loss": 3.4926499262631427,
      "tokens_seen": 1619078144
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025732196589769307,
      "loss": 2.6676,
      "theoretical_loss": 3.4926376437109785,
      "tokens_seen": 1619143680
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002573119358074223,
      "loss": 2.7692,
      "theoretical_loss": 3.492625361795144,
      "tokens_seen": 1619209216
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025730190571715143,
      "loss": 2.8743,
      "theoretical_loss": 3.4926130805155795,
      "tokens_seen": 1619274752
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025729187562688067,
      "loss": 2.8537,
      "theoretical_loss": 3.492600799872227,
      "tokens_seen": 1619340288
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025728184553660985,
      "loss": 2.6658,
      "theoretical_loss": 3.492588519865028,
      "tokens_seen": 1619405824
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025727181544633903,
      "loss": 2.6737,
      "theoretical_loss": 3.4925762404939227,
      "tokens_seen": 1619471360
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002572617853560682,
      "loss": 2.8957,
      "theoretical_loss": 3.4925639617588535,
      "tokens_seen": 1619536896
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002572517552657974,
      "loss": 2.6378,
      "theoretical_loss": 3.4925516836597614,
      "tokens_seen": 1619602432
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002572417251755266,
      "loss": 2.6016,
      "theoretical_loss": 3.4925394061965873,
      "tokens_seen": 1619667968
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002572316950852558,
      "loss": 2.5372,
      "theoretical_loss": 3.4925271293692735,
      "tokens_seen": 1619733504
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025722166499498494,
      "loss": 2.428,
      "theoretical_loss": 3.4925148531777603,
      "tokens_seen": 1619799040
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025721163490471417,
      "loss": 2.5885,
      "theoretical_loss": 3.4925025776219893,
      "tokens_seen": 1619864576
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002572016048144433,
      "loss": 2.6786,
      "theoretical_loss": 3.4924903027019027,
      "tokens_seen": 1619930112
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025719157472417253,
      "loss": 2.5506,
      "theoretical_loss": 3.492478028417441,
      "tokens_seen": 1619995648
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002571815446339017,
      "loss": 2.4361,
      "theoretical_loss": 3.4924657547685456,
      "tokens_seen": 1620061184
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002571715145436309,
      "loss": 2.6449,
      "theoretical_loss": 3.492453481755158,
      "tokens_seen": 1620126720
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002571614844533601,
      "loss": 2.548,
      "theoretical_loss": 3.49244120937722,
      "tokens_seen": 1620192256
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002571514543630893,
      "loss": 2.6244,
      "theoretical_loss": 3.4924289376346724,
      "tokens_seen": 1620257792
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025714142427281844,
      "loss": 2.7454,
      "theoretical_loss": 3.4924166665274567,
      "tokens_seen": 1620323328
    },
    {
      "epoch": 5.04,
      "objective/train/docs_used": 1817999,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.959540843963623,
      "objective/train/theoretical_loss": 3.4924074636139464,
      "objective/train/tokens_used": 1640832480,
      "theoretical_loss": 3.4924074636139464,
      "tokens_seen": 1620372480
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002571313941825477,
      "loss": 2.6683,
      "theoretical_loss": 3.4924043960555147,
      "tokens_seen": 1620388864
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002571213640922768,
      "loss": 2.671,
      "theoretical_loss": 3.4923921262187876,
      "tokens_seen": 1620454400
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025711133400200604,
      "loss": 2.8518,
      "theoretical_loss": 3.492379857017217,
      "tokens_seen": 1620519936
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002571013039117352,
      "loss": 2.3591,
      "theoretical_loss": 3.492367588450744,
      "tokens_seen": 1620585472
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002570912738214644,
      "loss": 2.4134,
      "theoretical_loss": 3.49235532051931,
      "tokens_seen": 1620651008
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002570812437311936,
      "loss": 2.6232,
      "theoretical_loss": 3.492343053222857,
      "tokens_seen": 1620716544
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025707121364092276,
      "loss": 2.6938,
      "theoretical_loss": 3.4923307865613262,
      "tokens_seen": 1620782080
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025706118355065194,
      "loss": 2.4861,
      "theoretical_loss": 3.492318520534658,
      "tokens_seen": 1620847616
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002570511534603812,
      "loss": 2.7885,
      "theoretical_loss": 3.492306255142796,
      "tokens_seen": 1620913152
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002570411233701103,
      "loss": 2.6434,
      "theoretical_loss": 3.4922939903856802,
      "tokens_seen": 1620978688
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025703109327983954,
      "loss": 2.5004,
      "theoretical_loss": 3.4922817262632524,
      "tokens_seen": 1621044224
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025702106318956867,
      "loss": 2.7335,
      "theoretical_loss": 3.4922694627754542,
      "tokens_seen": 1621109760
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002570110330992979,
      "loss": 2.734,
      "theoretical_loss": 3.4922571999222267,
      "tokens_seen": 1621175296
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002570010030090271,
      "loss": 2.7062,
      "theoretical_loss": 3.4922449377035125,
      "tokens_seen": 1621240832
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025699097291875626,
      "loss": 2.5759,
      "theoretical_loss": 3.4922326761192517,
      "tokens_seen": 1621306368
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025698094282848545,
      "loss": 2.4875,
      "theoretical_loss": 3.4922204151693865,
      "tokens_seen": 1621371904
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002569709127382147,
      "loss": 2.8393,
      "theoretical_loss": 3.492208154853859,
      "tokens_seen": 1621437440
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002569608826479438,
      "loss": 2.6645,
      "theoretical_loss": 3.4921958951726095,
      "tokens_seen": 1621502976
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025695085255767304,
      "loss": 2.8177,
      "theoretical_loss": 3.4921836361255805,
      "tokens_seen": 1621568512
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025694082246740217,
      "loss": 2.7412,
      "theoretical_loss": 3.4921713777127135,
      "tokens_seen": 1621634048
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002569307923771314,
      "loss": 2.534,
      "theoretical_loss": 3.4921591199339495,
      "tokens_seen": 1621699584
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002569207622868606,
      "loss": 2.5425,
      "theoretical_loss": 3.4921468627892307,
      "tokens_seen": 1621765120
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025691073219658977,
      "loss": 2.6161,
      "theoretical_loss": 3.4921346062784986,
      "tokens_seen": 1621830656
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025690070210631895,
      "loss": 2.3787,
      "theoretical_loss": 3.4921223504016945,
      "tokens_seen": 1621896192
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025689067201604813,
      "loss": 2.6759,
      "theoretical_loss": 3.49211009515876,
      "tokens_seen": 1621961728
    },
    {
      "epoch": 5.04,
      "objective/train/docs_used": 1818775,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.1677393913269043,
      "objective/train/theoretical_loss": 3.492100904142501,
      "objective/train/tokens_used": 1642470880,
      "theoretical_loss": 3.492100904142501,
      "tokens_seen": 1622010880
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002568806419257773,
      "loss": 2.469,
      "theoretical_loss": 3.4920978405496363,
      "tokens_seen": 1622027264
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025687061183550655,
      "loss": 2.6083,
      "theoretical_loss": 3.4920855865742664,
      "tokens_seen": 1622092800
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002568605817452357,
      "loss": 2.5079,
      "theoretical_loss": 3.4920733332325904,
      "tokens_seen": 1622158336
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002568505516549649,
      "loss": 2.6879,
      "theoretical_loss": 3.492061080524551,
      "tokens_seen": 1622223872
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025684052156469404,
      "loss": 2.5899,
      "theoretical_loss": 3.4920488284500895,
      "tokens_seen": 1622289408
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025683049147442327,
      "loss": 2.4714,
      "theoretical_loss": 3.4920365770091473,
      "tokens_seen": 1622354944
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025682046138415245,
      "loss": 2.527,
      "theoretical_loss": 3.4920243262016664,
      "tokens_seen": 1622420480
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025681043129388163,
      "loss": 2.808,
      "theoretical_loss": 3.4920120760275877,
      "tokens_seen": 1622486016
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002568004012036108,
      "loss": 2.5658,
      "theoretical_loss": 3.491999826486854,
      "tokens_seen": 1622551552
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025679037111334005,
      "loss": 2.596,
      "theoretical_loss": 3.4919875775794065,
      "tokens_seen": 1622617088
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002567803410230692,
      "loss": 2.5577,
      "theoretical_loss": 3.4919753293051867,
      "tokens_seen": 1622682624
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002567703109327984,
      "loss": 2.8376,
      "theoretical_loss": 3.491963081664136,
      "tokens_seen": 1622748160
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025676028084252754,
      "loss": 2.6855,
      "theoretical_loss": 3.4919508346561967,
      "tokens_seen": 1622813696
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002567502507522568,
      "loss": 2.5951,
      "theoretical_loss": 3.491938588281311,
      "tokens_seen": 1622879232
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.000256740220661986,
      "loss": 2.5642,
      "theoretical_loss": 3.491926342539419,
      "tokens_seen": 1622944768
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025673019057171514,
      "loss": 2.6147,
      "theoretical_loss": 3.491914097430464,
      "tokens_seen": 1623010304
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025672016048144437,
