diff --git "a/trainer_state.json" "b/trainer_state.json"
new file mode 100644--- /dev/null
+++ "b/trainer_state.json"
@@ -0,0 +1,4086 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 1.0,
+  "eval_steps": 100,
+  "global_step": 1563,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.0,
+      "grad_norm": 42.15456008911133,
+      "kl": 0.017818570137023926,
+      "learning_rate": 1.592356687898089e-08,
+      "logps/chosen": -285.75128173828125,
+      "logps/rejected": -254.7062530517578,
+      "loss": 0.4999,
+      "rewards/chosen": 0.004669209010899067,
+      "rewards/margins": 0.0025329389609396458,
+      "rewards/rejected": 0.0021362705156207085,
+      "step": 5
+    },
+    {
+      "epoch": 0.01,
+      "grad_norm": 44.65619659423828,
+      "kl": 0.05461766570806503,
+      "learning_rate": 3.184713375796178e-08,
+      "logps/chosen": -286.11944580078125,
+      "logps/rejected": -276.1832580566406,
+      "loss": 0.496,
+      "rewards/chosen": 0.03117586299777031,
+      "rewards/margins": 0.033454541116952896,
+      "rewards/rejected": -0.00227867579087615,
+      "step": 10
+    },
+    {
+      "epoch": 0.01,
+      "grad_norm": 39.61723327636719,
+      "kl": 0.44255906343460083,
+      "learning_rate": 4.777070063694268e-08,
+      "logps/chosen": -269.203125,
+      "logps/rejected": -260.71966552734375,
+      "loss": 0.4849,
+      "rewards/chosen": 0.13460782170295715,
+      "rewards/margins": 0.11511580646038055,
+      "rewards/rejected": 0.0194920115172863,
+      "step": 15
+    },
+    {
+      "epoch": 0.01,
+      "grad_norm": 36.42387771606445,
+      "kl": 0.871076226234436,
+      "learning_rate": 6.369426751592356e-08,
+      "logps/chosen": -244.55203247070312,
+      "logps/rejected": -259.15496826171875,
+      "loss": 0.4573,
+      "rewards/chosen": 0.3324822187423706,
+      "rewards/margins": 0.3418889045715332,
+      "rewards/rejected": -0.00940666627138853,
+      "step": 20
+    },
+    {
+      "epoch": 0.02,
+      "grad_norm": 35.794044494628906,
+      "kl": 0.3753414750099182,
+      "learning_rate": 7.961783439490445e-08,
+      "logps/chosen": -261.01800537109375,
+      "logps/rejected": -271.74786376953125,
+      "loss": 0.4164,
+      "rewards/chosen": 0.48356980085372925,
+      "rewards/margins": 0.7098164558410645,
+      "rewards/rejected": -0.22624659538269043,
+      "step": 25
+    },
+    {
+      "epoch": 0.02,
+      "grad_norm": 34.50333786010742,
+      "kl": 0.0,
+      "learning_rate": 9.554140127388536e-08,
+      "logps/chosen": -306.0127868652344,
+      "logps/rejected": -244.1356658935547,
+      "loss": 0.3649,
+      "rewards/chosen": 0.746610164642334,
+      "rewards/margins": 1.2027219533920288,
+      "rewards/rejected": -0.4561118483543396,
+      "step": 30
+    },
+    {
+      "epoch": 0.02,
+      "grad_norm": 30.50051498413086,
+      "kl": 0.0,
+      "learning_rate": 1.1146496815286624e-07,
+      "logps/chosen": -280.65692138671875,
+      "logps/rejected": -262.2502136230469,
+      "loss": 0.3171,
+      "rewards/chosen": 0.774597704410553,
+      "rewards/margins": 1.728833794593811,
+      "rewards/rejected": -0.9542360305786133,
+      "step": 35
+    },
+    {
+      "epoch": 0.03,
+      "grad_norm": 29.806577682495117,
+      "kl": 0.0,
+      "learning_rate": 1.2738853503184713e-07,
+      "logps/chosen": -276.9497375488281,
+      "logps/rejected": -259.6419982910156,
+      "loss": 0.2884,
+      "rewards/chosen": 0.6780191659927368,
+      "rewards/margins": 2.165379047393799,
+      "rewards/rejected": -1.4873597621917725,
+      "step": 40
+    },
+    {
+      "epoch": 0.03,
+      "grad_norm": 22.508939743041992,
+      "kl": 0.0,
+      "learning_rate": 1.43312101910828e-07,
+      "logps/chosen": -253.7576141357422,
+      "logps/rejected": -268.8052062988281,
+      "loss": 0.2596,
+      "rewards/chosen": 1.205094575881958,
+      "rewards/margins": 2.685385227203369,
+      "rewards/rejected": -1.4802907705307007,
+      "step": 45
+    },
+    {
+      "epoch": 0.03,
+      "grad_norm": 29.29058074951172,
+      "kl": 0.0,
+      "learning_rate": 1.592356687898089e-07,
+      "logps/chosen": -284.4584045410156,
+      "logps/rejected": -254.226806640625,
+      "loss": 0.2614,
+      "rewards/chosen": 1.0118796825408936,
+      "rewards/margins": 2.8269615173339844,
+      "rewards/rejected": -1.8150818347930908,
+      "step": 50
+    },
+    {
+      "epoch": 0.04,
+      "grad_norm": 29.25771141052246,
+      "kl": 0.0,
+      "learning_rate": 1.7515923566878978e-07,
+      "logps/chosen": -266.92254638671875,
+      "logps/rejected": -260.49261474609375,
+      "loss": 0.233,
+      "rewards/chosen": 1.16305673122406,
+      "rewards/margins": 3.4210219383239746,
+      "rewards/rejected": -2.257965326309204,
+      "step": 55
+    },
+    {
+      "epoch": 0.04,
+      "grad_norm": 20.642614364624023,
+      "kl": 0.0,
+      "learning_rate": 1.9108280254777072e-07,
+      "logps/chosen": -231.07168579101562,
+      "logps/rejected": -258.5835876464844,
+      "loss": 0.241,
+      "rewards/chosen": 1.3277983665466309,
+      "rewards/margins": 3.532343626022339,
+      "rewards/rejected": -2.204545497894287,
+      "step": 60
+    },
+    {
+      "epoch": 0.04,
+      "grad_norm": 23.786853790283203,
+      "kl": 0.0,
+      "learning_rate": 2.070063694267516e-07,
+      "logps/chosen": -251.5561981201172,
+      "logps/rejected": -267.85986328125,
+      "loss": 0.2376,
+      "rewards/chosen": 1.2313438653945923,
+      "rewards/margins": 3.837125062942505,
+      "rewards/rejected": -2.605781316757202,
+      "step": 65
+    },
+    {
+      "epoch": 0.04,
+      "grad_norm": 24.48412322998047,
+      "kl": 0.0,
+      "learning_rate": 2.2292993630573247e-07,
+      "logps/chosen": -244.3961639404297,
+      "logps/rejected": -248.7190704345703,
+      "loss": 0.2304,
+      "rewards/chosen": 1.338683843612671,
+      "rewards/margins": 3.6262855529785156,
+      "rewards/rejected": -2.287601947784424,
+      "step": 70
+    },
+    {
+      "epoch": 0.05,
+      "grad_norm": 16.654287338256836,
+      "kl": 0.0,
+      "learning_rate": 2.388535031847134e-07,
+      "logps/chosen": -261.95819091796875,
+      "logps/rejected": -291.40020751953125,
+      "loss": 0.2003,
+      "rewards/chosen": 1.3882195949554443,
+      "rewards/margins": 4.626662254333496,
+      "rewards/rejected": -3.238443374633789,
+      "step": 75
+    },
+    {
+      "epoch": 0.05,
+      "grad_norm": 30.89764976501465,
+      "kl": 0.0,
+      "learning_rate": 2.5477707006369425e-07,
+      "logps/chosen": -267.4027099609375,
+      "logps/rejected": -271.3486328125,
+      "loss": 0.1985,
+      "rewards/chosen": 1.4253151416778564,
+      "rewards/margins": 4.816943645477295,
+      "rewards/rejected": -3.3916287422180176,
+      "step": 80
+    },
+    {
+      "epoch": 0.05,
+      "grad_norm": 55.57185363769531,
+      "kl": 0.0,
+      "learning_rate": 2.7070063694267513e-07,
+      "logps/chosen": -289.416259765625,
+      "logps/rejected": -282.303955078125,
+      "loss": 0.2075,
+      "rewards/chosen": 1.308215856552124,
+      "rewards/margins": 5.0450119972229,
+      "rewards/rejected": -3.7367959022521973,
+      "step": 85
+    },
+    {
+      "epoch": 0.06,
+      "grad_norm": 34.31930923461914,
+      "kl": 0.0,
+      "learning_rate": 2.86624203821656e-07,
+      "logps/chosen": -255.4053192138672,
+      "logps/rejected": -293.8216247558594,
+      "loss": 0.1877,
+      "rewards/chosen": 1.560572862625122,
+      "rewards/margins": 5.238338470458984,
+      "rewards/rejected": -3.6777656078338623,
+      "step": 90
+    },
+    {
+      "epoch": 0.06,
+      "grad_norm": 22.21489143371582,
+      "kl": 0.0,
+      "learning_rate": 3.0254777070063694e-07,
+      "logps/chosen": -286.2869873046875,
+      "logps/rejected": -271.2530212402344,
+      "loss": 0.1985,
+      "rewards/chosen": 1.478058099746704,
+      "rewards/margins": 5.051484107971191,
+      "rewards/rejected": -3.5734260082244873,
+      "step": 95
+    },
+    {
+      "epoch": 0.06,
+      "grad_norm": 29.9346981048584,
+      "kl": 0.0,
+      "learning_rate": 3.184713375796178e-07,
+      "logps/chosen": -245.080810546875,
+      "logps/rejected": -287.15435791015625,
+      "loss": 0.1656,
+      "rewards/chosen": 1.4543142318725586,
+      "rewards/margins": 5.874560356140137,
+      "rewards/rejected": -4.420246124267578,
+      "step": 100
+    },
+    {
+      "epoch": 0.07,
+      "grad_norm": 22.431930541992188,
+      "kl": 0.0,
+      "learning_rate": 3.343949044585987e-07,
+      "logps/chosen": -259.4337463378906,
+      "logps/rejected": -309.25164794921875,
+      "loss": 0.1767,
+      "rewards/chosen": 1.4657680988311768,
+      "rewards/margins": 5.859877109527588,
+      "rewards/rejected": -4.39410924911499,
+      "step": 105
+    },
+    {
+      "epoch": 0.07,
+      "grad_norm": 36.517799377441406,
+      "kl": 0.0,
+      "learning_rate": 3.5031847133757957e-07,
+      "logps/chosen": -264.08306884765625,
+      "logps/rejected": -295.64788818359375,
+      "loss": 0.1599,
+      "rewards/chosen": 1.291338562965393,
+      "rewards/margins": 5.8188042640686035,
+      "rewards/rejected": -4.5274658203125,
+      "step": 110
+    },
+    {
+      "epoch": 0.07,
+      "grad_norm": 25.18623924255371,
+      "kl": 0.0,
+      "learning_rate": 3.6624203821656045e-07,
