{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.9988571428571429,
  "eval_steps": 50,
  "global_step": 437,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.022857142857142857,
      "grad_norm": 2.9285875213659387,
      "learning_rate": 1.1363636363636363e-07,
      "logits/chosen": -2.6540167331695557,
      "logits/rejected": -2.6346607208251953,
      "logps/chosen": -287.88873291015625,
      "logps/rejected": -266.501953125,
      "loss": 0.0456,
      "rewards/accuracies": 0.4000000059604645,
      "rewards/chosen": 0.0002630561648402363,
      "rewards/margins": 0.000395975133869797,
      "rewards/rejected": -0.00013291896902956069,
      "step": 10
    },
    {
      "epoch": 0.045714285714285714,
      "grad_norm": 2.6703796663983237,
      "learning_rate": 2.2727272727272726e-07,
      "logits/chosen": -2.6553053855895996,
      "logits/rejected": -2.62392258644104,
      "logps/chosen": -271.43328857421875,
      "logps/rejected": -260.1735534667969,
      "loss": 0.0465,
      "rewards/accuracies": 0.550000011920929,
      "rewards/chosen": 0.0027677875477820635,
      "rewards/margins": 0.001332368585281074,
      "rewards/rejected": 0.0014354187296703458,
      "step": 20
    },
    {
      "epoch": 0.06857142857142857,
      "grad_norm": 2.4626318533205516,
      "learning_rate": 3.4090909090909085e-07,
      "logits/chosen": -2.65043568611145,
      "logits/rejected": -2.5593643188476562,
      "logps/chosen": -271.85040283203125,
      "logps/rejected": -254.72055053710938,
      "loss": 0.0437,
      "rewards/accuracies": 0.7562500238418579,
      "rewards/chosen": 0.020383944734930992,
      "rewards/margins": 0.012471409514546394,
      "rewards/rejected": 0.007912535220384598,
      "step": 30
    },
    {
      "epoch": 0.09142857142857143,
      "grad_norm": 2.1219802043829175,
      "learning_rate": 4.545454545454545e-07,
      "logits/chosen": -2.6147427558898926,
      "logits/rejected": -2.6001462936401367,
      "logps/chosen": -261.89312744140625,
      "logps/rejected": -255.43844604492188,
      "loss": 0.0409,
      "rewards/accuracies": 0.699999988079071,
      "rewards/chosen": 0.03881374001502991,
      "rewards/margins": 0.03865280747413635,
      "rewards/rejected": 0.0001609304890735075,
      "step": 40
    },
    {
      "epoch": 0.11428571428571428,
      "grad_norm": 2.648446772286267,
      "learning_rate": 4.997124959943201e-07,
      "logits/chosen": -2.5977413654327393,
      "logits/rejected": -2.530109405517578,
      "logps/chosen": -287.03973388671875,
      "logps/rejected": -261.2123718261719,
      "loss": 0.0371,
      "rewards/accuracies": 0.768750011920929,
      "rewards/chosen": 0.024105368182063103,
      "rewards/margins": 0.09380187839269638,
      "rewards/rejected": -0.06969650834798813,
      "step": 50
    },
    {
      "epoch": 0.11428571428571428,
      "eval_logits/chosen": -2.6169474124908447,
      "eval_logits/rejected": -2.5727975368499756,
      "eval_logps/chosen": -258.3393249511719,
      "eval_logps/rejected": -256.71771240234375,
      "eval_loss": 0.032985787838697433,
      "eval_rewards/accuracies": 0.6637930870056152,
      "eval_rewards/chosen": 0.01621999777853489,
      "eval_rewards/margins": 0.10062520951032639,
      "eval_rewards/rejected": -0.08440519869327545,
      "eval_runtime": 90.7209,
      "eval_samples_per_second": 20.183,
      "eval_steps_per_second": 0.32,
      "step": 50
    },
    {
      "epoch": 0.13714285714285715,
      "grad_norm": 4.351801812595937,