      "loss": 2.5273,
      "theoretical_loss": 3.4919018529543866,
      "tokens_seen": 1623075840
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002567101303911735,
      "loss": 2.6065,
      "theoretical_loss": 3.491889609111129,
      "tokens_seen": 1623141376
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025670010030090273,
      "loss": 2.5336,
      "theoretical_loss": 3.491877365900633,
      "tokens_seen": 1623206912
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002566900702106319,
      "loss": 2.6442,
      "theoretical_loss": 3.4918651233228406,
      "tokens_seen": 1623272448
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002566800401203611,
      "loss": 2.6089,
      "theoretical_loss": 3.4918528813776932,
      "tokens_seen": 1623337984
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002566700100300903,
      "loss": 2.3904,
      "theoretical_loss": 3.4918406400651327,
      "tokens_seen": 1623403520
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002566599799398195,
      "loss": 2.7396,
      "theoretical_loss": 3.491828399385101,
      "tokens_seen": 1623469056
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025664994984954864,
      "loss": 2.7331,
      "theoretical_loss": 3.491816159337539,
      "tokens_seen": 1623534592
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002566399197592779,
      "loss": 2.762,
      "theoretical_loss": 3.49180391992239,
      "tokens_seen": 1623600128
    },
    {
      "epoch": 5.04,
      "objective/train/docs_used": 1820300,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7725954055786133,
      "objective/train/theoretical_loss": 3.4917947407760135,
      "objective/train/tokens_used": 1644109280,
      "theoretical_loss": 3.4917947407760135,
      "tokens_seen": 1623649280
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.000256629889669007,
      "loss": 2.5771,
      "theoretical_loss": 3.4917916811395946,
      "tokens_seen": 1623665664
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025661985957873624,
      "loss": 2.6068,
      "theoretical_loss": 3.491779442989095,
      "tokens_seen": 1623731200
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002566098294884654,
      "loss": 2.5716,
      "theoretical_loss": 3.491767205470833,
      "tokens_seen": 1623796736
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002565997993981946,
      "loss": 2.5825,
      "theoretical_loss": 3.4917549685847504,
      "tokens_seen": 1623862272
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002565897693079238,
      "loss": 2.7518,
      "theoretical_loss": 3.4917427323307892,
      "tokens_seen": 1623927808
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025657973921765296,
      "loss": 2.507,
      "theoretical_loss": 3.4917304967088914,
      "tokens_seen": 1623993344
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025656970912738214,
      "loss": 2.7368,
      "theoretical_loss": 3.491718261718998,
      "tokens_seen": 1624058880
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002565596790371114,
      "loss": 2.7704,
      "theoretical_loss": 3.491706027361052,
      "tokens_seen": 1624124416
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002565496489468405,
      "loss": 2.5858,
      "theoretical_loss": 3.4916937936349943,
      "tokens_seen": 1624189952
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025653961885656974,
      "loss": 2.8094,
      "theoretical_loss": 3.4916815605407674,
      "tokens_seen": 1624255488
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025652958876629887,
      "loss": 2.7634,
      "theoretical_loss": 3.4916693280783124,
      "tokens_seen": 1624321024
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002565195586760281,
      "loss": 2.5955,
      "theoretical_loss": 3.491657096247572,
      "tokens_seen": 1624386560
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002565095285857573,
      "loss": 2.5747,
      "theoretical_loss": 3.491644865048488,
      "tokens_seen": 1624452096
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025649949849548647,
      "loss": 2.7707,
      "theoretical_loss": 3.491632634481002,
      "tokens_seen": 1624517632
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025648946840521565,
      "loss": 2.6173,
      "theoretical_loss": 3.491620404545056,
      "tokens_seen": 1624583168
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002564794383149449,
      "loss": 2.6951,
      "theoretical_loss": 3.4916081752405916,
      "tokens_seen": 1624648704
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.000256469408224674,
      "loss": 2.7188,
      "theoretical_loss": 3.4915959465675517,
      "tokens_seen": 1624714240
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025645937813440324,
      "loss": 2.4266,
      "theoretical_loss": 3.4915837185258765,
      "tokens_seen": 1624779776
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025644934804413237,
      "loss": 2.6004,
      "theoretical_loss": 3.49157149111551,
      "tokens_seen": 1624845312
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002564393179538616,
      "loss": 2.5292,
      "theoretical_loss": 3.491559264336393,
      "tokens_seen": 1624910848
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002564292878635908,
      "loss": 2.4791,
      "theoretical_loss": 3.4915470381884672,
      "tokens_seen": 1624976384
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025641925777331997,
      "loss": 2.6726,
      "theoretical_loss": 3.4915348126716754,
      "tokens_seen": 1625041920
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025640922768304915,
      "loss": 2.8152,
      "theoretical_loss": 3.4915225877859593,
      "tokens_seen": 1625107456
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025639919759277833,
      "loss": 2.5362,
      "theoretical_loss": 3.49151036353126,
      "tokens_seen": 1625172992
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002563891675025075,
      "loss": 2.5838,
      "theoretical_loss": 3.4914981399075207,
      "tokens_seen": 1625238528
    },
    {
      "epoch": 5.04,
      "objective/train/docs_used": 1820966,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.0666143894195557,
      "objective/train/theoretical_loss": 3.491488972603748,
      "objective/train/tokens_used": 1645747680,
      "theoretical_loss": 3.491488972603748,
      "tokens_seen": 1625287680
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025637913741223675,
      "loss": 2.6085,
      "theoretical_loss": 3.491485916914683,
      "tokens_seen": 1625304064
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002563691073219659,
      "loss": 2.7686,
      "theoretical_loss": 3.4914736945526883,
      "tokens_seen": 1625369600
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002563590772316951,
      "loss": 2.6789,
      "theoretical_loss": 3.491461472821479,
      "tokens_seen": 1625435136
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025634904714142424,
      "loss": 2.4061,
      "theoretical_loss": 3.491449251720998,
      "tokens_seen": 1625500672
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025633901705115347,
      "loss": 2.522,
      "theoretical_loss": 3.491437031251186,
      "tokens_seen": 1625566208
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025632898696088265,
      "loss": 2.5905,
      "theoretical_loss": 3.491424811411986,
      "tokens_seen": 1625631744
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025631895687061183,
      "loss": 2.5403,
      "theoretical_loss": 3.491412592203339,
      "tokens_seen": 1625697280
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.000256308926780341,
      "loss": 2.7574,
      "theoretical_loss": 3.491400373625188,
      "tokens_seen": 1625762816
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025629889669007025,
      "loss": 2.646,
      "theoretical_loss": 3.491388155677475,
      "tokens_seen": 1625828352
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002562888665997994,
      "loss": 2.9001,
      "theoretical_loss": 3.4913759383601413,
      "tokens_seen": 1625893888
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002562788365095286,
      "loss": 2.7254,
      "theoretical_loss": 3.4913637216731295,
      "tokens_seen": 1625959424
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025626880641925774,
      "loss": 2.7043,
      "theoretical_loss": 3.4913515056163815,
      "tokens_seen": 1626024960
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.000256258776328987,
      "loss": 2.6255,
      "theoretical_loss": 3.49133929018984,
      "tokens_seen": 1626090496
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025624874623871616,
      "loss": 2.5757,
      "theoretical_loss": 3.4913270753934462,
      "tokens_seen": 1626156032
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025623871614844534,
      "loss": 2.5782,
      "theoretical_loss": 3.491314861227143,
      "tokens_seen": 1626221568
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002562286860581745,
      "loss": 2.4161,
      "theoretical_loss": 3.4913026476908717,
      "tokens_seen": 1626287104
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002562186559679037,
      "loss": 2.7689,
      "theoretical_loss": 3.491290434784575,
      "tokens_seen": 1626352640
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002562086258776329,
      "loss": 2.7879,
      "theoretical_loss": 3.4912782225081944,
      "tokens_seen": 1626418176
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002561985957873621,
      "loss": 2.5714,
      "theoretical_loss": 3.491266010861673,
      "tokens_seen": 1626483712
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025618856569709124,
      "loss": 2.4168,
      "theoretical_loss": 3.4912537998449515,
      "tokens_seen": 1626549248
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002561785356068205,
      "loss": 2.7801,
      "theoretical_loss": 3.491241589457974,
      "tokens_seen": 1626614784
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025616850551654966,
      "loss": 2.4912,
      "theoretical_loss": 3.4912293797006813,
      "tokens_seen": 1626680320
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025615847542627884,
      "loss": 2.6434,
      "theoretical_loss": 3.491217170573016,
      "tokens_seen": 1626745856
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.000256148445336008,
      "loss": 2.7688,
      "theoretical_loss": 3.49120496207492,
      "tokens_seen": 1626811392
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002561384152457372,
      "loss": 2.777,
      "theoretical_loss": 3.4911927542063355,
      "tokens_seen": 1626876928
    },
    {
      "epoch": 5.04,
      "objective/train/docs_used": 1822305,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.456148147583008,
      "objective/train/theoretical_loss": 3.4911835987179787,
      "objective/train/tokens_used": 1647386080,
      "theoretical_loss": 3.4911835987179787,
      "tokens_seen": 1626926080