+      "logps/chosen": -268.96527099609375,
+      "logps/rejected": -278.4827575683594,
+      "loss": 0.1735,
+      "rewards/chosen": 1.4270073175430298,
+      "rewards/margins": 5.516595363616943,
+      "rewards/rejected": -4.0895891189575195,
+      "step": 115
+    },
+    {
+      "epoch": 0.08,
+      "grad_norm": 24.899450302124023,
+      "kl": 0.0,
+      "learning_rate": 3.8216560509554143e-07,
+      "logps/chosen": -257.46807861328125,
+      "logps/rejected": -306.6875305175781,
+      "loss": 0.1675,
+      "rewards/chosen": 1.363029956817627,
+      "rewards/margins": 5.9437384605407715,
+      "rewards/rejected": -4.580708980560303,
+      "step": 120
+    },
+    {
+      "epoch": 0.08,
+      "grad_norm": 18.1636962890625,
+      "kl": 0.0,
+      "learning_rate": 3.980891719745223e-07,
+      "logps/chosen": -274.8599548339844,
+      "logps/rejected": -316.78057861328125,
+      "loss": 0.1509,
+      "rewards/chosen": 1.4044160842895508,
+      "rewards/margins": 6.668715000152588,
+      "rewards/rejected": -5.264299392700195,
+      "step": 125
+    },
+    {
+      "epoch": 0.08,
+      "grad_norm": 53.83689498901367,
+      "kl": 0.0,
+      "learning_rate": 4.140127388535032e-07,
+      "logps/chosen": -316.65460205078125,
+      "logps/rejected": -318.2517395019531,
+      "loss": 0.1667,
+      "rewards/chosen": 1.331627368927002,
+      "rewards/margins": 6.6092848777771,
+      "rewards/rejected": -5.277657508850098,
+      "step": 130
+    },
+    {
+      "epoch": 0.09,
+      "grad_norm": 16.44365692138672,
+      "kl": 0.0,
+      "learning_rate": 4.2993630573248406e-07,
+      "logps/chosen": -242.02651977539062,
+      "logps/rejected": -283.90106201171875,
+      "loss": 0.1762,
+      "rewards/chosen": 1.6127967834472656,
+      "rewards/margins": 6.091832160949707,
+      "rewards/rejected": -4.479035377502441,
+      "step": 135
+    },
+    {
+      "epoch": 0.09,
+      "grad_norm": 19.764270782470703,
+      "kl": 0.0,
+      "learning_rate": 4.4585987261146494e-07,
+      "logps/chosen": -252.82180786132812,
+      "logps/rejected": -319.73199462890625,
+      "loss": 0.1468,
+      "rewards/chosen": 1.5330384969711304,
+      "rewards/margins": 6.702307224273682,
+      "rewards/rejected": -5.169268608093262,
+      "step": 140
+    },
+    {
+      "epoch": 0.09,
+      "grad_norm": 21.560876846313477,
+      "kl": 0.0,
+      "learning_rate": 4.6178343949044587e-07,
+      "logps/chosen": -247.28427124023438,
+      "logps/rejected": -282.38848876953125,
+      "loss": 0.1868,
+      "rewards/chosen": 1.350294828414917,
+      "rewards/margins": 5.6418867111206055,
+      "rewards/rejected": -4.291592597961426,
+      "step": 145
+    },
+    {
+      "epoch": 0.1,
+      "grad_norm": 21.553871154785156,
+      "kl": 0.0,
+      "learning_rate": 4.777070063694267e-07,
+      "logps/chosen": -286.14312744140625,
+      "logps/rejected": -281.7044372558594,
+      "loss": 0.168,
+      "rewards/chosen": 1.5479779243469238,
+      "rewards/margins": 6.192603588104248,
+      "rewards/rejected": -4.644625663757324,
+      "step": 150
+    },
+    {
+      "epoch": 0.1,
+      "grad_norm": 24.488561630249023,
+      "kl": 0.0,
+      "learning_rate": 4.936305732484076e-07,
+      "logps/chosen": -272.4622497558594,
+      "logps/rejected": -323.91082763671875,
+      "loss": 0.1648,
+      "rewards/chosen": 1.3805065155029297,
+      "rewards/margins": 6.456129550933838,
+      "rewards/rejected": -5.07562255859375,
+      "step": 155
+    },
+    {
+      "epoch": 0.1,
+      "grad_norm": 21.614229202270508,
+      "kl": 0.0,
+      "learning_rate": 4.989331436699858e-07,
+      "logps/chosen": -207.8739776611328,
+      "logps/rejected": -324.73565673828125,
+      "loss": 0.1435,
+      "rewards/chosen": 1.465820074081421,
+      "rewards/margins": 6.953179836273193,
+      "rewards/rejected": -5.487359046936035,
+      "step": 160
+    },
+    {
+      "epoch": 0.11,
+      "grad_norm": 20.425729751586914,
+      "kl": 0.0,
+      "learning_rate": 4.971550497866287e-07,
+      "logps/chosen": -287.088623046875,
+      "logps/rejected": -310.1844787597656,
+      "loss": 0.1568,
+      "rewards/chosen": 1.5799640417099,
+      "rewards/margins": 6.89548397064209,
+      "rewards/rejected": -5.315520286560059,
+      "step": 165
+    },
+    {
+      "epoch": 0.11,
+      "grad_norm": 21.86062240600586,
+      "kl": 0.0,
+      "learning_rate": 4.953769559032717e-07,
+      "logps/chosen": -282.91741943359375,
+      "logps/rejected": -315.76470947265625,
+      "loss": 0.1516,
+      "rewards/chosen": 1.442077875137329,
+      "rewards/margins": 7.074441432952881,
+      "rewards/rejected": -5.632363796234131,
+      "step": 170
+    },
+    {
+      "epoch": 0.11,
+      "grad_norm": 20.31084632873535,
+      "kl": 0.0,
+      "learning_rate": 4.935988620199146e-07,
+      "logps/chosen": -274.0014953613281,
+      "logps/rejected": -310.8323059082031,
+      "loss": 0.1477,
+      "rewards/chosen": 1.4551314115524292,
+      "rewards/margins": 6.99865198135376,
+      "rewards/rejected": -5.543520450592041,
+      "step": 175
+    },
+    {
+      "epoch": 0.12,
+      "grad_norm": 20.067346572875977,
+      "kl": 0.0,
+      "learning_rate": 4.918207681365576e-07,
+      "logps/chosen": -274.91705322265625,
+      "logps/rejected": -330.29315185546875,
+      "loss": 0.1491,
+      "rewards/chosen": 1.5545661449432373,
+      "rewards/margins": 7.674098014831543,
+      "rewards/rejected": -6.119531631469727,
+      "step": 180
+    },
+    {
+      "epoch": 0.12,
+      "grad_norm": 26.703292846679688,
+      "kl": 0.0,
+      "learning_rate": 4.900426742532006e-07,
+      "logps/chosen": -267.2247619628906,
+      "logps/rejected": -308.6265563964844,
+      "loss": 0.142,
+      "rewards/chosen": 1.4495770931243896,
+      "rewards/margins": 7.4614386558532715,
+      "rewards/rejected": -6.011861324310303,
+      "step": 185
+    },
+    {
+      "epoch": 0.12,
+      "grad_norm": 22.81442642211914,
+      "kl": 0.0,
+      "learning_rate": 4.882645803698435e-07,
+      "logps/chosen": -243.8881072998047,
+      "logps/rejected": -318.4960632324219,
+      "loss": 0.1679,
+      "rewards/chosen": 1.2335823774337769,
+      "rewards/margins": 7.564291954040527,
+      "rewards/rejected": -6.330709934234619,
+      "step": 190
+    },
+    {
+      "epoch": 0.12,
+      "grad_norm": 17.875883102416992,
+      "kl": 0.0,
+      "learning_rate": 4.864864864864865e-07,
+      "logps/chosen": -276.0986328125,
+      "logps/rejected": -317.36163330078125,
+      "loss": 0.1442,
+      "rewards/chosen": 1.5666391849517822,
+      "rewards/margins": 7.757128715515137,
+      "rewards/rejected": -6.190489768981934,
+      "step": 195
+    },
+    {
+      "epoch": 0.13,
+      "grad_norm": 22.31620216369629,
+      "kl": 0.0,
+      "learning_rate": 4.847083926031294e-07,
+      "logps/chosen": -273.843994140625,
+      "logps/rejected": -329.43475341796875,
+      "loss": 0.154,
+      "rewards/chosen": 1.6012461185455322,
+      "rewards/margins": 7.519808292388916,
+      "rewards/rejected": -5.918562412261963,
+      "step": 200
+    },
+    {
+      "epoch": 0.13,
+      "grad_norm": 19.976451873779297,
+      "kl": 0.0,
+      "learning_rate": 4.829302987197724e-07,
+      "logps/chosen": -284.374755859375,
+      "logps/rejected": -326.24371337890625,
+      "loss": 0.1378,
+      "rewards/chosen": 1.5799314975738525,
+      "rewards/margins": 7.78595495223999,
+      "rewards/rejected": -6.206023216247559,
+      "step": 205
+    },
+    {
+      "epoch": 0.13,
+      "grad_norm": 23.919170379638672,
+      "kl": 0.0,
+      "learning_rate": 4.811522048364154e-07,
+      "logps/chosen": -236.8705291748047,
+      "logps/rejected": -316.30413818359375,
+      "loss": 0.1523,
+      "rewards/chosen": 1.5169861316680908,
+      "rewards/margins": 7.513286590576172,
+      "rewards/rejected": -5.99630069732666,
+      "step": 210
+    },
+    {
+      "epoch": 0.14,
+      "grad_norm": 25.041824340820312,
+      "kl": 0.0,
+      "learning_rate": 4.793741109530583e-07,
+      "logps/chosen": -219.86117553710938,
+      "logps/rejected": -331.60418701171875,
+      "loss": 0.1516,
+      "rewards/chosen": 1.5909864902496338,
+      "rewards/margins": 7.9639410972595215,
+      "rewards/rejected": -6.372954845428467,
+      "step": 215
+    },
+    {
+      "epoch": 0.14,
+      "grad_norm": 25.45568084716797,
+      "kl": 0.0,
+      "learning_rate": 4.775960170697012e-07,
+      "logps/chosen": -271.017333984375,
+      "logps/rejected": -333.2148132324219,
+      "loss": 0.1628,
+      "rewards/chosen": 1.5948388576507568,
+      "rewards/margins": 7.8658246994018555,
+      "rewards/rejected": -6.2709856033325195,
+      "step": 220
+    },
+    {
+      "epoch": 0.14,
+      "grad_norm": 22.259382247924805,
+      "kl": 0.0,
+      "learning_rate": 4.7581792318634425e-07,
+      "logps/chosen": -253.55789184570312,
+      "logps/rejected": -298.8450622558594,
+      "loss": 0.1452,
+      "rewards/chosen": 1.640355110168457,
+      "rewards/margins": 8.043792724609375,
+      "rewards/rejected": -6.403438568115234,
+      "step": 225
+    },
+    {
+      "epoch": 0.15,
+      "grad_norm": 26.135997772216797,
+      "kl": 0.0,
+      "learning_rate": 4.7403982930298717e-07,
+      "logps/chosen": -235.11349487304688,
+      "logps/rejected": -344.72369384765625,
+      "loss": 0.1472,