      "learning_rate": 4.979579212164186e-07,
      "logits/chosen": -2.5900657176971436,
      "logits/rejected": -2.5341601371765137,
      "logps/chosen": -294.68328857421875,
      "logps/rejected": -255.0167999267578,
      "loss": 0.0338,
      "rewards/accuracies": 0.71875,
      "rewards/chosen": 0.048027873039245605,
      "rewards/margins": 0.13361422717571259,
      "rewards/rejected": -0.08558637648820877,
      "step": 60
    },
    {
      "epoch": 0.16,
      "grad_norm": 3.0187175488356703,
      "learning_rate": 4.946196886175515e-07,
      "logits/chosen": -2.539175510406494,
      "logits/rejected": -2.4328207969665527,
      "logps/chosen": -299.51702880859375,
      "logps/rejected": -263.10400390625,
      "loss": 0.0324,
      "rewards/accuracies": 0.706250011920929,
      "rewards/chosen": 0.03838000446557999,
      "rewards/margins": 0.1820719689130783,
      "rewards/rejected": -0.14369197189807892,
      "step": 70
    },
    {
      "epoch": 0.18285714285714286,
      "grad_norm": 2.829278868602528,
      "learning_rate": 4.897191188239667e-07,
      "logits/chosen": -2.503023147583008,
      "logits/rejected": -2.4128475189208984,
      "logps/chosen": -271.768310546875,
      "logps/rejected": -266.82537841796875,
      "loss": 0.0321,
      "rewards/accuracies": 0.6875,
      "rewards/chosen": -0.04572376608848572,
      "rewards/margins": 0.15518848598003387,
      "rewards/rejected": -0.2009122371673584,
      "step": 80
    },
    {
      "epoch": 0.2057142857142857,
      "grad_norm": 2.659426058179347,
      "learning_rate": 4.832875107981763e-07,
      "logits/chosen": -2.122408390045166,
      "logits/rejected": -2.066514730453491,
      "logps/chosen": -249.37106323242188,
      "logps/rejected": -279.49212646484375,
      "loss": 0.0311,
      "rewards/accuracies": 0.762499988079071,
      "rewards/chosen": -0.07777222990989685,
      "rewards/margins": 0.199189230799675,
      "rewards/rejected": -0.27696144580841064,
      "step": 90
    },
    {
      "epoch": 0.22857142857142856,
      "grad_norm": 3.4560236437659873,
      "learning_rate": 4.753659419387223e-07,
      "logits/chosen": -2.338256359100342,
      "logits/rejected": -2.29487943649292,
      "logps/chosen": -266.89556884765625,
      "logps/rejected": -254.134521484375,
      "loss": 0.0326,
      "rewards/accuracies": 0.75,
      "rewards/chosen": -0.056456904858350754,
      "rewards/margins": 0.17859290540218353,
      "rewards/rejected": -0.23504984378814697,
      "step": 100
    },
    {
      "epoch": 0.22857142857142856,
      "eval_logits/chosen": -2.3468899726867676,
      "eval_logits/rejected": -2.2908225059509277,
      "eval_logps/chosen": -276.0060729980469,
      "eval_logps/rejected": -285.9020080566406,
      "eval_loss": 0.02837423048913479,
      "eval_rewards/accuracies": 0.7068965435028076,
      "eval_rewards/chosen": -0.16044749319553375,
      "eval_rewards/margins": 0.21580053865909576,
      "eval_rewards/rejected": -0.3762480318546295,
      "eval_runtime": 90.6197,
      "eval_samples_per_second": 20.205,
      "eval_steps_per_second": 0.32,
      "step": 100
    },
    {
      "epoch": 0.25142857142857145,
      "grad_norm": 3.986741503860454,
      "learning_rate": 4.660050057270191e-07,
      "logits/chosen": -2.414931058883667,
      "logits/rejected": -2.3160946369171143,
      "logps/chosen": -314.08258056640625,
      "logps/rejected": -264.74102783203125,
      "loss": 0.0274,
      "rewards/accuracies": 0.768750011920929,
      "rewards/chosen": -0.07821311056613922,