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002561283851554664,
      "loss": 2.8612,
      "theoretical_loss": 3.4911805469672044,
      "tokens_seen": 1626942464
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002561183550651956,
      "loss": 2.6333,
      "theoretical_loss": 3.49116834035747,
      "tokens_seen": 1627008000
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025610832497492475,
      "loss": 2.8309,
      "theoretical_loss": 3.491156134377073,
      "tokens_seen": 1627073536
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.000256098294884654,
      "loss": 2.5449,
      "theoretical_loss": 3.4911439290259567,
      "tokens_seen": 1627139072
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002560882647943831,
      "loss": 2.7455,
      "theoretical_loss": 3.491131724304063,
      "tokens_seen": 1627204608
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025607823470411234,
      "loss": 2.7514,
      "theoretical_loss": 3.4911195202113348,
      "tokens_seen": 1627270144
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002560682046138415,
      "loss": 2.5495,
      "theoretical_loss": 3.491107316747713,
      "tokens_seen": 1627335680
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002560581745235707,
      "loss": 2.5578,
      "theoretical_loss": 3.4910951139131408,
      "tokens_seen": 1627401216
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002560481444332999,
      "loss": 2.5765,
      "theoretical_loss": 3.4910829117075597,
      "tokens_seen": 1627466752
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025603811434302907,
      "loss": 2.5857,
      "theoretical_loss": 3.4910707101309124,
      "tokens_seen": 1627532288
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025602808425275825,
      "loss": 2.7842,
      "theoretical_loss": 3.491058509183141,
      "tokens_seen": 1627597824
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002560180541624875,
      "loss": 2.6268,
      "theoretical_loss": 3.4910463088641883,
      "tokens_seen": 1627663360
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002560080240722166,
      "loss": 2.7106,
      "theoretical_loss": 3.491034109173996,
      "tokens_seen": 1627728896
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025599799398194585,
      "loss": 2.7822,
      "theoretical_loss": 3.491021910112507,
      "tokens_seen": 1627794432
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002559879638916751,
      "loss": 2.4951,
      "theoretical_loss": 3.491009711679663,
      "tokens_seen": 1627859968
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002559779338014042,
      "loss": 2.7125,
      "theoretical_loss": 3.490997513875406,
      "tokens_seen": 1627925504
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025596790371113344,
      "loss": 2.4878,
      "theoretical_loss": 3.490985316699679,
      "tokens_seen": 1627991040
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025595787362086257,
      "loss": 2.5812,
      "theoretical_loss": 3.4909731201524243,
      "tokens_seen": 1628056576
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002559478435305918,
      "loss": 2.5068,
      "theoretical_loss": 3.4909609242335833,
      "tokens_seen": 1628122112
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.000255937813440321,
      "loss": 2.596,
      "theoretical_loss": 3.4909487289430996,
      "tokens_seen": 1628187648
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025592778335005017,
      "loss": 2.5647,
      "theoretical_loss": 3.490936534280915,
      "tokens_seen": 1628253184
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025591775325977935,
      "loss": 2.7733,
      "theoretical_loss": 3.4909243402469716,
      "tokens_seen": 1628318720
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025590772316950853,
      "loss": 2.5484,
      "theoretical_loss": 3.490912146841212,
      "tokens_seen": 1628384256
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002558976930792377,
      "loss": 2.493,
      "theoretical_loss": 3.490899954063578,
      "tokens_seen": 1628449792
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025588766298896695,
      "loss": 2.6368,
      "theoretical_loss": 3.4908877619140135,
      "tokens_seen": 1628515328
    },
    {
      "epoch": 5.04,
      "objective/train/docs_used": 1822948,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.1905274391174316,
      "objective/train/theoretical_loss": 3.4908786182139746,
      "objective/train/tokens_used": 1649024480,
      "theoretical_loss": 3.4908786182139746,
      "tokens_seen": 1628564480
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002558776328986961,
      "loss": 2.5279,
      "theoretical_loss": 3.4908755703924594,
      "tokens_seen": 1628580864
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002558676028084253,
      "loss": 2.841,
      "theoretical_loss": 3.490863379498858,
      "tokens_seen": 1628646400
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025585757271815444,
      "loss": 2.432,
      "theoretical_loss": 3.490851189233153,
      "tokens_seen": 1628711936
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025584754262788367,
      "loss": 2.8448,
      "theoretical_loss": 3.4908389995952858,
      "tokens_seen": 1628777472
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025583751253761285,
      "loss": 2.4533,
      "theoretical_loss": 3.4908268105851987,
      "tokens_seen": 1628843008
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025582748244734203,
      "loss": 2.639,
      "theoretical_loss": 3.4908146222028344,
      "tokens_seen": 1628908544
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002558174523570712,
      "loss": 2.7056,
      "theoretical_loss": 3.4908024344481356,
      "tokens_seen": 1628974080
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025580742226680045,
      "loss": 2.5056,
      "theoretical_loss": 3.490790247321045,
      "tokens_seen": 1629039616
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002557973921765296,
      "loss": 2.5828,
      "theoretical_loss": 3.490778060821504,
      "tokens_seen": 1629105152
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002557873620862588,
      "loss": 2.7358,
      "theoretical_loss": 3.4907658749494552,
      "tokens_seen": 1629170688
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025577733199598794,
      "loss": 2.3925,
      "theoretical_loss": 3.490753689704842,
      "tokens_seen": 1629236224
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002557673019057172,
      "loss": 2.5133,
      "theoretical_loss": 3.490741505087606,
      "tokens_seen": 1629301760
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025575727181544636,
      "loss": 2.7957,
      "theoretical_loss": 3.4907293210976897,
      "tokens_seen": 1629367296
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025574724172517554,
      "loss": 2.4953,
      "theoretical_loss": 3.490717137735036,
      "tokens_seen": 1629432832
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002557372116349047,
      "loss": 2.4671,
      "theoretical_loss": 3.4907049549995874,
      "tokens_seen": 1629498368
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002557271815446339,
      "loss": 2.3538,
      "theoretical_loss": 3.490692772891286,
      "tokens_seen": 1629563904
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002557171514543631,
      "loss": 2.63,
      "theoretical_loss": 3.4906805914100745,
      "tokens_seen": 1629629440
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002557071213640923,
      "loss": 2.5732,
      "theoretical_loss": 3.4906684105558954,
      "tokens_seen": 1629694976
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025569709127382144,
      "loss": 2.575,
      "theoretical_loss": 3.4906562303286908,
      "tokens_seen": 1629760512
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002556870611835507,
      "loss": 2.9597,
      "theoretical_loss": 3.4906440507284042,
      "tokens_seen": 1629826048
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025567703109327986,
      "loss": 2.5434,
      "theoretical_loss": 3.490631871754977,
      "tokens_seen": 1629891584
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025566700100300904,
      "loss": 2.3853,
      "theoretical_loss": 3.4906196934083527,
      "tokens_seen": 1629957120
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002556569709127382,
      "loss": 2.6846,
      "theoretical_loss": 3.490607515688473,
      "tokens_seen": 1630022656
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002556469408224674,
      "loss": 2.4938,
      "theoretical_loss": 3.490595338595281,
      "tokens_seen": 1630088192
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002556369107321966,
      "loss": 2.77,
      "theoretical_loss": 3.4905831621287193,
      "tokens_seen": 1630153728
    },
    {
      "epoch": 5.04,
      "objective/train/docs_used": 1824579,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.721491575241089,
      "objective/train/theoretical_loss": 3.4905740301899892,
      "objective/train/tokens_used": 1650662880,
      "theoretical_loss": 3.4905740301899892,
      "tokens_seen": 1630202880
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002556268806419258,
      "loss": 2.7009,
      "theoretical_loss": 3.4905709862887297,
      "tokens_seen": 1630219264
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025561685055165495,
      "loss": 2.9586,
      "theoretical_loss": 3.490558811075256,
      "tokens_seen": 1630284800
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002556068204613842,
      "loss": 2.7171,
      "theoretical_loss": 3.49054663648824,
      "tokens_seen": 1630350336
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002555967903711133,
      "loss": 2.6523,
      "theoretical_loss": 3.4905344625276244,
      "tokens_seen": 1630415872
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025558676028084254,
      "loss": 2.5221,
      "theoretical_loss": 3.4905222891933514,
      "tokens_seen": 1630481408
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002555767301905717,
      "loss": 2.6267,
      "theoretical_loss": 3.4905101164853645,
      "tokens_seen": 1630546944
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002555667001003009,
      "loss": 2.6661,
      "theoretical_loss": 3.4904979444036055,
      "tokens_seen": 1630612480
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002555566700100301,
      "loss": 2.7051,
      "theoretical_loss": 3.490485772948017,
      "tokens_seen": 1630678016
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025554663991975927,
      "loss": 2.7112,
      "theoretical_loss": 3.4904736021185427,
      "tokens_seen": 1630743552
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025553660982948845,
      "loss": 2.7162,
      "theoretical_loss": 3.490461431915124,
      "tokens_seen": 1630809088
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002555265797392177,
      "loss": 2.6602,