+      "rewards/chosen": 1.559660792350769,
+      "rewards/margins": 7.8217010498046875,
+      "rewards/rejected": -6.262040615081787,
+      "step": 230
+    },
+    {
+      "epoch": 0.15,
+      "grad_norm": 20.143089294433594,
+      "kl": 0.0,
+      "learning_rate": 4.7226173541963014e-07,
+      "logps/chosen": -249.42196655273438,
+      "logps/rejected": -277.12554931640625,
+      "loss": 0.1529,
+      "rewards/chosen": 1.6460959911346436,
+      "rewards/margins": 7.348568916320801,
+      "rewards/rejected": -5.702473163604736,
+      "step": 235
+    },
+    {
+      "epoch": 0.15,
+      "grad_norm": 23.44059181213379,
+      "kl": 0.0,
+      "learning_rate": 4.7048364153627306e-07,
+      "logps/chosen": -281.3484802246094,
+      "logps/rejected": -307.25457763671875,
+      "loss": 0.1387,
+      "rewards/chosen": 1.7489140033721924,
+      "rewards/margins": 7.860198974609375,
+      "rewards/rejected": -6.111284255981445,
+      "step": 240
+    },
+    {
+      "epoch": 0.16,
+      "grad_norm": 23.201915740966797,
+      "kl": 0.0,
+      "learning_rate": 4.6870554765291604e-07,
+      "logps/chosen": -272.64031982421875,
+      "logps/rejected": -310.50909423828125,
+      "loss": 0.1356,
+      "rewards/chosen": 1.6725397109985352,
+      "rewards/margins": 8.012718200683594,
+      "rewards/rejected": -6.340178489685059,
+      "step": 245
+    },
+    {
+      "epoch": 0.16,
+      "grad_norm": 19.45538902282715,
+      "kl": 0.0,
+      "learning_rate": 4.66927453769559e-07,
+      "logps/chosen": -272.3523254394531,
+      "logps/rejected": -338.20867919921875,
+      "loss": 0.1246,
+      "rewards/chosen": 1.727900743484497,
+      "rewards/margins": 8.646397590637207,
+      "rewards/rejected": -6.918497562408447,
+      "step": 250
+    },
+    {
+      "epoch": 0.16,
+      "grad_norm": 24.896251678466797,
+      "kl": 0.0,
+      "learning_rate": 4.65149359886202e-07,
+      "logps/chosen": -269.20794677734375,
+      "logps/rejected": -328.73486328125,
+      "loss": 0.1281,
+      "rewards/chosen": 1.6262487173080444,
+      "rewards/margins": 9.223333358764648,
+      "rewards/rejected": -7.597084999084473,
+      "step": 255
+    },
+    {
+      "epoch": 0.17,
+      "grad_norm": 20.800025939941406,
+      "kl": 0.0,
+      "learning_rate": 4.633712660028449e-07,
+      "logps/chosen": -270.3649597167969,
+      "logps/rejected": -330.872314453125,
+      "loss": 0.1279,
+      "rewards/chosen": 1.5803369283676147,
+      "rewards/margins": 9.708114624023438,
+      "rewards/rejected": -8.127778053283691,
+      "step": 260
+    },
+    {
+      "epoch": 0.17,
+      "grad_norm": 20.315540313720703,
+      "kl": 0.0,
+      "learning_rate": 4.615931721194879e-07,
+      "logps/chosen": -277.21807861328125,
+      "logps/rejected": -313.4653625488281,
+      "loss": 0.1461,
+      "rewards/chosen": 1.6411361694335938,
+      "rewards/margins": 7.905195713043213,
+      "rewards/rejected": -6.264059543609619,
+      "step": 265
+    },
+    {
+      "epoch": 0.17,
+      "grad_norm": 19.68859100341797,
+      "kl": 0.0,
+      "learning_rate": 4.5981507823613085e-07,
+      "logps/chosen": -275.69696044921875,
+      "logps/rejected": -295.79400634765625,
+      "loss": 0.1547,
+      "rewards/chosen": 1.6596572399139404,
+      "rewards/margins": 7.9982404708862305,
+      "rewards/rejected": -6.338583469390869,
+      "step": 270
+    },
+    {
+      "epoch": 0.18,
+      "grad_norm": 27.070371627807617,
+      "kl": 0.0,
+      "learning_rate": 4.580369843527738e-07,
+      "logps/chosen": -272.72735595703125,
+      "logps/rejected": -310.302734375,
+      "loss": 0.1673,
+      "rewards/chosen": 1.2246049642562866,
+      "rewards/margins": 8.44649600982666,
+      "rewards/rejected": -7.221890449523926,
+      "step": 275
+    },
+    {
+      "epoch": 0.18,
+      "grad_norm": 15.166617393493652,
+      "kl": 0.0,
+      "learning_rate": 4.562588904694168e-07,
+      "logps/chosen": -267.9317932128906,
+      "logps/rejected": -316.60479736328125,
+      "loss": 0.1304,
+      "rewards/chosen": 1.7944562435150146,
+      "rewards/margins": 8.752721786499023,
+      "rewards/rejected": -6.958265781402588,
+      "step": 280
+    },
+    {
+      "epoch": 0.18,
+      "grad_norm": 27.143291473388672,
+      "kl": 0.0,
+      "learning_rate": 4.544807965860597e-07,
+      "logps/chosen": -257.7388916015625,
+      "logps/rejected": -308.5174560546875,
+      "loss": 0.1421,
+      "rewards/chosen": 1.6366589069366455,
+      "rewards/margins": 8.678224563598633,
+      "rewards/rejected": -7.04156494140625,
+      "step": 285
+    },
+    {
+      "epoch": 0.19,
+      "grad_norm": 22.1202392578125,
+      "kl": 0.0,
+      "learning_rate": 4.5270270270270264e-07,
+      "logps/chosen": -266.8901062011719,
+      "logps/rejected": -322.6683654785156,
+      "loss": 0.1447,
+      "rewards/chosen": 1.787255883216858,
+      "rewards/margins": 9.206243515014648,
+      "rewards/rejected": -7.418987274169922,
+      "step": 290
+    },
+    {
+      "epoch": 0.19,
+      "grad_norm": 24.7782039642334,
+      "kl": 0.0,
+      "learning_rate": 4.509246088193456e-07,
+      "logps/chosen": -274.30438232421875,
+      "logps/rejected": -342.98541259765625,
+      "loss": 0.1421,
+      "rewards/chosen": 1.776602029800415,
+      "rewards/margins": 9.053030967712402,
+      "rewards/rejected": -7.276429176330566,
+      "step": 295
+    },
+    {
+      "epoch": 0.19,
+      "grad_norm": 23.54313087463379,
+      "kl": 0.0,
+      "learning_rate": 4.491465149359886e-07,
+      "logps/chosen": -215.9270477294922,
+      "logps/rejected": -303.5273742675781,
+      "loss": 0.1438,
+      "rewards/chosen": 1.868506669998169,
+      "rewards/margins": 9.129117965698242,
+      "rewards/rejected": -7.260611534118652,
+      "step": 300
+    },
+    {
+      "epoch": 0.2,
+      "grad_norm": 17.70969009399414,
+      "kl": 0.0,
+      "learning_rate": 4.4736842105263156e-07,
+      "logps/chosen": -240.4123992919922,
+      "logps/rejected": -336.1444396972656,
+      "loss": 0.1403,
+      "rewards/chosen": 1.7039588689804077,
+      "rewards/margins": 9.00100326538086,
+      "rewards/rejected": -7.297044277191162,
+      "step": 305
+    },
+    {
+      "epoch": 0.2,
+      "grad_norm": 19.908315658569336,
+      "kl": 0.0,
+      "learning_rate": 4.4559032716927454e-07,
+      "logps/chosen": -212.1704864501953,
+      "logps/rejected": -340.3092041015625,
+      "loss": 0.1338,
+      "rewards/chosen": 1.6462271213531494,
+      "rewards/margins": 8.726274490356445,
+      "rewards/rejected": -7.0800461769104,
+      "step": 310
+    },
+    {
+      "epoch": 0.2,
+      "grad_norm": 22.928499221801758,
+      "kl": 0.0,
+      "learning_rate": 4.438122332859175e-07,
+      "logps/chosen": -243.7815704345703,
+      "logps/rejected": -324.60772705078125,
+      "loss": 0.1323,
+      "rewards/chosen": 1.6998401880264282,
+      "rewards/margins": 8.809240341186523,
+      "rewards/rejected": -7.109400749206543,
+      "step": 315
+    },
+    {
+      "epoch": 0.2,
+      "grad_norm": 20.08190155029297,
+      "kl": 0.0,
+      "learning_rate": 4.420341394025605e-07,
+      "logps/chosen": -263.37725830078125,
+      "logps/rejected": -317.4496154785156,
+      "loss": 0.1443,
+      "rewards/chosen": 1.706053376197815,
+      "rewards/margins": 8.708757400512695,
+      "rewards/rejected": -7.002703666687012,
+      "step": 320
+    },
+    {
+      "epoch": 0.21,
+      "grad_norm": 25.728233337402344,
+      "kl": 0.0,
+      "learning_rate": 4.4025604551920335e-07,
+      "logps/chosen": -282.44989013671875,
+      "logps/rejected": -325.9697265625,
+      "loss": 0.1351,
+      "rewards/chosen": 1.6976518630981445,
+      "rewards/margins": 9.016552925109863,
+      "rewards/rejected": -7.318901062011719,
+      "step": 325
+    },
+    {
+      "epoch": 0.21,
+      "grad_norm": 22.311450958251953,
+      "kl": 0.0,
+      "learning_rate": 4.384779516358463e-07,
+      "logps/chosen": -260.39447021484375,
+      "logps/rejected": -309.620849609375,
+      "loss": 0.1311,
+      "rewards/chosen": 1.8077850341796875,
+      "rewards/margins": 9.406209945678711,
+      "rewards/rejected": -7.598425388336182,
+      "step": 330
+    },
+    {
+      "epoch": 0.21,
+      "grad_norm": 23.20633316040039,
+      "kl": 0.0,
+      "learning_rate": 4.366998577524893e-07,
+      "logps/chosen": -302.47509765625,
+      "logps/rejected": -318.767578125,
+      "loss": 0.128,
+      "rewards/chosen": 1.7027689218521118,
+      "rewards/margins": 10.0870943069458,
+      "rewards/rejected": -8.38432502746582,
+      "step": 335
+    },
+    {
+      "epoch": 0.22,
+      "grad_norm": 25.405933380126953,
+      "kl": 0.0,
+      "learning_rate": 4.3492176386913227e-07,
+      "logps/chosen": -238.94827270507812,
+      "logps/rejected": -326.12786865234375,
+      "loss": 0.1505,
+      "rewards/chosen": 1.7603483200073242,
+      "rewards/margins": 9.71554183959961,
+      "rewards/rejected": -7.955193519592285,
+      "step": 340
+    },
+    {
+      "epoch": 0.22,
+      "grad_norm": 22.41493034362793,
+      "kl": 0.0,
+      "learning_rate": 4.3314366998577524e-07,
+      "logps/chosen": -253.826416015625,
+      "logps/rejected": -330.7270812988281,
+      "loss": 0.1413,
+      "rewards/chosen": 1.8580402135849,
+      "rewards/margins": 9.722744941711426,