      "rewards/margins": 0.22488529980182648,
      "rewards/rejected": -0.3030984103679657,
      "step": 110
    },
    {
      "epoch": 0.2742857142857143,
      "grad_norm": 3.0310867108282418,
      "learning_rate": 4.5526448859687144e-07,
      "logits/chosen": -2.4880690574645996,
      "logits/rejected": -2.4331300258636475,
      "logps/chosen": -281.57952880859375,
      "logps/rejected": -279.35137939453125,
      "loss": 0.0292,
      "rewards/accuracies": 0.7437499761581421,
      "rewards/chosen": -0.08748000860214233,
      "rewards/margins": 0.20896463096141815,
      "rewards/rejected": -0.2964446544647217,
      "step": 120
    },
    {
      "epoch": 0.29714285714285715,
      "grad_norm": 2.686160441272964,
      "learning_rate": 4.432129880904388e-07,
      "logits/chosen": -2.505074977874756,
      "logits/rejected": -2.4108517169952393,
      "logps/chosen": -281.3280944824219,
      "logps/rejected": -264.3760986328125,
      "loss": 0.0271,
      "rewards/accuracies": 0.625,
      "rewards/chosen": -0.07846375554800034,
      "rewards/margins": 0.17637231945991516,
      "rewards/rejected": -0.2548360228538513,
      "step": 130
    },
    {
      "epoch": 0.32,
      "grad_norm": 2.7176410561935582,
      "learning_rate": 4.299274747394055e-07,
      "logits/chosen": -2.41599702835083,
      "logits/rejected": -2.3766045570373535,
      "logps/chosen": -328.3199157714844,
      "logps/rejected": -292.5992736816406,
      "loss": 0.0269,
      "rewards/accuracies": 0.7749999761581421,
      "rewards/chosen": -0.10486980527639389,
      "rewards/margins": 0.21833749115467072,
      "rewards/rejected": -0.323207288980484,
      "step": 140
    },
    {
      "epoch": 0.34285714285714286,
      "grad_norm": 2.624843592752447,
      "learning_rate": 4.1549280046953653e-07,
      "logits/chosen": -2.360866069793701,
      "logits/rejected": -2.3315412998199463,
      "logps/chosen": -309.66278076171875,
      "logps/rejected": -299.77325439453125,
      "loss": 0.0276,
      "rewards/accuracies": 0.7562500238418579,
      "rewards/chosen": -0.13608722388744354,
      "rewards/margins": 0.20585337281227112,
      "rewards/rejected": -0.34194058179855347,
      "step": 150
    },
    {
      "epoch": 0.34285714285714286,
      "eval_logits/chosen": -2.3992269039154053,
      "eval_logits/rejected": -2.353783369064331,
      "eval_logps/chosen": -274.219970703125,
      "eval_logps/rejected": -284.3463439941406,
      "eval_loss": 0.026143357157707214,
      "eval_rewards/accuracies": 0.7198275923728943,
      "eval_rewards/chosen": -0.1425865888595581,
      "eval_rewards/margins": 0.21810482442378998,
      "eval_rewards/rejected": -0.3606913983821869,
      "eval_runtime": 91.9073,
      "eval_samples_per_second": 19.922,
      "eval_steps_per_second": 0.316,
      "step": 150
    },
    {
      "epoch": 0.3657142857142857,
      "grad_norm": 5.477307060879475,
      "learning_rate": 4.000011566683401e-07,
      "logits/chosen": -2.4539685249328613,
      "logits/rejected": -2.354241371154785,
      "logps/chosen": -321.659423828125,
      "logps/rejected": -301.2185974121094,
      "loss": 0.0272,
      "rewards/accuracies": 0.7437499761581421,
      "rewards/chosen": -0.12505923211574554,
      "rewards/margins": 0.18981832265853882,
      "rewards/rejected": -0.31487756967544556,
      "step": 160
    },
    {
      "epoch": 0.38857142857142857,
      "grad_norm": 3.291959162082193,
      "learning_rate": 3.8355148537705047e-07,
      "logits/chosen": -2.4768872261047363,