      "theoretical_loss": 3.4904492623377044,
      "tokens_seen": 1630874624
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002555165496489468,
      "loss": 2.6757,
      "theoretical_loss": 3.490437093386226,
      "tokens_seen": 1630940160
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025550651955867605,
      "loss": 2.5184,
      "theoretical_loss": 3.4904249250606316,
      "tokens_seen": 1631005696
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025549648946840523,
      "loss": 2.65,
      "theoretical_loss": 3.4904127573608634,
      "tokens_seen": 1631071232
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002554864593781344,
      "loss": 2.6598,
      "theoretical_loss": 3.490400590286865,
      "tokens_seen": 1631136768
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002554764292878636,
      "loss": 2.4051,
      "theoretical_loss": 3.4903884238385787,
      "tokens_seen": 1631202304
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025546639919759277,
      "loss": 2.7635,
      "theoretical_loss": 3.4903762580159468,
      "tokens_seen": 1631267840
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025545636910732195,
      "loss": 2.6505,
      "theoretical_loss": 3.4903640928189126,
      "tokens_seen": 1631333376
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002554463390170512,
      "loss": 2.6143,
      "theoretical_loss": 3.4903519282474185,
      "tokens_seen": 1631398912
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002554363089267803,
      "loss": 2.8442,
      "theoretical_loss": 3.490339764301407,
      "tokens_seen": 1631464448
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025542627883650955,
      "loss": 2.6405,
      "theoretical_loss": 3.4903276009808213,
      "tokens_seen": 1631529984
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002554162487462387,
      "loss": 2.4992,
      "theoretical_loss": 3.4903154382856036,
      "tokens_seen": 1631595520
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002554062186559679,
      "loss": 2.5556,
      "theoretical_loss": 3.4903032762156974,
      "tokens_seen": 1631661056
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002553961885656971,
      "loss": 2.648,
      "theoretical_loss": 3.4902911147710443,
      "tokens_seen": 1631726592
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002553861584754263,
      "loss": 2.6144,
      "theoretical_loss": 3.490278953951588,
      "tokens_seen": 1631792128
    },
    {
      "epoch": 5.04,
      "objective/train/docs_used": 1825363,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6945481300354004,
      "objective/train/theoretical_loss": 3.490269833747246,
      "objective/train/tokens_used": 1652301280,
      "theoretical_loss": 3.490269833747246,
      "tokens_seen": 1631841280
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025537612838515546,
      "loss": 2.5007,
      "theoretical_loss": 3.4902667937572707,
      "tokens_seen": 1631857664
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025536609829488464,
      "loss": 2.9637,
      "theoretical_loss": 3.4902546341880356,
      "tokens_seen": 1631923200
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002553560682046138,
      "loss": 2.6689,
      "theoretical_loss": 3.490242475243825,
      "tokens_seen": 1631988736
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.00025534603811434305,
      "loss": 2.5679,
      "theoretical_loss": 3.490230316924582,
      "tokens_seen": 1632054272
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002553360080240722,
      "loss": 2.5961,
      "theoretical_loss": 3.490218159230249,
      "tokens_seen": 1632119808
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002553259779338014,
      "loss": 2.5681,
      "theoretical_loss": 3.49020600216077,
      "tokens_seen": 1632185344
    },
    {
      "epoch": 5.04,
      "learning_rate": 0.0002553159478435306,
      "loss": 2.5314,
      "theoretical_loss": 3.490193845716086,
      "tokens_seen": 1632250880
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002553059177532598,
      "loss": 2.8687,
      "theoretical_loss": 3.490181689896141,
      "tokens_seen": 1632316416
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.00025529588766298896,
      "loss": 2.5809,
      "theoretical_loss": 3.4901695347008777,
      "tokens_seen": 1632381952
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.00025528585757271814,
      "loss": 2.4463,
      "theoretical_loss": 3.4901573801302384,
      "tokens_seen": 1632447488
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002552758274824473,
      "loss": 2.5499,
      "theoretical_loss": 3.4901452261841666,
      "tokens_seen": 1632513024
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.00025526579739217656,
      "loss": 2.5645,
      "theoretical_loss": 3.4901330728626045,
      "tokens_seen": 1632578560
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002552557673019057,
      "loss": 2.341,
      "theoretical_loss": 3.490120920165496,
      "tokens_seen": 1632644096
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002552457372116349,
      "loss": 2.5379,
      "theoretical_loss": 3.490108768092782,
      "tokens_seen": 1632709632
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002552357071213641,
      "loss": 2.6763,
      "theoretical_loss": 3.490096616644407,
      "tokens_seen": 1632775168
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002552256770310933,
      "loss": 2.8082,
      "theoretical_loss": 3.4900844658203134,
      "tokens_seen": 1632840704
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002552156469408225,
      "loss": 2.6397,
      "theoretical_loss": 3.4900723156204445,
      "tokens_seen": 1632906240
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.00025520561685055164,
      "loss": 2.4946,
      "theoretical_loss": 3.4900601660447426,
      "tokens_seen": 1632971776
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002551955867602809,
      "loss": 2.6199,
      "theoretical_loss": 3.4900480170931507,
      "tokens_seen": 1633037312
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.00025518555667001006,
      "loss": 2.5152,
      "theoretical_loss": 3.4900358687656112,
      "tokens_seen": 1633102848
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.00025517552657973924,
      "loss": 2.6778,
      "theoretical_loss": 3.4900237210620686,
      "tokens_seen": 1633168384
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002551654964894684,
      "loss": 2.4758,
      "theoretical_loss": 3.4900115739824638,
      "tokens_seen": 1633233920
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002551554663991976,
      "loss": 2.4492,
      "theoretical_loss": 3.489999427526741,
      "tokens_seen": 1633299456
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002551454363089268,
      "loss": 2.8651,
      "theoretical_loss": 3.4899872816948427,
      "tokens_seen": 1633364992
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.000255135406218656,
      "loss": 2.7476,
      "theoretical_loss": 3.4899751364867124,
      "tokens_seen": 1633430528
    },
    {
      "epoch": 5.05,
      "objective/train/docs_used": 1826734,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7838215827941895,
      "objective/train/theoretical_loss": 3.4899660279899276,
      "objective/train/tokens_used": 1653939680,
      "theoretical_loss": 3.4899660279899276,
      "tokens_seen": 1633479680
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.00025512537612838515,
      "loss": 2.687,
      "theoretical_loss": 3.489962991902292,
      "tokens_seen": 1633496064
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002551153460381144,
      "loss": 2.6393,
      "theoretical_loss": 3.4899508479415253,
      "tokens_seen": 1633561600
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002551053159478435,
      "loss": 2.8947,
      "theoretical_loss": 3.489938704604355,
      "tokens_seen": 1633627136
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.00025509528585757274,
      "loss": 2.7382,
      "theoretical_loss": 3.4899265618907243,
      "tokens_seen": 1633692672
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002550852557673019,
      "loss": 2.7037,
      "theoretical_loss": 3.4899144198005754,
      "tokens_seen": 1633758208
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002550752256770311,
      "loss": 2.8429,
      "theoretical_loss": 3.4899022783338522,
      "tokens_seen": 1633823744
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002550651955867603,
      "loss": 2.5956,
      "theoretical_loss": 3.489890137490497,
      "tokens_seen": 1633889280
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.00025505516549648947,
      "loss": 2.5744,
      "theoretical_loss": 3.4898779972704537,
      "tokens_seen": 1633954816
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.00025504513540621865,
      "loss": 2.344,
      "theoretical_loss": 3.489865857673664,
      "tokens_seen": 1634020352
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002550351053159479,
      "loss": 2.459,
      "theoretical_loss": 3.489853718700072,
      "tokens_seen": 1634085888
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.000255025075225677,
      "loss": 2.4011,
      "theoretical_loss": 3.48984158034962,
      "tokens_seen": 1634151424
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.00025501504513540625,
      "loss": 2.6465,
      "theoretical_loss": 3.489829442622251,
      "tokens_seen": 1634216960
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.00025500501504513543,
      "loss": 2.6639,
      "theoretical_loss": 3.489817305517909,
      "tokens_seen": 1634282496
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002549949849548646,
      "loss": 2.5077,
      "theoretical_loss": 3.4898051690365364,
      "tokens_seen": 1634348032
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002549849548645938,
      "loss": 2.3277,
      "theoretical_loss": 3.489793033178076,
      "tokens_seen": 1634413568
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.00025497492477432297,
      "loss": 2.8964,
      "theoretical_loss": 3.489780897942471,
      "tokens_seen": 1634479104
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.00025496489468405215,
      "loss": 2.5993,
      "theoretical_loss": 3.4897687633296646,
      "tokens_seen": 1634544640
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002549548645937814,
      "loss": 2.5346,
      "theoretical_loss": 3.4897566293396,
      "tokens_seen": 1634610176
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002549448345035105,
      "loss": 2.6303,
      "theoretical_loss": 3.4897444959722197,
      "tokens_seen": 1634675712
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.00025493480441323975,
      "loss": 2.3833,
      "theoretical_loss": 3.4897323632274677,
      "tokens_seen": 1634741248
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002549247743229689,
      "loss": 2.7033,
      "theoretical_loss": 3.489720231105286,