+      "rewards/rejected": -7.8647050857543945,
+      "step": 345
+    },
+    {
+      "epoch": 0.22,
+      "grad_norm": 26.315628051757812,
+      "kl": 0.0,
+      "learning_rate": 4.313655761024182e-07,
+      "logps/chosen": -286.4155578613281,
+      "logps/rejected": -342.62652587890625,
+      "loss": 0.1422,
+      "rewards/chosen": 1.7754218578338623,
+      "rewards/margins": 10.7272367477417,
+      "rewards/rejected": -8.951814651489258,
+      "step": 350
+    },
+    {
+      "epoch": 0.23,
+      "grad_norm": 22.07290267944336,
+      "kl": 0.0,
+      "learning_rate": 4.2958748221906114e-07,
+      "logps/chosen": -239.4450225830078,
+      "logps/rejected": -310.7589111328125,
+      "loss": 0.1108,
+      "rewards/chosen": 1.9980005025863647,
+      "rewards/margins": 10.0792818069458,
+      "rewards/rejected": -8.081281661987305,
+      "step": 355
+    },
+    {
+      "epoch": 0.23,
+      "grad_norm": 22.56456756591797,
+      "kl": 0.0,
+      "learning_rate": 4.278093883357041e-07,
+      "logps/chosen": -269.43560791015625,
+      "logps/rejected": -328.5326843261719,
+      "loss": 0.1524,
+      "rewards/chosen": 1.747078537940979,
+      "rewards/margins": 9.907144546508789,
+      "rewards/rejected": -8.160065650939941,
+      "step": 360
+    },
+    {
+      "epoch": 0.23,
+      "grad_norm": 24.159225463867188,
+      "kl": 0.0,
+      "learning_rate": 4.260312944523471e-07,
+      "logps/chosen": -241.72286987304688,
+      "logps/rejected": -353.99493408203125,
+      "loss": 0.1185,
+      "rewards/chosen": 2.0929551124572754,
+      "rewards/margins": 10.861230850219727,
+      "rewards/rejected": -8.768275260925293,
+      "step": 365
+    },
+    {
+      "epoch": 0.24,
+      "grad_norm": 21.119632720947266,
+      "kl": 0.0,
+      "learning_rate": 4.2425320056899e-07,
+      "logps/chosen": -263.73370361328125,
+      "logps/rejected": -343.69866943359375,
+      "loss": 0.1164,
+      "rewards/chosen": 1.9714298248291016,
+      "rewards/margins": 10.647806167602539,
+      "rewards/rejected": -8.676377296447754,
+      "step": 370
+    },
+    {
+      "epoch": 0.24,
+      "grad_norm": 23.59329605102539,
+      "kl": 0.0,
+      "learning_rate": 4.22475106685633e-07,
+      "logps/chosen": -244.0771026611328,
+      "logps/rejected": -348.35693359375,
+      "loss": 0.1218,
+      "rewards/chosen": 1.8124616146087646,
+      "rewards/margins": 10.569868087768555,
+      "rewards/rejected": -8.757406234741211,
+      "step": 375
+    },
+    {
+      "epoch": 0.24,
+      "grad_norm": 23.84940528869629,
+      "kl": 0.0,
+      "learning_rate": 4.2069701280227595e-07,
+      "logps/chosen": -241.4117889404297,
+      "logps/rejected": -336.4566345214844,
+      "loss": 0.1294,
+      "rewards/chosen": 1.8767503499984741,
+      "rewards/margins": 10.489054679870605,
+      "rewards/rejected": -8.612303733825684,
+      "step": 380
+    },
+    {
+      "epoch": 0.25,
+      "grad_norm": 16.005657196044922,
+      "kl": 0.0,
+      "learning_rate": 4.189189189189189e-07,
+      "logps/chosen": -307.21600341796875,
+      "logps/rejected": -341.0569152832031,
+      "loss": 0.1154,
+      "rewards/chosen": 1.901063323020935,
+      "rewards/margins": 10.400434494018555,
+      "rewards/rejected": -8.499369621276855,
+      "step": 385
+    },
+    {
+      "epoch": 0.25,
+      "grad_norm": 16.274534225463867,
+      "kl": 0.0,
+      "learning_rate": 4.1714082503556185e-07,
+      "logps/chosen": -253.8114471435547,
+      "logps/rejected": -333.0570983886719,
+      "loss": 0.1327,
+      "rewards/chosen": 1.955413818359375,
+      "rewards/margins": 10.895428657531738,
+      "rewards/rejected": -8.940014839172363,
+      "step": 390
+    },
+    {
+      "epoch": 0.25,
+      "grad_norm": 20.72968101501465,
+      "kl": 0.0,
+      "learning_rate": 4.153627311522048e-07,
+      "logps/chosen": -261.703857421875,
+      "logps/rejected": -339.4162902832031,
+      "loss": 0.1308,
+      "rewards/chosen": 1.7927815914154053,
+      "rewards/margins": 10.807384490966797,
+      "rewards/rejected": -9.014602661132812,
+      "step": 395
+    },
+    {
+      "epoch": 0.26,
+      "grad_norm": 21.173845291137695,
+      "kl": 0.0,
+      "learning_rate": 4.135846372688478e-07,
+      "logps/chosen": -260.5010681152344,
+      "logps/rejected": -359.41851806640625,
+      "loss": 0.1131,
+      "rewards/chosen": 1.8654544353485107,
+      "rewards/margins": 11.833639144897461,
+      "rewards/rejected": -9.968184471130371,
+      "step": 400
+    },
+    {
+      "epoch": 0.26,
+      "grad_norm": 17.008028030395508,
+      "kl": 0.0,
+      "learning_rate": 4.1180654338549077e-07,
+      "logps/chosen": -234.20315551757812,
+      "logps/rejected": -342.5408020019531,
+      "loss": 0.1224,
+      "rewards/chosen": 2.086550235748291,
+      "rewards/margins": 10.627596855163574,
+      "rewards/rejected": -8.541048049926758,
+      "step": 405
+    },
+    {
+      "epoch": 0.26,
+      "grad_norm": 19.24855613708496,
+      "kl": 0.0,
+      "learning_rate": 4.100284495021337e-07,
+      "logps/chosen": -282.03106689453125,
+      "logps/rejected": -347.0184631347656,
+      "loss": 0.1212,
+      "rewards/chosen": 1.9099693298339844,
+      "rewards/margins": 10.274763107299805,
+      "rewards/rejected": -8.36479377746582,
+      "step": 410
+    },
+    {
+      "epoch": 0.27,
+      "grad_norm": 19.914691925048828,
+      "kl": 0.0,
+      "learning_rate": 4.082503556187766e-07,
+      "logps/chosen": -278.811767578125,
+      "logps/rejected": -348.43096923828125,
+      "loss": 0.1209,
+      "rewards/chosen": 1.7562357187271118,
+      "rewards/margins": 11.304061889648438,
+      "rewards/rejected": -9.547826766967773,
+      "step": 415
+    },
+    {
+      "epoch": 0.27,
+      "grad_norm": 32.35190963745117,
+      "kl": 0.0,
+      "learning_rate": 4.064722617354196e-07,
+      "logps/chosen": -272.9192199707031,
+      "logps/rejected": -339.4073181152344,
+      "loss": 0.1005,
+      "rewards/chosen": 2.1069161891937256,
+      "rewards/margins": 11.689797401428223,
+      "rewards/rejected": -9.582880973815918,
+      "step": 420
+    },
+    {
+      "epoch": 0.27,
+      "grad_norm": 16.27385902404785,
+      "kl": 0.0,
+      "learning_rate": 4.0469416785206256e-07,
+      "logps/chosen": -233.26083374023438,
+      "logps/rejected": -329.8385314941406,
+      "loss": 0.1135,
+      "rewards/chosen": 2.1503446102142334,
+      "rewards/margins": 10.886509895324707,
+      "rewards/rejected": -8.736165046691895,
+      "step": 425
+    },
+    {
+      "epoch": 0.28,
+      "grad_norm": 22.10724449157715,
+      "kl": 0.0,
+      "learning_rate": 4.0291607396870553e-07,
+      "logps/chosen": -267.9533996582031,
+      "logps/rejected": -359.1152038574219,
+      "loss": 0.1197,
+      "rewards/chosen": 1.8767788410186768,
+      "rewards/margins": 11.441519737243652,
+      "rewards/rejected": -9.564741134643555,
+      "step": 430
+    },
+    {
+      "epoch": 0.28,
+      "grad_norm": 29.721330642700195,
+      "kl": 0.0,
+      "learning_rate": 4.011379800853485e-07,
+      "logps/chosen": -268.9942932128906,
+      "logps/rejected": -377.97308349609375,
+      "loss": 0.1061,
+      "rewards/chosen": 2.016845703125,
+      "rewards/margins": 11.474628448486328,
+      "rewards/rejected": -9.457781791687012,
+      "step": 435
+    },
+    {
+      "epoch": 0.28,
+      "grad_norm": 24.95069122314453,
+      "kl": 0.0,
+      "learning_rate": 3.993598862019915e-07,
+      "logps/chosen": -279.73944091796875,
+      "logps/rejected": -370.55474853515625,
+      "loss": 0.1045,
+      "rewards/chosen": 2.340444803237915,
+      "rewards/margins": 11.733491897583008,
+      "rewards/rejected": -9.393046379089355,
+      "step": 440
+    },
+    {
+      "epoch": 0.28,
+      "grad_norm": 26.99854278564453,
+      "kl": 0.0,
+      "learning_rate": 3.975817923186344e-07,
+      "logps/chosen": -245.24417114257812,
+      "logps/rejected": -336.7696533203125,
+      "loss": 0.1234,
+      "rewards/chosen": 2.2276217937469482,
+      "rewards/margins": 10.63569164276123,
+      "rewards/rejected": -8.40807056427002,
+      "step": 445
+    },
+    {
+      "epoch": 0.29,
+      "grad_norm": 22.343507766723633,
+      "kl": 0.0,
+      "learning_rate": 3.9580369843527737e-07,
+      "logps/chosen": -250.68197631835938,
+      "logps/rejected": -364.1970520019531,
+      "loss": 0.1342,
+      "rewards/chosen": 2.1794769763946533,
+      "rewards/margins": 10.827537536621094,
+      "rewards/rejected": -8.64806079864502,
+      "step": 450
+    },
+    {
+      "epoch": 0.29,
+      "grad_norm": 16.532583236694336,
+      "kl": 0.0,
+      "learning_rate": 3.940256045519203e-07,
+      "logps/chosen": -246.8504638671875,
+      "logps/rejected": -336.2493896484375,
+      "loss": 0.1151,
+      "rewards/chosen": 1.9471546411514282,
+      "rewards/margins": 11.059728622436523,
+      "rewards/rejected": -9.112573623657227,
+      "step": 455
+    },
+    {
+      "epoch": 0.29,
+      "grad_norm": 19.48154067993164,
+      "kl": 0.0,
+      "learning_rate": 3.9224751066856327e-07,
+      "logps/chosen": -246.47671508789062,
+      "logps/rejected": -328.21038818359375,
+      "loss": 0.1156,
+      "rewards/chosen": 2.022249221801758,
+      "rewards/margins": 11.125974655151367,
+      "rewards/rejected": -9.103724479675293,
+      "step": 460
+    },
+    {