      "logits/rejected": -2.4600863456726074,
      "logps/chosen": -300.3105773925781,
      "logps/rejected": -303.3681335449219,
      "loss": 0.0265,
      "rewards/accuracies": 0.78125,
      "rewards/chosen": -0.06613761186599731,
      "rewards/margins": 0.22872081398963928,
      "rewards/rejected": -0.2948583960533142,
      "step": 170
    },
    {
      "epoch": 0.4114285714285714,
      "grad_norm": 3.4863952801217724,
      "learning_rate": 3.662488473675315e-07,
      "logits/chosen": -2.496547222137451,
      "logits/rejected": -2.414828062057495,
      "logps/chosen": -312.77105712890625,
      "logps/rejected": -283.5912170410156,
      "loss": 0.0277,
      "rewards/accuracies": 0.8062499761581421,
      "rewards/chosen": -0.1033080667257309,
      "rewards/margins": 0.268028199672699,
      "rewards/rejected": -0.37133628129959106,
      "step": 180
    },
    {
      "epoch": 0.4342857142857143,
      "grad_norm": 2.6342821858302345,
      "learning_rate": 3.48203751140067e-07,
      "logits/chosen": -2.5662384033203125,
      "logits/rejected": -2.477693796157837,
      "logps/chosen": -274.94903564453125,
      "logps/rejected": -275.15032958984375,
      "loss": 0.0267,
      "rewards/accuracies": 0.800000011920929,
      "rewards/chosen": -0.14113394916057587,
      "rewards/margins": 0.26104897260665894,
      "rewards/rejected": -0.4021829068660736,
      "step": 190
    },
    {
      "epoch": 0.45714285714285713,
      "grad_norm": 2.36382999827445,
      "learning_rate": 3.2953144712759537e-07,
      "logits/chosen": -2.5955493450164795,
      "logits/rejected": -2.541722059249878,
      "logps/chosen": -258.66326904296875,
      "logps/rejected": -270.30609130859375,
      "loss": 0.0257,
      "rewards/accuracies": 0.7875000238418579,
      "rewards/chosen": -0.08326815813779831,
      "rewards/margins": 0.23202641308307648,
      "rewards/rejected": -0.3152945637702942,
      "step": 200
    },
    {
      "epoch": 0.45714285714285713,
      "eval_logits/chosen": -2.505389451980591,
      "eval_logits/rejected": -2.466869831085205,
      "eval_logps/chosen": -272.464599609375,
      "eval_logps/rejected": -282.8442077636719,
      "eval_loss": 0.025467805564403534,
      "eval_rewards/accuracies": 0.7370689511299133,
      "eval_rewards/chosen": -0.12503290176391602,
      "eval_rewards/margins": 0.22063730657100677,
      "eval_rewards/rejected": -0.3456701934337616,
      "eval_runtime": 92.5527,
      "eval_samples_per_second": 19.783,
      "eval_steps_per_second": 0.313,
      "step": 200
    },
    {
      "epoch": 0.48,
      "grad_norm": 2.537677375344279,
      "learning_rate": 3.103511916141658e-07,
      "logits/chosen": -2.515646457672119,
      "logits/rejected": -2.4341044425964355,
      "logps/chosen": -296.09417724609375,
      "logps/rejected": -296.98992919921875,
      "loss": 0.0255,
      "rewards/accuracies": 0.762499988079071,
      "rewards/chosen": -0.10250511020421982,
      "rewards/margins": 0.24679327011108398,
      "rewards/rejected": -0.3492983877658844,
      "step": 210
    },
    {
      "epoch": 0.5028571428571429,
      "grad_norm": 3.017540832091563,
      "learning_rate": 2.9078548506882117e-07,
      "logits/chosen": -2.455840826034546,
      "logits/rejected": -2.4036200046539307,
      "logps/chosen": -314.2608642578125,
      "logps/rejected": -324.8536071777344,
      "loss": 0.0268,
      "rewards/accuracies": 0.7437499761581421,
      "rewards/chosen": -0.09878429025411606,
      "rewards/margins": 0.24836881458759308,