      "tokens_seen": 1634806784
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002549147442326981,
      "loss": 2.5677,
      "theoretical_loss": 3.489708099605619,
      "tokens_seen": 1634872320
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002549047141424273,
      "loss": 2.3206,
      "theoretical_loss": 3.489695968728409,
      "tokens_seen": 1634937856
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002548946840521565,
      "loss": 2.5654,
      "theoretical_loss": 3.4896838384735984,
      "tokens_seen": 1635003392
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.00025488465396188566,
      "loss": 2.5729,
      "theoretical_loss": 3.4896717088411315,
      "tokens_seen": 1635068928
    },
    {
      "epoch": 5.05,
      "objective/train/docs_used": 1827286,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7116994857788086,
      "objective/train/theoretical_loss": 3.4896626120251604,
      "objective/train/tokens_used": 1655578080,
      "theoretical_loss": 3.4896626120251604,
      "tokens_seen": 1635118080
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.00025487462387161484,
      "loss": 2.5476,
      "theoretical_loss": 3.4896595798309513,
      "tokens_seen": 1635134464
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.000254864593781344,
      "loss": 2.3753,
      "theoretical_loss": 3.4896474514430005,
      "tokens_seen": 1635200000
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.00025485456369107325,
      "loss": 2.6086,
      "theoretical_loss": 3.4896353236772226,
      "tokens_seen": 1635265536
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002548445336008024,
      "loss": 2.5868,
      "theoretical_loss": 3.48962319653356,
      "tokens_seen": 1635331072
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002548345035105316,
      "loss": 2.7026,
      "theoretical_loss": 3.489611070011957,
      "tokens_seen": 1635396608
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002548244734202608,
      "loss": 2.662,
      "theoretical_loss": 3.489598944112356,
      "tokens_seen": 1635462144
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.00025481444332999,
      "loss": 2.7653,
      "theoretical_loss": 3.4895868188347,
      "tokens_seen": 1635527680
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.00025480441323971916,
      "loss": 2.5967,
      "theoretical_loss": 3.489574694178933,
      "tokens_seen": 1635593216
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.00025479438314944834,
      "loss": 2.6639,
      "theoretical_loss": 3.4895625701449977,
      "tokens_seen": 1635658752
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002547843530591775,
      "loss": 2.5723,
      "theoretical_loss": 3.4895504467328373,
      "tokens_seen": 1635724288
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.00025477432296890676,
      "loss": 2.4181,
      "theoretical_loss": 3.489538323942395,
      "tokens_seen": 1635789824
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002547642928786359,
      "loss": 2.5984,
      "theoretical_loss": 3.489526201773614,
      "tokens_seen": 1635855360
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002547542627883651,
      "loss": 2.7379,
      "theoretical_loss": 3.4895140802264377,
      "tokens_seen": 1635920896
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.00025474423269809425,
      "loss": 2.6937,
      "theoretical_loss": 3.489501959300809,
      "tokens_seen": 1635986432
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002547342026078235,
      "loss": 2.6661,
      "theoretical_loss": 3.489489838996671,
      "tokens_seen": 1636051968
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.00025472417251755266,
      "loss": 2.8623,
      "theoretical_loss": 3.4894777193139683,
      "tokens_seen": 1636117504
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.00025471414242728184,
      "loss": 2.8822,
      "theoretical_loss": 3.4894656002526423,
      "tokens_seen": 1636183040
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.000254704112337011,
      "loss": 2.4939,
      "theoretical_loss": 3.489453481812637,
      "tokens_seen": 1636248576
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.00025469408224674026,
      "loss": 2.9008,
      "theoretical_loss": 3.4894413639938957,
      "tokens_seen": 1636314112
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002546840521564694,
      "loss": 2.754,
      "theoretical_loss": 3.4894292467963623,
      "tokens_seen": 1636379648
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002546740220661986,
      "loss": 2.4276,
      "theoretical_loss": 3.4894171302199792,
      "tokens_seen": 1636445184
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.00025466399197592775,
      "loss": 2.67,
      "theoretical_loss": 3.4894050142646895,
      "tokens_seen": 1636510720
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.000254653961885657,
      "loss": 2.4721,
      "theoretical_loss": 3.4893928989304372,
      "tokens_seen": 1636576256
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.00025464393179538616,
      "loss": 2.9586,
      "theoretical_loss": 3.4893807842171656,
      "tokens_seen": 1636641792
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.00025463390170511535,
      "loss": 2.6099,
      "theoretical_loss": 3.4893686701248177,
      "tokens_seen": 1636707328
    },
    {
      "epoch": 5.05,
      "objective/train/docs_used": 1827288,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.952646493911743,
      "objective/train/theoretical_loss": 3.4893595849630037,
      "objective/train/tokens_used": 1657216480,
      "theoretical_loss": 3.4893595849630037,
      "tokens_seen": 1636756480
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.00025462387161484453,
      "loss": 2.7477,
      "theoretical_loss": 3.4893565566533367,
      "tokens_seen": 1636772864
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002546138415245737,
      "loss": 2.6074,
      "theoretical_loss": 3.489344443802666,
      "tokens_seen": 1636838400
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002546038114343029,
      "loss": 2.8813,
      "theoretical_loss": 3.4893323315727494,
      "tokens_seen": 1636903936
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002545937813440321,
      "loss": 2.5295,
      "theoretical_loss": 3.4893202199635294,
      "tokens_seen": 1636969472
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.00025458375125376125,
      "loss": 2.7456,
      "theoretical_loss": 3.4893081089749503,
      "tokens_seen": 1637035008
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002545737211634905,
      "loss": 2.6556,
      "theoretical_loss": 3.4892959986069547,
      "tokens_seen": 1637100544
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002545636910732196,
      "loss": 2.7824,
      "theoretical_loss": 3.489283888859486,
      "tokens_seen": 1637166080
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.00025455366098294885,
      "loss": 2.7233,
      "theoretical_loss": 3.489271779732488,
      "tokens_seen": 1637231616
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.00025454363089267803,
      "loss": 2.649,
      "theoretical_loss": 3.489259671225904,
      "tokens_seen": 1637297152
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002545336008024072,
      "loss": 2.5388,
      "theoretical_loss": 3.489247563339677,
      "tokens_seen": 1637362688
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002545235707121364,
      "loss": 2.5585,
      "theoretical_loss": 3.48923545607375,
      "tokens_seen": 1637428224
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.00025451354062186563,
      "loss": 2.5326,
      "theoretical_loss": 3.489223349428068,
      "tokens_seen": 1637493760
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002545035105315948,
      "loss": 2.5587,
      "theoretical_loss": 3.4892112434025733,
      "tokens_seen": 1637559296
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.000254493480441324,
      "loss": 2.8104,
      "theoretical_loss": 3.4891991379972094,
      "tokens_seen": 1637624832
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.00025448345035105317,
      "loss": 2.6764,
      "theoretical_loss": 3.4891870332119193,
      "tokens_seen": 1637690368
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.00025447342026078235,
      "loss": 2.5624,
      "theoretical_loss": 3.489174929046647,
      "tokens_seen": 1637755904
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002544633901705116,
      "loss": 2.5065,
      "theoretical_loss": 3.489162825501336,
      "tokens_seen": 1637821440
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002544533600802407,
      "loss": 2.4074,
      "theoretical_loss": 3.4891507225759297,
      "tokens_seen": 1637886976
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.00025444332998996995,
      "loss": 2.6044,
      "theoretical_loss": 3.4891386202703716,
      "tokens_seen": 1637952512
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002544332998996991,
      "loss": 2.516,
      "theoretical_loss": 3.4891265185846043,
      "tokens_seen": 1638018048
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002544232698094283,
      "loss": 2.4602,
      "theoretical_loss": 3.4891144175185724,
      "tokens_seen": 1638083584
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002544132397191575,
      "loss": 2.6896,
      "theoretical_loss": 3.4891023170722186,
      "tokens_seen": 1638149120
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002544032096288867,
      "loss": 2.7171,
      "theoretical_loss": 3.4890902172454865,
      "tokens_seen": 1638214656
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.00025439317953861586,
      "loss": 2.6357,
      "theoretical_loss": 3.4890781180383206,
      "tokens_seen": 1638280192
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.00025438314944834504,
      "loss": 2.5777,
      "theoretical_loss": 3.4890660194506626,
      "tokens_seen": 1638345728
    },
    {
      "debugging/Self-BLEU-5": 0.28035211984762975,
      "debugging/distinct-1-grams": 0.8180070273723525,
      "debugging/distinct-2-grams": 0.9661205773003526,
      "debugging/entropy-1-grams": 5.418256584988372,
      "debugging/entropy-2-grams": 5.950145975632809,
      "debugging/length": 475.1666666666667,
      "debugging/num_segments": 6,
      "debugging/score": 0.006501734263504407,
      "debugging/score_std": 0.00880767366589508,
      "epoch": 5.05,
      "objective/train/docs_used": 1828052,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.91074538230896,
      "objective/train/theoretical_loss": 3.4890569459164382,
      "objective/train/tokens_used": 1658854880,
      "theoretical_loss": 3.4890569459164382,
      "tokens_seen": 1638394880
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002543731193580742,
      "loss": 2.7252,
      "theoretical_loss": 3.4890539214824576,
      "tokens_seen": 1638411264
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.00025436308926780345,