+      "epoch": 0.3,
+      "grad_norm": 17.954296112060547,
+      "kl": 0.0,
+      "learning_rate": 3.9046941678520624e-07,
+      "logps/chosen": -254.69808959960938,
+      "logps/rejected": -347.45758056640625,
+      "loss": 0.1254,
+      "rewards/chosen": 2.0911641120910645,
+      "rewards/margins": 11.680787086486816,
+      "rewards/rejected": -9.58962345123291,
+      "step": 465
+    },
+    {
+      "epoch": 0.3,
+      "grad_norm": 21.16987419128418,
+      "kl": 0.0,
+      "learning_rate": 3.886913229018492e-07,
+      "logps/chosen": -226.58358764648438,
+      "logps/rejected": -334.48138427734375,
+      "loss": 0.1437,
+      "rewards/chosen": 1.8157398700714111,
+      "rewards/margins": 10.399962425231934,
+      "rewards/rejected": -8.584221839904785,
+      "step": 470
+    },
+    {
+      "epoch": 0.3,
+      "grad_norm": 18.404508590698242,
+      "kl": 0.0,
+      "learning_rate": 3.8691322901849213e-07,
+      "logps/chosen": -263.36395263671875,
+      "logps/rejected": -337.2552795410156,
+      "loss": 0.1078,
+      "rewards/chosen": 2.1132290363311768,
+      "rewards/margins": 11.090356826782227,
+      "rewards/rejected": -8.977127075195312,
+      "step": 475
+    },
+    {
+      "epoch": 0.31,
+      "grad_norm": 18.355857849121094,
+      "kl": 0.0,
+      "learning_rate": 3.851351351351351e-07,
+      "logps/chosen": -256.04412841796875,
+      "logps/rejected": -356.82073974609375,
+      "loss": 0.1145,
+      "rewards/chosen": 1.9654195308685303,
+      "rewards/margins": 11.159137725830078,
+      "rewards/rejected": -9.193717956542969,
+      "step": 480
+    },
+    {
+      "epoch": 0.31,
+      "grad_norm": 16.297401428222656,
+      "kl": 0.0,
+      "learning_rate": 3.833570412517781e-07,
+      "logps/chosen": -330.457275390625,
+      "logps/rejected": -367.7776794433594,
+      "loss": 0.113,
+      "rewards/chosen": 1.8794314861297607,
+      "rewards/margins": 11.851155281066895,
+      "rewards/rejected": -9.971722602844238,
+      "step": 485
+    },
+    {
+      "epoch": 0.31,
+      "grad_norm": 22.856266021728516,
+      "kl": 0.0,
+      "learning_rate": 3.8157894736842105e-07,
+      "logps/chosen": -254.3659210205078,
+      "logps/rejected": -343.2192687988281,
+      "loss": 0.1138,
+      "rewards/chosen": 2.034006118774414,
+      "rewards/margins": 11.903759002685547,
+      "rewards/rejected": -9.869752883911133,
+      "step": 490
+    },
+    {
+      "epoch": 0.32,
+      "grad_norm": 21.443723678588867,
+      "kl": 0.0,
+      "learning_rate": 3.7980085348506403e-07,
+      "logps/chosen": -278.9993591308594,
+      "logps/rejected": -333.3291320800781,
+      "loss": 0.1287,
+      "rewards/chosen": 1.906224012374878,
+      "rewards/margins": 10.766765594482422,
+      "rewards/rejected": -8.860541343688965,
+      "step": 495
+    },
+    {
+      "epoch": 0.32,
+      "grad_norm": 24.97509002685547,
+      "kl": 0.0,
+      "learning_rate": 3.7802275960170695e-07,
+      "logps/chosen": -238.44290161132812,
+      "logps/rejected": -338.4499206542969,
+      "loss": 0.1142,
+      "rewards/chosen": 2.103523015975952,
+      "rewards/margins": 11.162189483642578,
+      "rewards/rejected": -9.058666229248047,
+      "step": 500
+    },
+    {
+      "epoch": 0.32,
+      "grad_norm": 22.563114166259766,
+      "kl": 0.0,
+      "learning_rate": 3.7624466571834987e-07,
+      "logps/chosen": -238.7085723876953,
+      "logps/rejected": -352.23663330078125,
+      "loss": 0.1396,
+      "rewards/chosen": 1.8689464330673218,
+      "rewards/margins": 10.945247650146484,
+      "rewards/rejected": -9.076301574707031,
+      "step": 505
+    },
+    {
+      "epoch": 0.33,
+      "grad_norm": 17.017030715942383,
+      "kl": 0.0,
+      "learning_rate": 3.7446657183499284e-07,
+      "logps/chosen": -241.0718231201172,
+      "logps/rejected": -340.98968505859375,
+      "loss": 0.1197,
+      "rewards/chosen": 1.9388542175292969,
+      "rewards/margins": 11.241914749145508,
+      "rewards/rejected": -9.303060531616211,
+      "step": 510
+    },
+    {
+      "epoch": 0.33,
+      "grad_norm": 16.066030502319336,
+      "kl": 0.0,
+      "learning_rate": 3.726884779516358e-07,
+      "logps/chosen": -224.85086059570312,
+      "logps/rejected": -349.51800537109375,
+      "loss": 0.1094,
+      "rewards/chosen": 2.1027920246124268,
+      "rewards/margins": 11.039840698242188,
+      "rewards/rejected": -8.937047958374023,
+      "step": 515
+    },
+    {
+      "epoch": 0.33,
+      "grad_norm": 17.983535766601562,
+      "kl": 0.0,
+      "learning_rate": 3.709103840682788e-07,
+      "logps/chosen": -269.41693115234375,
+      "logps/rejected": -347.45477294921875,
+      "loss": 0.1234,
+      "rewards/chosen": 2.2909629344940186,
+      "rewards/margins": 11.061802864074707,
+      "rewards/rejected": -8.770838737487793,
+      "step": 520
+    },
+    {
+      "epoch": 0.34,
+      "grad_norm": 32.42300796508789,
+      "kl": 0.0,
+      "learning_rate": 3.6913229018492176e-07,
+      "logps/chosen": -274.3470764160156,
+      "logps/rejected": -335.10345458984375,
+      "loss": 0.1186,
+      "rewards/chosen": 2.1239845752716064,
+      "rewards/margins": 11.3440523147583,
+      "rewards/rejected": -9.220067977905273,
+      "step": 525
+    },
+    {
+      "epoch": 0.34,
+      "grad_norm": 16.601402282714844,
+      "kl": 0.0,
+      "learning_rate": 3.6735419630156474e-07,
+      "logps/chosen": -232.43392944335938,
+      "logps/rejected": -354.2206726074219,
+      "loss": 0.1218,
+      "rewards/chosen": 2.117833137512207,
+      "rewards/margins": 11.192573547363281,
+      "rewards/rejected": -9.07474136352539,
+      "step": 530
+    },
+    {
+      "epoch": 0.34,
+      "grad_norm": 24.04266357421875,
+      "kl": 0.0,
+      "learning_rate": 3.655761024182077e-07,
+      "logps/chosen": -277.41241455078125,
+      "logps/rejected": -334.04815673828125,
+      "loss": 0.1025,
+      "rewards/chosen": 2.3604273796081543,
+      "rewards/margins": 11.987722396850586,
+      "rewards/rejected": -9.627294540405273,
+      "step": 535
+    },
+    {
+      "epoch": 0.35,
+      "grad_norm": 15.607158660888672,
+      "kl": 0.0,
+      "learning_rate": 3.637980085348506e-07,
+      "logps/chosen": -286.7027282714844,
+      "logps/rejected": -319.68896484375,
+      "loss": 0.1124,
+      "rewards/chosen": 2.4556756019592285,
+      "rewards/margins": 11.582249641418457,
+      "rewards/rejected": -9.126574516296387,
+      "step": 540
+    },
+    {
+      "epoch": 0.35,
+      "grad_norm": 22.26024055480957,
+      "kl": 0.0,
+      "learning_rate": 3.6201991465149355e-07,
+      "logps/chosen": -239.4503173828125,
+      "logps/rejected": -345.2344970703125,
+      "loss": 0.112,
+      "rewards/chosen": 1.9915698766708374,
+      "rewards/margins": 11.853796005249023,
+      "rewards/rejected": -9.862226486206055,
+      "step": 545
+    },
+    {
+      "epoch": 0.35,
+      "grad_norm": 29.981088638305664,
+      "kl": 0.0,
+      "learning_rate": 3.602418207681365e-07,
+      "logps/chosen": -241.8583984375,
+      "logps/rejected": -365.6163024902344,
+      "loss": 0.1207,
+      "rewards/chosen": 1.9029747247695923,
+      "rewards/margins": 11.538789749145508,
+      "rewards/rejected": -9.635814666748047,
+      "step": 550
+    },
+    {
+      "epoch": 0.36,
+      "grad_norm": 17.248044967651367,
+      "kl": 0.0,
+      "learning_rate": 3.584637268847795e-07,
+      "logps/chosen": -255.35635375976562,
+      "logps/rejected": -348.6636962890625,
+      "loss": 0.1267,
+      "rewards/chosen": 2.0199761390686035,
+      "rewards/margins": 11.458740234375,
+      "rewards/rejected": -9.438763618469238,
+      "step": 555
+    },
+    {
+      "epoch": 0.36,
+      "grad_norm": 20.46364974975586,
+      "kl": 0.0,
+      "learning_rate": 3.5668563300142247e-07,
+      "logps/chosen": -271.98846435546875,
+      "logps/rejected": -319.4902038574219,
+      "loss": 0.1209,
+      "rewards/chosen": 2.140854597091675,
+      "rewards/margins": 11.506881713867188,
+      "rewards/rejected": -9.36602783203125,
+      "step": 560
+    },
+    {
+      "epoch": 0.36,
+      "grad_norm": 13.751622200012207,
+      "kl": 0.0,
+      "learning_rate": 3.5490753911806545e-07,
+      "logps/chosen": -249.31838989257812,
+      "logps/rejected": -363.65313720703125,
+      "loss": 0.1066,
+      "rewards/chosen": 2.302562713623047,
+      "rewards/margins": 12.324169158935547,
+      "rewards/rejected": -10.0216064453125,
+      "step": 565
+    },
+    {
+      "epoch": 0.36,
+      "grad_norm": 23.35749053955078,
+      "kl": 0.0,
+      "learning_rate": 3.5312944523470837e-07,
+      "logps/chosen": -242.62533569335938,
+      "logps/rejected": -326.2989807128906,
+      "loss": 0.1213,
+      "rewards/chosen": 2.164386749267578,
+      "rewards/margins": 11.125585556030273,
+      "rewards/rejected": -8.961198806762695,
+      "step": 570
+    },
+    {
+      "epoch": 0.37,
+      "grad_norm": 22.92085075378418,
+      "kl": 0.0,
+      "learning_rate": 3.5135135135135134e-07,
+      "logps/chosen": -215.1080322265625,
+      "logps/rejected": -343.92315673828125,
+      "loss": 0.1096,
+      "rewards/chosen": 2.084261417388916,
+      "rewards/margins": 12.25261402130127,
+      "rewards/rejected": -10.168352127075195,
+      "step": 575
+    },
+    {
+      "epoch": 0.37,
+      "grad_norm": 19.450267791748047,