      "rewards/rejected": -0.3471531271934509,
      "step": 220
    },
    {
      "epoch": 0.5257142857142857,
      "grad_norm": 2.5806806439798318,
      "learning_rate": 2.709592897595191e-07,
      "logits/chosen": -2.5932087898254395,
      "logits/rejected": -2.5489110946655273,
      "logps/chosen": -291.34063720703125,
      "logps/rejected": -303.5892028808594,
      "loss": 0.027,
      "rewards/accuracies": 0.7124999761581421,
      "rewards/chosen": -0.06546103954315186,
      "rewards/margins": 0.18855178356170654,
      "rewards/rejected": -0.2540128231048584,
      "step": 230
    },
    {
      "epoch": 0.5485714285714286,
      "grad_norm": 2.6133927348580595,
      "learning_rate": 2.509992316440332e-07,
      "logits/chosen": -2.605285167694092,
      "logits/rejected": -2.5287115573883057,
      "logps/chosen": -316.57171630859375,
      "logps/rejected": -295.0510559082031,
      "loss": 0.0264,
      "rewards/accuracies": 0.7875000238418579,
      "rewards/chosen": -0.08883465826511383,
      "rewards/margins": 0.24537017941474915,
      "rewards/rejected": -0.3342048227787018,
      "step": 240
    },
    {
      "epoch": 0.5714285714285714,
      "grad_norm": 2.534725769830134,
      "learning_rate": 2.3103279163519918e-07,
      "logits/chosen": -2.5661091804504395,
      "logits/rejected": -2.5358757972717285,
      "logps/chosen": -294.138427734375,
      "logps/rejected": -314.68145751953125,
      "loss": 0.0259,
      "rewards/accuracies": 0.675000011920929,
      "rewards/chosen": -0.16875380277633667,
      "rewards/margins": 0.16810715198516846,
      "rewards/rejected": -0.3368609547615051,
      "step": 250
    },
    {
      "epoch": 0.5714285714285714,
      "eval_logits/chosen": -2.609321117401123,
      "eval_logits/rejected": -2.5650839805603027,
      "eval_logps/chosen": -274.17425537109375,
      "eval_logps/rejected": -285.8867492675781,
      "eval_loss": 0.02494961954653263,
      "eval_rewards/accuracies": 0.7456896305084229,
      "eval_rewards/chosen": -0.14212976396083832,
      "eval_rewards/margins": 0.23396606743335724,
      "eval_rewards/rejected": -0.37609583139419556,
      "eval_runtime": 92.1346,
      "eval_samples_per_second": 19.873,
      "eval_steps_per_second": 0.315,
      "step": 250
    },
    {
      "epoch": 0.5942857142857143,
      "grad_norm": 2.414707825956405,
      "learning_rate": 2.1118749140573358e-07,
      "logits/chosen": -2.580400228500366,
      "logits/rejected": -2.5031521320343018,
      "logps/chosen": -283.9717712402344,
      "logps/rejected": -252.01821899414062,
      "loss": 0.0261,
      "rewards/accuracies": 0.7124999761581421,
      "rewards/chosen": -0.0814717561006546,
      "rewards/margins": 0.21402212977409363,
      "rewards/rejected": -0.2954939305782318,
      "step": 260
    },
    {
      "epoch": 0.6171428571428571,
      "grad_norm": 2.5939402682824984,
      "learning_rate": 1.9159007893272703e-07,
      "logits/chosen": -2.6335721015930176,
      "logits/rejected": -2.547212600708008,
      "logps/chosen": -297.3409423828125,
      "logps/rejected": -272.5791931152344,
      "loss": 0.026,
      "rewards/accuracies": 0.637499988079071,
      "rewards/chosen": -0.09214051067829132,
      "rewards/margins": 0.1732146292924881,
      "rewards/rejected": -0.26535511016845703,
      "step": 270
    },
    {
      "epoch": 0.64,
      "grad_norm": 2.2469749248074895,
      "learning_rate": 1.7236571898357766e-07,
      "logits/chosen": -2.4656567573547363,