      "loss": 2.8064,
      "theoretical_loss": 3.489041824133648,
      "tokens_seen": 1638476800
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002543530591775326,
      "loss": 2.5308,
      "theoretical_loss": 3.489029727404178,
      "tokens_seen": 1638542336
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002543430290872618,
      "loss": 2.651,
      "theoretical_loss": 3.489017631293991,
      "tokens_seen": 1638607872
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.000254332998996991,
      "loss": 2.3064,
      "theoretical_loss": 3.48900553580303,
      "tokens_seen": 1638673408
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002543229689067202,
      "loss": 2.8929,
      "theoretical_loss": 3.4889934409312398,
      "tokens_seen": 1638738944
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.00025431293881644936,
      "loss": 2.516,
      "theoretical_loss": 3.488981346678562,
      "tokens_seen": 1638804480
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.00025430290872617854,
      "loss": 2.5252,
      "theoretical_loss": 3.4889692530449423,
      "tokens_seen": 1638870016
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002542928786359077,
      "loss": 2.4958,
      "theoretical_loss": 3.4889571600303224,
      "tokens_seen": 1638935552
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.00025428284854563696,
      "loss": 2.6536,
      "theoretical_loss": 3.4889450676346474,
      "tokens_seen": 1639001088
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002542728184553661,
      "loss": 2.7862,
      "theoretical_loss": 3.4889329758578596,
      "tokens_seen": 1639066624
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002542627883650953,
      "loss": 2.6614,
      "theoretical_loss": 3.4889208846999034,
      "tokens_seen": 1639132160
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.00025425275827482445,
      "loss": 2.631,
      "theoretical_loss": 3.488908794160722,
      "tokens_seen": 1639197696
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002542427281845537,
      "loss": 2.7136,
      "theoretical_loss": 3.4888967042402594,
      "tokens_seen": 1639263232
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.00025423269809428286,
      "loss": 2.5916,
      "theoretical_loss": 3.4888846149384585,
      "tokens_seen": 1639328768
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.00025422266800401204,
      "loss": 2.4989,
      "theoretical_loss": 3.4888725262552636,
      "tokens_seen": 1639394304
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002542126379137412,
      "loss": 2.7752,
      "theoretical_loss": 3.488860438190618,
      "tokens_seen": 1639459840
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.00025420260782347046,
      "loss": 2.7072,
      "theoretical_loss": 3.4888483507444654,
      "tokens_seen": 1639525376
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002541925777331996,
      "loss": 2.6852,
      "theoretical_loss": 3.4888362639167494,
      "tokens_seen": 1639590912
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002541825476429288,
      "loss": 2.4922,
      "theoretical_loss": 3.488824177707414,
      "tokens_seen": 1639656448
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.00025417251755265795,
      "loss": 2.611,
      "theoretical_loss": 3.488812092116402,
      "tokens_seen": 1639721984
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002541624874623872,
      "loss": 2.8802,
      "theoretical_loss": 3.488800007143658,
      "tokens_seen": 1639787520
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.00025415245737211637,
      "loss": 2.7579,
      "theoretical_loss": 3.488787922789125,
      "tokens_seen": 1639853056
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.00025414242728184555,
      "loss": 2.868,
      "theoretical_loss": 3.488775839052747,
      "tokens_seen": 1639918592
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.00025413239719157473,
      "loss": 2.8075,
      "theoretical_loss": 3.4887637559344675,
      "tokens_seen": 1639984128
    },
    {
      "epoch": 5.05,
      "objective/train/docs_used": 1828793,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.076061725616455,
      "objective/train/theoretical_loss": 3.4887546940013507,
      "objective/train/tokens_used": 1660493280,
      "theoretical_loss": 3.4887546940013507,
      "tokens_seen": 1640033280
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002541223671013039,
      "loss": 2.544,
      "theoretical_loss": 3.48875167343423,
      "tokens_seen": 1640049664
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002541123370110331,
      "loss": 2.5412,
      "theoretical_loss": 3.4887395915519788,
      "tokens_seen": 1640115200
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002541023069207623,
      "loss": 2.5361,
      "theoretical_loss": 3.4887275102876574,
      "tokens_seen": 1640180736
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.00025409227683049145,
      "loss": 2.9203,
      "theoretical_loss": 3.488715429641209,
      "tokens_seen": 1640246272
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002540822467402207,
      "loss": 2.5023,
      "theoretical_loss": 3.488703349612578,
      "tokens_seen": 1640311808
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002540722166499498,
      "loss": 2.7101,
      "theoretical_loss": 3.4886912702017074,
      "tokens_seen": 1640377344
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.00025406218655967905,
      "loss": 2.7177,
      "theoretical_loss": 3.4886791914085418,
      "tokens_seen": 1640442880
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.00025405215646940823,
      "loss": 2.6426,
      "theoretical_loss": 3.488667113233024,
      "tokens_seen": 1640508416
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002540421263791374,
      "loss": 2.8069,
      "theoretical_loss": 3.488655035675098,
      "tokens_seen": 1640573952
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002540320962888666,
      "loss": 2.714,
      "theoretical_loss": 3.4886429587347085,
      "tokens_seen": 1640639488
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.00025402206619859583,
      "loss": 2.6145,
      "theoretical_loss": 3.4886308824117984,
      "tokens_seen": 1640705024
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.00025401203610832495,
      "loss": 2.6574,
      "theoretical_loss": 3.4886188067063113,
      "tokens_seen": 1640770560
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002540020060180542,
      "loss": 2.4431,
      "theoretical_loss": 3.4886067316181917,
      "tokens_seen": 1640836096
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002539919759277833,
      "loss": 2.7695,
      "theoretical_loss": 3.4885946571473827,
      "tokens_seen": 1640901632
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.00025398194583751255,
      "loss": 2.7228,
      "theoretical_loss": 3.488582583293828,
      "tokens_seen": 1640967168
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.00025397191574724173,
      "loss": 2.6124,
      "theoretical_loss": 3.488570510057472,
      "tokens_seen": 1641032704
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002539618856569709,
      "loss": 2.6302,
      "theoretical_loss": 3.4885584374382583,
      "tokens_seen": 1641098240
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002539518555667001,
      "loss": 2.6826,
      "theoretical_loss": 3.4885463654361306,
      "tokens_seen": 1641163776
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002539418254764293,
      "loss": 2.5072,
      "theoretical_loss": 3.488534294051033,
      "tokens_seen": 1641229312
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.00025393179538615846,
      "loss": 2.5015,
      "theoretical_loss": 3.4885222232829083,
      "tokens_seen": 1641294848
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002539217652958877,
      "loss": 2.7234,
      "theoretical_loss": 3.4885101531317018,
      "tokens_seen": 1641360384
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002539117352056168,
      "loss": 2.8709,
      "theoretical_loss": 3.4884980835973565,
      "tokens_seen": 1641425920
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.00025390170511534606,
      "loss": 2.6931,
      "theoretical_loss": 3.4884860146798164,
      "tokens_seen": 1641491456
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002538916750250752,
      "loss": 2.4736,
      "theoretical_loss": 3.488473946379025,
      "tokens_seen": 1641556992
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002538816449348044,
      "loss": 2.6352,
      "theoretical_loss": 3.488461878694927,
      "tokens_seen": 1641622528
    },
    {
      "epoch": 5.05,
      "objective/train/docs_used": 1830032,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.0165646076202393,
      "objective/train/theoretical_loss": 3.4884528283365244,
      "objective/train/tokens_used": 1662131680,
      "theoretical_loss": 3.4884528283365244,
      "tokens_seen": 1641671680
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002538716148445336,
      "loss": 2.8231,
      "theoretical_loss": 3.488449811627466,
      "tokens_seen": 1641688064
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002538615847542628,
      "loss": 2.7286,
      "theoretical_loss": 3.4884377451765847,
      "tokens_seen": 1641753600
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.00025385155466399196,
      "loss": 2.6162,
      "theoretical_loss": 3.488425679342229,
      "tokens_seen": 1641819136
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002538415245737212,
      "loss": 2.6865,
      "theoretical_loss": 3.488413614124341,
      "tokens_seen": 1641884672
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002538314944834503,
      "loss": 2.6419,
      "theoretical_loss": 3.488401549522866,
      "tokens_seen": 1641950208
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.00025382146439317956,
      "loss": 2.6924,
      "theoretical_loss": 3.4883894855377466,
      "tokens_seen": 1642015744
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002538114343029087,
      "loss": 2.498,
      "theoretical_loss": 3.488377422168928,
      "tokens_seen": 1642081280
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002538014042126379,
      "loss": 2.7008,
      "theoretical_loss": 3.4883653594163535,
      "tokens_seen": 1642146816
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002537913741223671,
      "loss": 2.6494,
      "theoretical_loss": 3.4883532972799665,
      "tokens_seen": 1642212352
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002537813440320963,
      "loss": 2.5609,
      "theoretical_loss": 3.4883412357597114,
      "tokens_seen": 1642277888
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.00025377131394182546,
      "loss": 2.6344,
      "theoretical_loss": 3.4883291748555325,
      "tokens_seen": 1642343424
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.00025376128385155465,
      "loss": 2.6554,