+      "kl": 0.0,
+      "learning_rate": 3.495732574679943e-07,
+      "logps/chosen": -217.7822723388672,
+      "logps/rejected": -329.91375732421875,
+      "loss": 0.1294,
+      "rewards/chosen": 2.3254265785217285,
+      "rewards/margins": 11.483414649963379,
+      "rewards/rejected": -9.157987594604492,
+      "step": 580
+    },
+    {
+      "epoch": 0.37,
+      "grad_norm": 18.078920364379883,
+      "kl": 0.0,
+      "learning_rate": 3.4779516358463724e-07,
+      "logps/chosen": -253.4072723388672,
+      "logps/rejected": -344.491455078125,
+      "loss": 0.0927,
+      "rewards/chosen": 2.3925743103027344,
+      "rewards/margins": 12.241676330566406,
+      "rewards/rejected": -9.849101066589355,
+      "step": 585
+    },
+    {
+      "epoch": 0.38,
+      "grad_norm": 17.867704391479492,
+      "kl": 0.0,
+      "learning_rate": 3.460170697012802e-07,
+      "logps/chosen": -253.86178588867188,
+      "logps/rejected": -367.8619689941406,
+      "loss": 0.1152,
+      "rewards/chosen": 2.281512975692749,
+      "rewards/margins": 12.078498840332031,
+      "rewards/rejected": -9.79698657989502,
+      "step": 590
+    },
+    {
+      "epoch": 0.38,
+      "grad_norm": 19.863170623779297,
+      "kl": 0.0,
+      "learning_rate": 3.442389758179232e-07,
+      "logps/chosen": -289.83990478515625,
+      "logps/rejected": -346.80987548828125,
+      "loss": 0.1098,
+      "rewards/chosen": 2.1612110137939453,
+      "rewards/margins": 12.505953788757324,
+      "rewards/rejected": -10.344742774963379,
+      "step": 595
+    },
+    {
+      "epoch": 0.38,
+      "grad_norm": 14.594950675964355,
+      "kl": 0.0,
+      "learning_rate": 3.424608819345661e-07,
+      "logps/chosen": -199.16708374023438,
+      "logps/rejected": -352.14447021484375,
+      "loss": 0.0929,
+      "rewards/chosen": 1.9940685033798218,
+      "rewards/margins": 12.916508674621582,
+      "rewards/rejected": -10.922439575195312,
+      "step": 600
+    },
+    {
+      "epoch": 0.39,
+      "grad_norm": 20.205915451049805,
+      "kl": 0.0,
+      "learning_rate": 3.406827880512091e-07,
+      "logps/chosen": -251.85546875,
+      "logps/rejected": -347.35113525390625,
+      "loss": 0.1137,
+      "rewards/chosen": 2.3019704818725586,
+      "rewards/margins": 12.349814414978027,
+      "rewards/rejected": -10.047843933105469,
+      "step": 605
+    },
+    {
+      "epoch": 0.39,
+      "grad_norm": 21.62274742126465,
+      "kl": 0.0,
+      "learning_rate": 3.3890469416785205e-07,
+      "logps/chosen": -241.4209442138672,
+      "logps/rejected": -366.311767578125,
+      "loss": 0.104,
+      "rewards/chosen": 2.3162879943847656,
+      "rewards/margins": 11.884176254272461,
+      "rewards/rejected": -9.567889213562012,
+      "step": 610
+    },
+    {
+      "epoch": 0.39,
+      "grad_norm": 18.622377395629883,
+      "kl": 0.0,
+      "learning_rate": 3.37126600284495e-07,
+      "logps/chosen": -270.2699279785156,
+      "logps/rejected": -338.7982482910156,
+      "loss": 0.1127,
+      "rewards/chosen": 2.2165682315826416,
+      "rewards/margins": 11.826861381530762,
+      "rewards/rejected": -9.6102933883667,
+      "step": 615
+    },
+    {
+      "epoch": 0.4,
+      "grad_norm": 19.930374145507812,
+      "kl": 0.0,
+      "learning_rate": 3.35348506401138e-07,
+      "logps/chosen": -265.74432373046875,
+      "logps/rejected": -366.4137268066406,
+      "loss": 0.1235,
+      "rewards/chosen": 2.140533924102783,
+      "rewards/margins": 12.077839851379395,
+      "rewards/rejected": -9.937305450439453,
+      "step": 620
+    },
+    {
+      "epoch": 0.4,
+      "grad_norm": 20.78235626220703,
+      "kl": 0.0,
+      "learning_rate": 3.335704125177809e-07,
+      "logps/chosen": -251.783935546875,
+      "logps/rejected": -321.7081298828125,
+      "loss": 0.1251,
+      "rewards/chosen": 2.4168410301208496,
+      "rewards/margins": 11.567142486572266,
+      "rewards/rejected": -9.150300979614258,
+      "step": 625
+    },
+    {
+      "epoch": 0.4,
+      "grad_norm": 18.487886428833008,
+      "kl": 0.0,
+      "learning_rate": 3.3179231863442384e-07,
+      "logps/chosen": -244.60250854492188,
+      "logps/rejected": -386.168212890625,
+      "loss": 0.0922,
+      "rewards/chosen": 2.371992588043213,
+      "rewards/margins": 13.357122421264648,
+      "rewards/rejected": -10.985128402709961,
+      "step": 630
+    },
+    {
+      "epoch": 0.41,
+      "grad_norm": 16.354122161865234,
+      "kl": 0.0,
+      "learning_rate": 3.300142247510668e-07,
+      "logps/chosen": -252.8701171875,
+      "logps/rejected": -334.31744384765625,
+      "loss": 0.1003,
+      "rewards/chosen": 2.335937023162842,
+      "rewards/margins": 12.5576753616333,
+      "rewards/rejected": -10.2217378616333,
+      "step": 635
+    },
+    {
+      "epoch": 0.41,
+      "grad_norm": 16.126724243164062,
+      "kl": 0.0,
+      "learning_rate": 3.282361308677098e-07,
+      "logps/chosen": -240.2421417236328,
+      "logps/rejected": -346.31414794921875,
+      "loss": 0.1069,
+      "rewards/chosen": 2.2422432899475098,
+      "rewards/margins": 11.99687671661377,
+      "rewards/rejected": -9.754633903503418,
+      "step": 640
+    },
+    {
+      "epoch": 0.41,
+      "grad_norm": 21.066267013549805,
+      "kl": 0.0,
+      "learning_rate": 3.2645803698435276e-07,
+      "logps/chosen": -264.81964111328125,
+      "logps/rejected": -338.95379638671875,
+      "loss": 0.1126,
+      "rewards/chosen": 2.3289685249328613,
+      "rewards/margins": 11.437704086303711,
+      "rewards/rejected": -9.108736038208008,
+      "step": 645
+    },
+    {
+      "epoch": 0.42,
+      "grad_norm": 14.892672538757324,
+      "kl": 0.0,
+      "learning_rate": 3.2467994310099573e-07,
+      "logps/chosen": -240.17236328125,
+      "logps/rejected": -355.543212890625,
+      "loss": 0.0884,
+      "rewards/chosen": 2.1047472953796387,
+      "rewards/margins": 12.319875717163086,
+      "rewards/rejected": -10.215127944946289,
+      "step": 650
+    },
+    {
+      "epoch": 0.42,
+      "grad_norm": 22.722187042236328,
+      "kl": 0.0,
+      "learning_rate": 3.229018492176387e-07,
+      "logps/chosen": -235.84341430664062,
+      "logps/rejected": -363.74041748046875,
+      "loss": 0.0958,
+      "rewards/chosen": 2.262721300125122,
+      "rewards/margins": 12.947868347167969,
+      "rewards/rejected": -10.685147285461426,
+      "step": 655
+    },
+    {
+      "epoch": 0.42,
+      "grad_norm": 16.061914443969727,
+      "kl": 0.0,
+      "learning_rate": 3.211237553342817e-07,
+      "logps/chosen": -232.480224609375,
+      "logps/rejected": -363.93792724609375,
+      "loss": 0.0947,
+      "rewards/chosen": 2.245720863342285,
+      "rewards/margins": 13.019895553588867,
+      "rewards/rejected": -10.774174690246582,
+      "step": 660
+    },
+    {
+      "epoch": 0.43,
+      "grad_norm": 17.95708656311035,
+      "kl": 0.0,
+      "learning_rate": 3.193456614509246e-07,
+      "logps/chosen": -266.2756652832031,
+      "logps/rejected": -349.5681457519531,
+      "loss": 0.0981,
+      "rewards/chosen": 2.0552451610565186,
+      "rewards/margins": 12.47476863861084,
+      "rewards/rejected": -10.419523239135742,
+      "step": 665
+    },
+    {
+      "epoch": 0.43,
+      "grad_norm": 22.32909393310547,
+      "kl": 0.0,
+      "learning_rate": 3.175675675675675e-07,
+      "logps/chosen": -262.5782165527344,
+      "logps/rejected": -350.5921936035156,
+      "loss": 0.1155,
+      "rewards/chosen": 2.324476718902588,
+      "rewards/margins": 12.327436447143555,
+      "rewards/rejected": -10.002958297729492,
+      "step": 670
+    },
+    {
+      "epoch": 0.43,
+      "grad_norm": 14.884044647216797,
+      "kl": 0.0,
+      "learning_rate": 3.157894736842105e-07,
+      "logps/chosen": -257.66094970703125,
+      "logps/rejected": -344.09869384765625,
+      "loss": 0.096,
+      "rewards/chosen": 2.2759206295013428,
+      "rewards/margins": 13.264117240905762,
+      "rewards/rejected": -10.988197326660156,
+      "step": 675
+    },
+    {
+      "epoch": 0.44,
+      "grad_norm": 18.041053771972656,
+      "kl": 0.0,
+      "learning_rate": 3.1401137980085347e-07,
+      "logps/chosen": -270.93646240234375,
+      "logps/rejected": -360.5249938964844,
+      "loss": 0.0958,
+      "rewards/chosen": 2.366556167602539,
+      "rewards/margins": 13.1105318069458,
+      "rewards/rejected": -10.743974685668945,
+      "step": 680
+    },
+    {
+      "epoch": 0.44,
+      "grad_norm": 17.0706729888916,
+      "kl": 0.0,
+      "learning_rate": 3.1223328591749644e-07,
+      "logps/chosen": -214.73690795898438,
+      "logps/rejected": -342.7330017089844,
+      "loss": 0.1107,
+      "rewards/chosen": 2.152832508087158,
+      "rewards/margins": 12.126806259155273,
+      "rewards/rejected": -9.973973274230957,
+      "step": 685
+    },
+    {
+      "epoch": 0.44,
+      "grad_norm": 25.26626968383789,
+      "kl": 0.0,
+      "learning_rate": 3.104551920341394e-07,
+      "logps/chosen": -293.3183898925781,
+      "logps/rejected": -356.9164123535156,
+      "loss": 0.0973,
+      "rewards/chosen": 2.3354926109313965,
+      "rewards/margins": 13.199078559875488,
+      "rewards/rejected": -10.863585472106934,
+      "step": 690
+    },
+    {
+      "epoch": 0.44,
+      "grad_norm": 28.42453384399414,
+      "kl": 0.0,
+      "learning_rate": 3.0867709815078234e-07,