      "logits/rejected": -2.41697359085083,
      "logps/chosen": -263.90411376953125,
      "logps/rejected": -272.6389465332031,
      "loss": 0.0259,
      "rewards/accuracies": 0.7562500238418579,
      "rewards/chosen": -0.1417647898197174,
      "rewards/margins": 0.24923905730247498,
      "rewards/rejected": -0.3910038471221924,
      "step": 280
    },
    {
      "epoch": 0.6628571428571428,
      "grad_norm": 2.6187338744329365,
      "learning_rate": 1.5363719371356882e-07,
      "logits/chosen": -2.569354295730591,
      "logits/rejected": -2.459669589996338,
      "logps/chosen": -297.64971923828125,
      "logps/rejected": -284.9115295410156,
      "loss": 0.0263,
      "rewards/accuracies": 0.7124999761581421,
      "rewards/chosen": -0.11722731590270996,
      "rewards/margins": 0.20035290718078613,
      "rewards/rejected": -0.3175801634788513,
      "step": 290
    },
    {
      "epoch": 0.6857142857142857,
      "grad_norm": 2.6936960874326923,
      "learning_rate": 1.3552411848071565e-07,
      "logits/chosen": -2.6174628734588623,
      "logits/rejected": -2.498176097869873,
      "logps/chosen": -322.8263854980469,
      "logps/rejected": -298.7721252441406,
      "loss": 0.0244,
      "rewards/accuracies": 0.762499988079071,
      "rewards/chosen": -0.09476250410079956,
      "rewards/margins": 0.24882373213768005,
      "rewards/rejected": -0.343586266040802,
      "step": 300
    },
    {
      "epoch": 0.6857142857142857,
      "eval_logits/chosen": -2.5555710792541504,
      "eval_logits/rejected": -2.4930553436279297,
      "eval_logps/chosen": -271.10662841796875,
      "eval_logps/rejected": -284.2852478027344,
      "eval_loss": 0.0244239903986454,
      "eval_rewards/accuracies": 0.732758641242981,
      "eval_rewards/chosen": -0.11145297437906265,
      "eval_rewards/margins": 0.24862772226333618,
      "eval_rewards/rejected": -0.3600807189941406,
      "eval_runtime": 91.0911,
      "eval_samples_per_second": 20.101,
      "eval_steps_per_second": 0.318,
      "step": 300
    },
    {
      "epoch": 0.7085714285714285,
      "grad_norm": 2.3645242243925644,
      "learning_rate": 1.1814217788631473e-07,
      "logits/chosen": -2.63641095161438,
      "logits/rejected": -2.527548313140869,
      "logps/chosen": -320.08880615234375,
      "logps/rejected": -300.36761474609375,
      "loss": 0.0242,
      "rewards/accuracies": 0.7875000238418579,
      "rewards/chosen": -0.0683048814535141,
      "rewards/margins": 0.2592974901199341,
      "rewards/rejected": -0.3276023864746094,
      "step": 310
    },
    {
      "epoch": 0.7314285714285714,
      "grad_norm": 2.7644839761717734,
      "learning_rate": 1.0160238692045331e-07,
      "logits/chosen": -2.511667251586914,
      "logits/rejected": -2.400014638900757,
      "logps/chosen": -307.7832336425781,
      "logps/rejected": -310.0841064453125,
      "loss": 0.0244,
      "rewards/accuracies": 0.793749988079071,
      "rewards/chosen": -0.10665328800678253,
      "rewards/margins": 0.24441497027873993,
      "rewards/rejected": -0.35106825828552246,
      "step": 320
    },
    {
      "epoch": 0.7542857142857143,
      "grad_norm": 2.781520358885941,
      "learning_rate": 8.601038193139438e-08,
      "logits/chosen": -2.4999566078186035,
      "logits/rejected": -2.454909563064575,
      "logps/chosen": -257.72882080078125,
      "logps/rejected": -273.76507568359375,
      "loss": 0.0274,
      "rewards/accuracies": 0.706250011920929,
      "rewards/chosen": -0.12040361016988754,
      "rewards/margins": 0.21321117877960205,