      "theoretical_loss": 3.4883171145673737,
      "tokens_seen": 1642408960
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002537512537612839,
      "loss": 2.6871,
      "theoretical_loss": 3.4883050548951777,
      "tokens_seen": 1642474496
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.00025374122367101306,
      "loss": 2.554,
      "theoretical_loss": 3.488292995838891,
      "tokens_seen": 1642540032
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.00025373119358074224,
      "loss": 2.6946,
      "theoretical_loss": 3.4882809373984545,
      "tokens_seen": 1642605568
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002537211634904714,
      "loss": 2.7,
      "theoretical_loss": 3.4882688795738144,
      "tokens_seen": 1642671104
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.00025371113340020066,
      "loss": 2.8572,
      "theoretical_loss": 3.488256822364914,
      "tokens_seen": 1642736640
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002537011033099298,
      "loss": 2.7233,
      "theoretical_loss": 3.4882447657716975,
      "tokens_seen": 1642802176
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.000253691073219659,
      "loss": 2.5799,
      "theoretical_loss": 3.4882327097941084,
      "tokens_seen": 1642867712
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.00025368104312938815,
      "loss": 2.6587,
      "theoretical_loss": 3.4882206544320913,
      "tokens_seen": 1642933248
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002536710130391174,
      "loss": 2.5166,
      "theoretical_loss": 3.4882085996855894,
      "tokens_seen": 1642998784
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.00025366098294884657,
      "loss": 2.5442,
      "theoretical_loss": 3.4881965455545476,
      "tokens_seen": 1643064320
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.00025365095285857575,
      "loss": 2.7119,
      "theoretical_loss": 3.4881844920389096,
      "tokens_seen": 1643129856
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.00025364092276830493,
      "loss": 2.648,
      "theoretical_loss": 3.4881724391386193,
      "tokens_seen": 1643195392
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002536308926780341,
      "loss": 2.7163,
      "theoretical_loss": 3.488160386853621,
      "tokens_seen": 1643260928
    },
    {
      "epoch": 5.05,
      "objective/train/docs_used": 1830619,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8538637161254883,
      "objective/train/theoretical_loss": 3.488151348043624,
      "objective/train/tokens_used": 1663770080,
      "theoretical_loss": 3.488151348043624,
      "tokens_seen": 1643310080
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002536208625877633,
      "loss": 2.7363,
      "theoretical_loss": 3.4881483351838583,
      "tokens_seen": 1643326464
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002536108324974925,
      "loss": 2.5418,
      "theoretical_loss": 3.488136284129276,
      "tokens_seen": 1643392000
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.00025360080240722165,
      "loss": 2.5203,
      "theoretical_loss": 3.4881242336898177,
      "tokens_seen": 1643457536
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002535907723169509,
      "loss": 2.7345,
      "theoretical_loss": 3.488112183865427,
      "tokens_seen": 1643523072
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.00025358074222668,
      "loss": 2.5824,
      "theoretical_loss": 3.488100134656049,
      "tokens_seen": 1643588608
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.00025357071213640925,
      "loss": 2.6666,
      "theoretical_loss": 3.4880880860616275,
      "tokens_seen": 1643654144
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.00025356068204613843,
      "loss": 2.7623,
      "theoretical_loss": 3.488076038082106,
      "tokens_seen": 1643719680
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002535506519558676,
      "loss": 2.6642,
      "theoretical_loss": 3.4880639907174285,
      "tokens_seen": 1643785216
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002535406218655968,
      "loss": 2.5737,
      "theoretical_loss": 3.48805194396754,
      "tokens_seen": 1643850752
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.00025353059177532603,
      "loss": 2.5808,
      "theoretical_loss": 3.488039897832385,
      "tokens_seen": 1643916288
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.00025352056168505516,
      "loss": 2.6431,
      "theoretical_loss": 3.488027852311906,
      "tokens_seen": 1643981824
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002535105315947844,
      "loss": 2.4755,
      "theoretical_loss": 3.488015807406048,
      "tokens_seen": 1644047360
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002535005015045135,
      "loss": 2.6564,
      "theoretical_loss": 3.4880037631147554,
      "tokens_seen": 1644112896
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.00025349047141424275,
      "loss": 2.6924,
      "theoretical_loss": 3.487991719437972,
      "tokens_seen": 1644178432
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.00025348044132397193,
      "loss": 2.6345,
      "theoretical_loss": 3.487979676375642,
      "tokens_seen": 1644243968
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002534704112337011,
      "loss": 2.4965,
      "theoretical_loss": 3.4879676339277097,
      "tokens_seen": 1644309504
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002534603811434303,
      "loss": 2.6393,
      "theoretical_loss": 3.487955592094119,
      "tokens_seen": 1644375040
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002534503510531595,
      "loss": 2.8411,
      "theoretical_loss": 3.487943550874814,
      "tokens_seen": 1644440576
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.00025344032096288866,
      "loss": 2.7036,
      "theoretical_loss": 3.487931510269739,
      "tokens_seen": 1644506112
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002534302908726179,
      "loss": 2.6083,
      "theoretical_loss": 3.4879194702788388,
      "tokens_seen": 1644571648
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.000253420260782347,
      "loss": 2.8604,
      "theoretical_loss": 3.487907430902057,
      "tokens_seen": 1644637184
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.00025341023069207626,
      "loss": 2.5739,
      "theoretical_loss": 3.487895392139338,
      "tokens_seen": 1644702720
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002534002006018054,
      "loss": 2.6247,
      "theoretical_loss": 3.4878833539906258,
      "tokens_seen": 1644768256
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002533901705115346,
      "loss": 2.6973,
      "theoretical_loss": 3.4878713164558643,
      "tokens_seen": 1644833792
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002533801404212638,
      "loss": 2.6411,
      "theoretical_loss": 3.487859279534999,
      "tokens_seen": 1644899328
    },
    {
      "epoch": 5.05,
      "objective/train/docs_used": 1831835,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8763232231140137,
      "objective/train/theoretical_loss": 3.487850252247185,
      "objective/train/tokens_used": 1665408480,
      "theoretical_loss": 3.487850252247185,
      "tokens_seen": 1644948480
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.000253370110330993,
      "loss": 2.8589,
      "theoretical_loss": 3.487847243227973,
      "tokens_seen": 1644964864
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.00025336008024072216,
      "loss": 2.5824,
      "theoretical_loss": 3.4878352075347308,
      "tokens_seen": 1645030400
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002533500501504514,
      "loss": 2.7122,
      "theoretical_loss": 3.487823172455217,
      "tokens_seen": 1645095936
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002533400200601805,
      "loss": 2.7071,
      "theoretical_loss": 3.487811137989375,
      "tokens_seen": 1645161472
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.00025332998996990976,
      "loss": 2.5932,
      "theoretical_loss": 3.4877991041371494,
      "tokens_seen": 1645227008
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002533199598796389,
      "loss": 2.7684,
      "theoretical_loss": 3.4877870708984853,
      "tokens_seen": 1645292544
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002533099297893681,
      "loss": 2.6536,
      "theoretical_loss": 3.4877750382733264,
      "tokens_seen": 1645358080
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002532998996990973,
      "loss": 2.5601,
      "theoretical_loss": 3.4877630062616167,
      "tokens_seen": 1645423616
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002532898696088265,
      "loss": 2.557,
      "theoretical_loss": 3.487750974863301,
      "tokens_seen": 1645489152
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.00025327983951855566,
      "loss": 2.7465,
      "theoretical_loss": 3.4877389440783233,
      "tokens_seen": 1645554688
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.00025326980942828485,
      "loss": 2.2815,
      "theoretical_loss": 3.4877269139066276,
      "tokens_seen": 1645620224
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.000253259779338014,
      "loss": 2.5423,
      "theoretical_loss": 3.4877148843481587,
      "tokens_seen": 1645685760
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.00025324974924774326,
      "loss": 2.7766,
      "theoretical_loss": 3.487702855402861,
      "tokens_seen": 1645751296
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002532397191574724,
      "loss": 2.895,
      "theoretical_loss": 3.4876908270706783,
      "tokens_seen": 1645816832
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002532296890672016,
      "loss": 2.8808,
      "theoretical_loss": 3.487678799351556,
      "tokens_seen": 1645882368
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.00025321965897693075,
      "loss": 2.5823,
      "theoretical_loss": 3.4876667722454373,
      "tokens_seen": 1645947904
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.00025320962888666,
      "loss": 2.641,
      "theoretical_loss": 3.487654745752267,
      "tokens_seen": 1646013440
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.00025319959879638917,
      "loss": 2.5569,
      "theoretical_loss": 3.487642719871989,
      "tokens_seen": 1646078976
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.00025318956870611835,
      "loss": 2.7729,
      "theoretical_loss": 3.4876306946045483,
      "tokens_seen": 1646144512
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.00025317953861584753,
      "loss": 2.7548,
      "theoretical_loss": 3.4876186699498897,
      "tokens_seen": 1646210048
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.00025316950852557677,
      "loss": 2.6345,
      "theoretical_loss": 3.487606645907956,
      "tokens_seen": 1646275584
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002531594784353059,
      "loss": 2.5346,
      "theoretical_loss": 3.487594622478693,