+      "logps/chosen": -285.2272644042969,
+      "logps/rejected": -367.3414001464844,
+      "loss": 0.1091,
+      "rewards/chosen": 2.12477707862854,
+      "rewards/margins": 13.724832534790039,
+      "rewards/rejected": -11.600054740905762,
+      "step": 695
+    },
+    {
+      "epoch": 0.45,
+      "grad_norm": 17.899433135986328,
+      "kl": 0.0,
+      "learning_rate": 3.068990042674253e-07,
+      "logps/chosen": -262.5505065917969,
+      "logps/rejected": -371.45025634765625,
+      "loss": 0.1083,
+      "rewards/chosen": 2.060147762298584,
+      "rewards/margins": 12.760570526123047,
+      "rewards/rejected": -10.700422286987305,
+      "step": 700
+    },
+    {
+      "epoch": 0.45,
+      "grad_norm": 19.646583557128906,
+      "kl": 0.0,
+      "learning_rate": 3.051209103840683e-07,
+      "logps/chosen": -274.9856262207031,
+      "logps/rejected": -360.3489685058594,
+      "loss": 0.1105,
+      "rewards/chosen": 2.4847919940948486,
+      "rewards/margins": 13.168169975280762,
+      "rewards/rejected": -10.683378219604492,
+      "step": 705
+    },
+    {
+      "epoch": 0.45,
+      "grad_norm": 14.362700462341309,
+      "kl": 0.0,
+      "learning_rate": 3.033428165007112e-07,
+      "logps/chosen": -253.5523681640625,
+      "logps/rejected": -336.803466796875,
+      "loss": 0.1024,
+      "rewards/chosen": 2.3810598850250244,
+      "rewards/margins": 12.508206367492676,
+      "rewards/rejected": -10.127145767211914,
+      "step": 710
+    },
+    {
+      "epoch": 0.46,
+      "grad_norm": 22.93763542175293,
+      "kl": 0.0,
+      "learning_rate": 3.015647226173542e-07,
+      "logps/chosen": -229.1265869140625,
+      "logps/rejected": -354.0566101074219,
+      "loss": 0.1135,
+      "rewards/chosen": 2.1574866771698,
+      "rewards/margins": 12.193730354309082,
+      "rewards/rejected": -10.036243438720703,
+      "step": 715
+    },
+    {
+      "epoch": 0.46,
+      "grad_norm": 22.314695358276367,
+      "kl": 0.0,
+      "learning_rate": 2.9978662873399715e-07,
+      "logps/chosen": -274.2774353027344,
+      "logps/rejected": -405.2040100097656,
+      "loss": 0.0953,
+      "rewards/chosen": 2.191157817840576,
+      "rewards/margins": 13.628583908081055,
+      "rewards/rejected": -11.43742561340332,
+      "step": 720
+    },
+    {
+      "epoch": 0.46,
+      "grad_norm": 20.03668975830078,
+      "kl": 0.0,
+      "learning_rate": 2.9800853485064007e-07,
+      "logps/chosen": -250.0018768310547,
+      "logps/rejected": -340.919677734375,
+      "loss": 0.1022,
+      "rewards/chosen": 2.4375367164611816,
+      "rewards/margins": 12.8548583984375,
+      "rewards/rejected": -10.417322158813477,
+      "step": 725
+    },
+    {
+      "epoch": 0.47,
+      "grad_norm": 20.21332359313965,
+      "kl": 0.0,
+      "learning_rate": 2.9623044096728305e-07,
+      "logps/chosen": -262.5707092285156,
+      "logps/rejected": -341.51556396484375,
+      "loss": 0.113,
+      "rewards/chosen": 2.2380969524383545,
+      "rewards/margins": 12.677546501159668,
+      "rewards/rejected": -10.439449310302734,
+      "step": 730
+    },
+    {
+      "epoch": 0.47,
+      "grad_norm": 18.686241149902344,
+      "kl": 0.0,
+      "learning_rate": 2.94452347083926e-07,
+      "logps/chosen": -259.5292053222656,
+      "logps/rejected": -340.3607482910156,
+      "loss": 0.1188,
+      "rewards/chosen": 2.1912286281585693,
+      "rewards/margins": 12.15953540802002,
+      "rewards/rejected": -9.968307495117188,
+      "step": 735
+    },
+    {
+      "epoch": 0.47,
+      "grad_norm": 19.728899002075195,
+      "kl": 0.0,
+      "learning_rate": 2.92674253200569e-07,
+      "logps/chosen": -249.8451385498047,
+      "logps/rejected": -381.3360290527344,
+      "loss": 0.0967,
+      "rewards/chosen": 2.2508890628814697,
+      "rewards/margins": 13.840009689331055,
+      "rewards/rejected": -11.589120864868164,
+      "step": 740
+    },
+    {
+      "epoch": 0.48,
+      "grad_norm": 17.338115692138672,
+      "kl": 0.0,
+      "learning_rate": 2.9089615931721197e-07,
+      "logps/chosen": -268.9515380859375,
+      "logps/rejected": -368.5078125,
+      "loss": 0.1015,
+      "rewards/chosen": 2.28879976272583,
+      "rewards/margins": 13.396432876586914,
+      "rewards/rejected": -11.107633590698242,
+      "step": 745
+    },
+    {
+      "epoch": 0.48,
+      "grad_norm": 19.87563133239746,
+      "kl": 0.0,
+      "learning_rate": 2.8911806543385494e-07,
+      "logps/chosen": -293.7141418457031,
+      "logps/rejected": -355.55535888671875,
+      "loss": 0.1071,
+      "rewards/chosen": 2.4517719745635986,
+      "rewards/margins": 13.449313163757324,
+      "rewards/rejected": -10.997541427612305,
+      "step": 750
+    },
+    {
+      "epoch": 0.48,
+      "grad_norm": 19.807565689086914,
+      "kl": 0.0,
+      "learning_rate": 2.873399715504978e-07,
+      "logps/chosen": -248.6322784423828,
+      "logps/rejected": -353.8228759765625,
+      "loss": 0.1088,
+      "rewards/chosen": 2.2985501289367676,
+      "rewards/margins": 12.722993850708008,
+      "rewards/rejected": -10.424444198608398,
+      "step": 755
+    },
+    {
+      "epoch": 0.49,
+      "grad_norm": 14.098715782165527,
+      "kl": 0.0,
+      "learning_rate": 2.855618776671408e-07,
+      "logps/chosen": -240.1453857421875,
+      "logps/rejected": -330.6163635253906,
+      "loss": 0.0891,
+      "rewards/chosen": 2.550762414932251,
+      "rewards/margins": 12.813929557800293,
+      "rewards/rejected": -10.263166427612305,
+      "step": 760
+    },
+    {
+      "epoch": 0.49,
+      "grad_norm": 16.266538619995117,
+      "kl": 0.0,
+      "learning_rate": 2.8378378378378376e-07,
+      "logps/chosen": -236.1558380126953,
+      "logps/rejected": -360.686767578125,
+      "loss": 0.0973,
+      "rewards/chosen": 2.2286760807037354,
+      "rewards/margins": 13.117490768432617,
+      "rewards/rejected": -10.888814926147461,
+      "step": 765
+    },
+    {
+      "epoch": 0.49,
+      "grad_norm": 16.414714813232422,
+      "kl": 0.0,
+      "learning_rate": 2.8200568990042673e-07,
+      "logps/chosen": -283.36944580078125,
+      "logps/rejected": -368.7194519042969,
+      "loss": 0.0967,
+      "rewards/chosen": 2.4076812267303467,
+      "rewards/margins": 12.956690788269043,
+      "rewards/rejected": -10.549009323120117,
+      "step": 770
+    },
+    {
+      "epoch": 0.5,
+      "grad_norm": 18.088735580444336,
+      "kl": 0.0,
+      "learning_rate": 2.802275960170697e-07,
+      "logps/chosen": -302.5070495605469,
+      "logps/rejected": -348.53289794921875,
+      "loss": 0.1047,
+      "rewards/chosen": 2.4432990550994873,
+      "rewards/margins": 12.935566902160645,
+      "rewards/rejected": -10.492268562316895,
+      "step": 775
+    },
+    {
+      "epoch": 0.5,
+      "grad_norm": 16.778833389282227,
+      "kl": 0.0,
+      "learning_rate": 2.784495021337127e-07,
+      "logps/chosen": -243.6858367919922,
+      "logps/rejected": -366.0279235839844,
+      "loss": 0.0852,
+      "rewards/chosen": 2.5511088371276855,
+      "rewards/margins": 13.161443710327148,
+      "rewards/rejected": -10.610334396362305,
+      "step": 780
+    },
+    {
+      "epoch": 0.5,
+      "grad_norm": 17.379335403442383,
+      "kl": 0.0,
+      "learning_rate": 2.766714082503556e-07,
+      "logps/chosen": -279.8948669433594,
+      "logps/rejected": -352.0287170410156,
+      "loss": 0.0957,
+      "rewards/chosen": 2.464888334274292,
+      "rewards/margins": 13.103485107421875,
+      "rewards/rejected": -10.638595581054688,
+      "step": 785
+    },
+    {
+      "epoch": 0.51,
+      "grad_norm": 21.737380981445312,
+      "kl": 0.0,
+      "learning_rate": 2.7489331436699857e-07,
+      "logps/chosen": -258.93865966796875,
+      "logps/rejected": -345.5055236816406,
+      "loss": 0.1177,
+      "rewards/chosen": 2.079983711242676,
+      "rewards/margins": 12.90088176727295,
+      "rewards/rejected": -10.820898056030273,
+      "step": 790
+    },
+    {
+      "epoch": 0.51,
+      "grad_norm": 15.227828025817871,
+      "kl": 0.0,
+      "learning_rate": 2.7311522048364154e-07,
+      "logps/chosen": -233.96194458007812,
+      "logps/rejected": -370.0096130371094,
+      "loss": 0.1054,
+      "rewards/chosen": 2.696063280105591,
+      "rewards/margins": 13.426877975463867,
+      "rewards/rejected": -10.730813980102539,
+      "step": 795
+    },
+    {
+      "epoch": 0.51,
+      "grad_norm": 13.938480377197266,
+      "kl": 0.0,
+      "learning_rate": 2.7133712660028446e-07,
+      "logps/chosen": -243.730224609375,
+      "logps/rejected": -350.44183349609375,
+      "loss": 0.0976,
+      "rewards/chosen": 2.3282103538513184,
+      "rewards/margins": 13.082319259643555,
+      "rewards/rejected": -10.754108428955078,
+      "step": 800
+    },
+    {
+      "epoch": 0.52,
+      "grad_norm": 29.015178680419922,
+      "kl": 0.0,
+      "learning_rate": 2.6955903271692744e-07,
+      "logps/chosen": -216.66488647460938,
+      "logps/rejected": -368.9279479980469,
+      "loss": 0.1045,
+      "rewards/chosen": 2.2219786643981934,
+      "rewards/margins": 13.255975723266602,
+      "rewards/rejected": -11.033994674682617,
+      "step": 805
+    },
+    {
+      "epoch": 0.52,
+      "grad_norm": 12.759780883789062,
+      "kl": 0.0,
+      "learning_rate": 2.677809388335704e-07,
+      "logps/chosen": -232.0436248779297,