      "rewards/rejected": -0.3336147665977478,
      "step": 330
    },
    {
      "epoch": 0.7771428571428571,
      "grad_norm": 2.7244436343518386,
      "learning_rate": 7.146574594727572e-08,
      "logits/chosen": -2.5485446453094482,
      "logits/rejected": -2.4302964210510254,
      "logps/chosen": -257.6943664550781,
      "logps/rejected": -247.9900665283203,
      "loss": 0.0256,
      "rewards/accuracies": 0.7875000238418579,
      "rewards/chosen": -0.1169048398733139,
      "rewards/margins": 0.22713470458984375,
      "rewards/rejected": -0.34403952956199646,
      "step": 340
    },
    {
      "epoch": 0.8,
      "grad_norm": 2.9548985236957557,
      "learning_rate": 5.8061372659157306e-08,
      "logits/chosen": -2.5826621055603027,
      "logits/rejected": -2.4442431926727295,
      "logps/chosen": -319.64111328125,
      "logps/rejected": -302.02886962890625,
      "loss": 0.0232,
      "rewards/accuracies": 0.762499988079071,
      "rewards/chosen": -0.0946800708770752,
      "rewards/margins": 0.24179792404174805,
      "rewards/rejected": -0.33647799491882324,
      "step": 350
    },
    {
      "epoch": 0.8,
      "eval_logits/chosen": -2.5542304515838623,
      "eval_logits/rejected": -2.493983268737793,
      "eval_logps/chosen": -271.9947204589844,
      "eval_logps/rejected": -285.1673889160156,
      "eval_loss": 0.024052616208791733,
      "eval_rewards/accuracies": 0.732758641242981,
      "eval_rewards/chosen": -0.12033439427614212,
      "eval_rewards/margins": 0.24856774508953094,
      "eval_rewards/rejected": -0.36890214681625366,
      "eval_runtime": 92.5532,
      "eval_samples_per_second": 19.783,
      "eval_steps_per_second": 0.313,
      "step": 350
    },
    {
      "epoch": 0.8228571428571428,
      "grad_norm": 2.6937215781427297,
      "learning_rate": 4.5882873127531614e-08,
      "logits/chosen": -2.5525920391082764,
      "logits/rejected": -2.4930379390716553,
      "logps/chosen": -303.6790771484375,
      "logps/rejected": -306.0064392089844,
      "loss": 0.0264,
      "rewards/accuracies": 0.75,
      "rewards/chosen": -0.11168299615383148,
      "rewards/margins": 0.2747558355331421,
      "rewards/rejected": -0.38643878698349,
      "step": 360
    },
    {
      "epoch": 0.8457142857142858,
      "grad_norm": 3.0478094770904853,
      "learning_rate": 3.500802900154412e-08,
      "logits/chosen": -2.560382604598999,
      "logits/rejected": -2.522139549255371,
      "logps/chosen": -266.638671875,
      "logps/rejected": -294.6788330078125,
      "loss": 0.0246,
      "rewards/accuracies": 0.75,
      "rewards/chosen": -0.10632523149251938,
      "rewards/margins": 0.21934013068675995,
      "rewards/rejected": -0.3256654143333435,
      "step": 370
    },
    {
      "epoch": 0.8685714285714285,
      "grad_norm": 2.925024120524796,
      "learning_rate": 2.550629574310309e-08,
      "logits/chosen": -2.484462261199951,
      "logits/rejected": -2.4379611015319824,
      "logps/chosen": -269.59234619140625,
      "logps/rejected": -288.28961181640625,
      "loss": 0.0253,
      "rewards/accuracies": 0.6937500238418579,
      "rewards/chosen": -0.13730758428573608,
      "rewards/margins": 0.16804398596286774,
      "rewards/rejected": -0.305351585149765,
      "step": 380
    },
    {
      "epoch": 0.8914285714285715,
      "grad_norm": 2.515719678530588,
      "learning_rate": 1.7438359028687983e-08,
      "logits/chosen": -2.5585129261016846,
      "logits/rejected": -2.4909610748291016,