      "tokens_seen": 1646341120
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.00025314944834503513,
      "loss": 2.7141,
      "theoretical_loss": 3.487582599662045,
      "tokens_seen": 1646406656
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.00025313941825476425,
      "loss": 2.7156,
      "theoretical_loss": 3.4875705774579555,
      "tokens_seen": 1646472192
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002531293881644935,
      "loss": 2.6339,
      "theoretical_loss": 3.4875585558663698,
      "tokens_seen": 1646537728
    },
    {
      "epoch": 5.05,
      "objective/train/docs_used": 1832689,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.549023389816284,
      "objective/train/theoretical_loss": 3.4875495400746024,
      "objective/train/tokens_used": 1667046880,
      "theoretical_loss": 3.4875495400746024,
      "tokens_seen": 1646586880
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.00025311935807422267,
      "loss": 2.5987,
      "theoretical_loss": 3.4875465348872314,
      "tokens_seen": 1646603264
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.00025310932798395185,
      "loss": 2.8123,
      "theoretical_loss": 3.487534514520486,
      "tokens_seen": 1646668800
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.00025309929789368103,
      "loss": 2.6352,
      "theoretical_loss": 3.487522494766077,
      "tokens_seen": 1646734336
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002530892678034102,
      "loss": 2.5625,
      "theoretical_loss": 3.4875104756239486,
      "tokens_seen": 1646799872
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002530792377131394,
      "loss": 2.5651,
      "theoretical_loss": 3.4874984570940466,
      "tokens_seen": 1646865408
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.00025306920762286863,
      "loss": 2.6599,
      "theoretical_loss": 3.4874864391763145,
      "tokens_seen": 1646930944
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.00025305917753259776,
      "loss": 2.5127,
      "theoretical_loss": 3.4874744218706963,
      "tokens_seen": 1646996480
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.000253049147442327,
      "loss": 2.4157,
      "theoretical_loss": 3.487462405177138,
      "tokens_seen": 1647062016
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002530391173520562,
      "loss": 2.5339,
      "theoretical_loss": 3.4874503890955824,
      "tokens_seen": 1647127552
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.00025302908726178536,
      "loss": 2.5977,
      "theoretical_loss": 3.487438373625975,
      "tokens_seen": 1647193088
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.00025301905717151454,
      "loss": 2.7327,
      "theoretical_loss": 3.48742635876826,
      "tokens_seen": 1647258624
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002530090270812437,
      "loss": 2.7865,
      "theoretical_loss": 3.4874143445223824,
      "tokens_seen": 1647324160
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.00025299899699097295,
      "loss": 2.6948,
      "theoretical_loss": 3.4874023308882856,
      "tokens_seen": 1647389696
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.00025298896690070213,
      "loss": 2.7051,
      "theoretical_loss": 3.487390317865915,
      "tokens_seen": 1647455232
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002529789368104313,
      "loss": 2.6647,
      "theoretical_loss": 3.4873783054552145,
      "tokens_seen": 1647520768
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002529689067201605,
      "loss": 2.564,
      "theoretical_loss": 3.4873662936561294,
      "tokens_seen": 1647586304
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002529588766298897,
      "loss": 2.7542,
      "theoretical_loss": 3.487354282468604,
      "tokens_seen": 1647651840
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.00025294884653961886,
      "loss": 2.3254,
      "theoretical_loss": 3.487342271892582,
      "tokens_seen": 1647717376
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002529388164493481,
      "loss": 2.5738,
      "theoretical_loss": 3.487330261928009,
      "tokens_seen": 1647782912
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002529287863590772,
      "loss": 2.679,
      "theoretical_loss": 3.4873182525748296,
      "tokens_seen": 1647848448
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.00025291875626880646,
      "loss": 2.4659,
      "theoretical_loss": 3.4873062438329874,
      "tokens_seen": 1647913984
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002529087261785356,
      "loss": 2.5778,
      "theoretical_loss": 3.4872942357024277,
      "tokens_seen": 1647979520
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002528986960882648,
      "loss": 2.4391,
      "theoretical_loss": 3.4872822281830946,
      "tokens_seen": 1648045056
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.000252888665997994,
      "loss": 2.6824,
      "theoretical_loss": 3.487270221274933,
      "tokens_seen": 1648110592
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002528786359077232,
      "loss": 2.7971,
      "theoretical_loss": 3.4872582149778877,
      "tokens_seen": 1648176128
    },
    {
      "epoch": 5.05,
      "objective/train/docs_used": 1834029,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5318849086761475,
      "objective/train/theoretical_loss": 3.487249210656115,
      "objective/train/tokens_used": 1668685280,
      "theoretical_loss": 3.487249210656115,
      "tokens_seen": 1648225280
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.00025286860581745236,
      "loss": 2.652,
      "theoretical_loss": 3.4872462092919028,
      "tokens_seen": 1648241664
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002528585757271816,
      "loss": 2.6296,
      "theoretical_loss": 3.4872342042169233,
      "tokens_seen": 1648307200
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002528485456369107,
      "loss": 2.4941,
      "theoretical_loss": 3.4872221997528934,
      "tokens_seen": 1648372736
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.00025283851554663996,
      "loss": 2.4465,
      "theoretical_loss": 3.4872101958997583,
      "tokens_seen": 1648438272
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002528284854563691,
      "loss": 2.3638,
      "theoretical_loss": 3.4871981926574622,
      "tokens_seen": 1648503808
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002528184553660983,
      "loss": 2.703,
      "theoretical_loss": 3.48718619002595,
      "tokens_seen": 1648569344
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002528084252758275,
      "loss": 2.6302,
      "theoretical_loss": 3.4871741880051657,
      "tokens_seen": 1648634880
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002527983951855567,
      "loss": 2.5522,
      "theoretical_loss": 3.487162186595055,
      "tokens_seen": 1648700416
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.00025278836509528586,
      "loss": 2.687,
      "theoretical_loss": 3.487150185795562,
      "tokens_seen": 1648765952
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.00025277833500501505,
      "loss": 2.5756,
      "theoretical_loss": 3.487138185606631,
      "tokens_seen": 1648831488
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002527683049147442,
      "loss": 2.7657,
      "theoretical_loss": 3.487126186028207,
      "tokens_seen": 1648897024
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.00025275827482447346,
      "loss": 2.6197,
      "theoretical_loss": 3.487114187060235,
      "tokens_seen": 1648962560
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002527482447342026,
      "loss": 2.3855,
      "theoretical_loss": 3.487102188702659,
      "tokens_seen": 1649028096
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002527382146439318,
      "loss": 2.6836,
      "theoretical_loss": 3.487090190955424,
      "tokens_seen": 1649093632
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.00025272818455366095,
      "loss": 2.6625,
      "theoretical_loss": 3.4870781938184754,
      "tokens_seen": 1649159168
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002527181544633902,
      "loss": 2.639,
      "theoretical_loss": 3.4870661972917567,
      "tokens_seen": 1649224704
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.00025270812437311937,
      "loss": 2.6155,
      "theoretical_loss": 3.4870542013752135,
      "tokens_seen": 1649290240
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.00025269809428284855,
      "loss": 2.753,
      "theoretical_loss": 3.4870422060687902,
      "tokens_seen": 1649355776
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.00025268806419257773,
      "loss": 2.4928,
      "theoretical_loss": 3.4870302113724314,
      "tokens_seen": 1649421312
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.00025267803410230697,
      "loss": 2.6719,
      "theoretical_loss": 3.4870182172860824,
      "tokens_seen": 1649486848
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002526680040120361,
      "loss": 2.5814,
      "theoretical_loss": 3.4870062238096873,
      "tokens_seen": 1649552384
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.00025265797392176533,
      "loss": 2.3619,
      "theoretical_loss": 3.486994230943191,
      "tokens_seen": 1649617920
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.00025264794383149445,
      "loss": 2.5583,
      "theoretical_loss": 3.4869822386865383,
      "tokens_seen": 1649683456
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002526379137412237,
      "loss": 2.5248,
      "theoretical_loss": 3.4869702470396735,
      "tokens_seen": 1649748992
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.00025262788365095287,
      "loss": 2.7687,
      "theoretical_loss": 3.4869582560025423,
      "tokens_seen": 1649814528
    },
    {
      "epoch": 5.05,
      "objective/train/docs_used": 1834679,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.2152130603790283,
      "objective/train/theoretical_loss": 3.4869492631247985,
      "objective/train/tokens_used": 1670323680,
      "theoretical_loss": 3.4869492631247985,
      "tokens_seen": 1649863680
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.00025261785356068205,
      "loss": 2.7612,
      "theoretical_loss": 3.4869462655750896,
      "tokens_seen": 1649880064
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.00025260782347041123,
      "loss": 2.6469,
      "theoretical_loss": 3.486934275757259,
      "tokens_seen": 1649945600
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002525977933801404,
      "loss": 2.5645,
      "theoretical_loss": 3.486922286548996,
      "tokens_seen": 1650011136
    },
    {
      "epoch": 5.05,
      "learning_rate": 0.0002525877632898696,
      "loss": 2.5006,
      "theoretical_loss": 3.4869102979502458,
      "tokens_seen": 1650076672
    }
  ],
  "max_steps": 50354,
  "num_train_epochs": 9223372036854775807,
  "total_flos": 8.42093528481792e+17,
  "trial_name": null,
  "trial_params": null
}