      "logps/chosen": -277.55438232421875,
      "logps/rejected": -285.38214111328125,
      "loss": 0.0251,
      "rewards/accuracies": 0.668749988079071,
      "rewards/chosen": -0.10983601957559586,
      "rewards/margins": 0.23024682700634003,
      "rewards/rejected": -0.3400828242301941,
      "step": 390
    },
    {
      "epoch": 0.9142857142857143,
      "grad_norm": 2.819162826588092,
      "learning_rate": 1.0855747162029361e-08,
      "logits/chosen": -2.538931131362915,
      "logits/rejected": -2.500331401824951,
      "logps/chosen": -285.94342041015625,
      "logps/rejected": -301.77130126953125,
      "loss": 0.0253,
      "rewards/accuracies": 0.737500011920929,
      "rewards/chosen": -0.11810322850942612,
      "rewards/margins": 0.22658781707286835,
      "rewards/rejected": -0.34469106793403625,
      "step": 400
    },
    {
      "epoch": 0.9142857142857143,
      "eval_logits/chosen": -2.5477468967437744,
      "eval_logits/rejected": -2.485684633255005,
      "eval_logps/chosen": -273.0142517089844,
      "eval_logps/rejected": -286.3204040527344,
      "eval_loss": 0.024040229618549347,
      "eval_rewards/accuracies": 0.732758641242981,
      "eval_rewards/chosen": -0.13052918016910553,
      "eval_rewards/margins": 0.2499030977487564,
      "eval_rewards/rejected": -0.38043221831321716,
      "eval_runtime": 91.0229,
      "eval_samples_per_second": 20.116,
      "eval_steps_per_second": 0.319,
      "step": 400
    },
    {
      "epoch": 0.9371428571428572,
      "grad_norm": 3.0051422552626748,
      "learning_rate": 5.8005019731033615e-09,
      "logits/chosen": -2.547819137573242,
      "logits/rejected": -2.4846596717834473,
      "logps/chosen": -271.83538818359375,
      "logps/rejected": -289.2705078125,
      "loss": 0.0246,
      "rewards/accuracies": 0.7749999761581421,
      "rewards/chosen": -0.11531362682580948,
      "rewards/margins": 0.20074260234832764,
      "rewards/rejected": -0.3160562515258789,
      "step": 410
    },
    {
      "epoch": 0.96,
      "grad_norm": 2.650270081198499,
      "learning_rate": 2.3049103053431886e-09,
      "logits/chosen": -2.6320762634277344,
      "logits/rejected": -2.5170094966888428,
      "logps/chosen": -321.63592529296875,
      "logps/rejected": -333.3614501953125,
      "loss": 0.0246,
      "rewards/accuracies": 0.768750011920929,
      "rewards/chosen": -0.1096896082162857,
      "rewards/margins": 0.2813798785209656,
      "rewards/rejected": -0.3910694718360901,
      "step": 420
    },
    {
      "epoch": 0.9828571428571429,
      "grad_norm": 2.8479881145572183,
      "learning_rate": 3.9129780600541397e-10,
      "logits/chosen": -2.5273232460021973,
      "logits/rejected": -2.4938735961914062,
      "logps/chosen": -295.13446044921875,
      "logps/rejected": -292.31854248046875,
      "loss": 0.025,
      "rewards/accuracies": 0.768750011920929,
      "rewards/chosen": -0.09540516138076782,
      "rewards/margins": 0.24863004684448242,
      "rewards/rejected": -0.34403520822525024,
      "step": 430
    },
    {
      "epoch": 0.9988571428571429,
      "step": 437,
      "total_flos": 0.0,
      "train_loss": 0.02869435827710263,
      "train_runtime": 11449.0759,
      "train_samples_per_second": 4.891,
      "train_steps_per_second": 0.038
    }
  ],
  "logging_steps": 10,
  "max_steps": 437,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 1,
  "save_steps": 100,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 0.0,
  "train_batch_size": 8,
  "trial_name": null,
  "trial_params": null
}