+      "logps/rejected": -358.2416076660156,
+      "loss": 0.0943,
+      "rewards/chosen": 2.235215425491333,
+      "rewards/margins": 12.901985168457031,
+      "rewards/rejected": -10.666769027709961,
+      "step": 810
+    },
+    {
+      "epoch": 0.52,
+      "grad_norm": 19.0715389251709,
+      "kl": 0.0,
+      "learning_rate": 2.6600284495021333e-07,
+      "logps/chosen": -245.0729217529297,
+      "logps/rejected": -379.9324645996094,
+      "loss": 0.1173,
+      "rewards/chosen": 2.423140287399292,
+      "rewards/margins": 13.03075885772705,
+      "rewards/rejected": -10.60761833190918,
+      "step": 815
+    },
+    {
+      "epoch": 0.52,
+      "grad_norm": 15.53042984008789,
+      "kl": 0.0,
+      "learning_rate": 2.642247510668563e-07,
+      "logps/chosen": -261.9339294433594,
+      "logps/rejected": -327.780517578125,
+      "loss": 0.0903,
+      "rewards/chosen": 2.4564433097839355,
+      "rewards/margins": 13.401847839355469,
+      "rewards/rejected": -10.945404052734375,
+      "step": 820
+    },
+    {
+      "epoch": 0.53,
+      "grad_norm": 26.67053985595703,
+      "kl": 0.0,
+      "learning_rate": 2.624466571834993e-07,
+      "logps/chosen": -248.34487915039062,
+      "logps/rejected": -345.79083251953125,
+      "loss": 0.1071,
+      "rewards/chosen": 2.359476089477539,
+      "rewards/margins": 13.113537788391113,
+      "rewards/rejected": -10.754061698913574,
+      "step": 825
+    },
+    {
+      "epoch": 0.53,
+      "grad_norm": 22.24759292602539,
+      "kl": 0.0,
+      "learning_rate": 2.6066856330014225e-07,
+      "logps/chosen": -262.4079895019531,
+      "logps/rejected": -364.2686462402344,
+      "loss": 0.1076,
+      "rewards/chosen": 2.37235689163208,
+      "rewards/margins": 13.127764701843262,
+      "rewards/rejected": -10.75540828704834,
+      "step": 830
+    },
+    {
+      "epoch": 0.53,
+      "grad_norm": 17.638994216918945,
+      "kl": 0.0,
+      "learning_rate": 2.5889046941678523e-07,
+      "logps/chosen": -273.74432373046875,
+      "logps/rejected": -383.0593566894531,
+      "loss": 0.0984,
+      "rewards/chosen": 2.275177478790283,
+      "rewards/margins": 14.033803939819336,
+      "rewards/rejected": -11.758626937866211,
+      "step": 835
+    },
+    {
+      "epoch": 0.54,
+      "grad_norm": 15.510226249694824,
+      "kl": 0.0,
+      "learning_rate": 2.5711237553342815e-07,
+      "logps/chosen": -268.3088073730469,
+      "logps/rejected": -372.16876220703125,
+      "loss": 0.075,
+      "rewards/chosen": 2.7448737621307373,
+      "rewards/margins": 14.005487442016602,
+      "rewards/rejected": -11.260614395141602,
+      "step": 840
+    },
+    {
+      "epoch": 0.54,
+      "grad_norm": 18.002506256103516,
+      "kl": 0.0,
+      "learning_rate": 2.5533428165007107e-07,
+      "logps/chosen": -214.6236572265625,
+      "logps/rejected": -379.697265625,
+      "loss": 0.1156,
+      "rewards/chosen": 2.090531349182129,
+      "rewards/margins": 14.037800788879395,
+      "rewards/rejected": -11.947270393371582,
+      "step": 845
+    },
+    {
+      "epoch": 0.54,
+      "grad_norm": 18.8648681640625,
+      "kl": 0.0,
+      "learning_rate": 2.5355618776671404e-07,
+      "logps/chosen": -230.46810913085938,
+      "logps/rejected": -345.8827209472656,
+      "loss": 0.1028,
+      "rewards/chosen": 2.3623595237731934,
+      "rewards/margins": 13.34724235534668,
+      "rewards/rejected": -10.984883308410645,
+      "step": 850
+    },
+    {
+      "epoch": 0.55,
+      "grad_norm": 15.514137268066406,
+      "kl": 0.0,
+      "learning_rate": 2.51778093883357e-07,
+      "logps/chosen": -224.1614227294922,
+      "logps/rejected": -365.08477783203125,
+      "loss": 0.0985,
+      "rewards/chosen": 2.2946102619171143,
+      "rewards/margins": 14.019018173217773,
+      "rewards/rejected": -11.724408149719238,
+      "step": 855
+    },
+    {
+      "epoch": 0.55,
+      "grad_norm": 17.679288864135742,
+      "kl": 0.0,
+      "learning_rate": 2.5e-07,
+      "logps/chosen": -231.95565795898438,
+      "logps/rejected": -349.056640625,
+      "loss": 0.0966,
+      "rewards/chosen": 2.3342556953430176,
+      "rewards/margins": 13.233372688293457,
+      "rewards/rejected": -10.899115562438965,
+      "step": 860
+    },
+    {
+      "epoch": 0.55,
+      "grad_norm": 19.04916000366211,
+      "kl": 0.0,
+      "learning_rate": 2.4822190611664296e-07,
+      "logps/chosen": -233.5281982421875,
+      "logps/rejected": -374.4831848144531,
+      "loss": 0.0922,
+      "rewards/chosen": 2.58280873298645,
+      "rewards/margins": 13.661170959472656,
+      "rewards/rejected": -11.078360557556152,
+      "step": 865
+    },
+    {
+      "epoch": 0.56,
+      "grad_norm": 16.834497451782227,
+      "kl": 0.0,
+      "learning_rate": 2.4644381223328594e-07,
+      "logps/chosen": -236.75106811523438,
+      "logps/rejected": -353.3365478515625,
+      "loss": 0.096,
+      "rewards/chosen": 2.7401375770568848,
+      "rewards/margins": 13.776086807250977,
+      "rewards/rejected": -11.035948753356934,
+      "step": 870
+    },
+    {
+      "epoch": 0.56,
+      "grad_norm": 24.789772033691406,
+      "kl": 0.0,
+      "learning_rate": 2.4466571834992886e-07,
+      "logps/chosen": -267.3653869628906,
+      "logps/rejected": -370.1556091308594,
+      "loss": 0.0949,
+      "rewards/chosen": 2.2669379711151123,
+      "rewards/margins": 13.626495361328125,
+      "rewards/rejected": -11.35955810546875,
+      "step": 875
+    },
+    {
+      "epoch": 0.56,
+      "grad_norm": 12.911144256591797,
+      "kl": 0.0,
+      "learning_rate": 2.4288762446657183e-07,
+      "logps/chosen": -250.23556518554688,
+      "logps/rejected": -363.0490417480469,
+      "loss": 0.0717,
+      "rewards/chosen": 2.6297707557678223,
+      "rewards/margins": 15.319549560546875,
+      "rewards/rejected": -12.689778327941895,
+      "step": 880
+    },
+    {
+      "epoch": 0.57,
+      "grad_norm": 18.159881591796875,
+      "kl": 0.0,
+      "learning_rate": 2.411095305832148e-07,
+      "logps/chosen": -276.918701171875,
+      "logps/rejected": -351.54132080078125,
+      "loss": 0.1062,
+      "rewards/chosen": 2.5637714862823486,
+      "rewards/margins": 13.2720365524292,
+      "rewards/rejected": -10.708267211914062,
+      "step": 885
+    },
+    {
+      "epoch": 0.57,
+      "grad_norm": 16.24278450012207,
+      "kl": 0.0,
+      "learning_rate": 2.393314366998578e-07,
+      "logps/chosen": -264.356689453125,
+      "logps/rejected": -345.72064208984375,
+      "loss": 0.1061,
+      "rewards/chosen": 2.57415509223938,
+      "rewards/margins": 13.096124649047852,
+      "rewards/rejected": -10.521968841552734,
+      "step": 890
+    },
+    {
+      "epoch": 0.57,
+      "grad_norm": 15.353395462036133,
+      "kl": 0.0,
+      "learning_rate": 2.375533428165007e-07,
+      "logps/chosen": -245.558349609375,
+      "logps/rejected": -346.11151123046875,
+      "loss": 0.1098,
+      "rewards/chosen": 2.0979220867156982,
+      "rewards/margins": 13.061151504516602,
+      "rewards/rejected": -10.963228225708008,
+      "step": 895
+    },
+    {
+      "epoch": 0.58,
+      "grad_norm": 15.291166305541992,
+      "kl": 0.0,
+      "learning_rate": 2.3577524893314365e-07,
+      "logps/chosen": -268.0149841308594,
+      "logps/rejected": -359.99676513671875,
+      "loss": 0.0935,
+      "rewards/chosen": 2.660017251968384,
+      "rewards/margins": 13.785112380981445,
+      "rewards/rejected": -11.125094413757324,
+      "step": 900
+    },
+    {
+      "epoch": 0.58,
+      "grad_norm": 27.338308334350586,
+      "kl": 0.0,
+      "learning_rate": 2.3399715504978662e-07,
+      "logps/chosen": -223.98062133789062,
+      "logps/rejected": -382.53924560546875,
+      "loss": 0.0733,
+      "rewards/chosen": 2.573464870452881,
+      "rewards/margins": 14.070528030395508,
+      "rewards/rejected": -11.497062683105469,
+      "step": 905
+    },
+    {
+      "epoch": 0.58,
+      "grad_norm": 28.126638412475586,
+      "kl": 0.0,
+      "learning_rate": 2.322190611664296e-07,
+      "logps/chosen": -225.90432739257812,
+      "logps/rejected": -369.765380859375,
+      "loss": 0.1068,
+      "rewards/chosen": 2.4907584190368652,
+      "rewards/margins": 13.410183906555176,
+      "rewards/rejected": -10.919425964355469,
+      "step": 910
+    },
+    {
+      "epoch": 0.59,
+      "grad_norm": 23.488555908203125,
+      "kl": 0.0,
+      "learning_rate": 2.304409672830725e-07,
+      "logps/chosen": -264.053466796875,
+      "logps/rejected": -347.4461364746094,
+      "loss": 0.0978,
+      "rewards/chosen": 2.3625271320343018,
+      "rewards/margins": 13.296697616577148,
+      "rewards/rejected": -10.934170722961426,
+      "step": 915
+    },
+    {
+      "epoch": 0.59,
+      "grad_norm": 16.538496017456055,
+      "kl": 0.0,
+      "learning_rate": 2.2866287339971549e-07,
+      "logps/chosen": -305.2149353027344,
+      "logps/rejected": -387.3183898925781,
+      "loss": 0.0895,
+      "rewards/chosen": 2.51884126663208,
+      "rewards/margins": 14.519546508789062,
+      "rewards/rejected": -12.000704765319824,
+      "step": 920
+    },
+    {
+      "epoch": 0.59,
+      "grad_norm": 14.570239067077637,
+      "kl": 0.0,
+      "learning_rate": 2.2688477951635846e-07,
+      "logps/chosen": -288.55859375,
+      "logps/rejected": -360.0328063964844,
+      "loss": 0.0891,
+      "rewards/chosen": 2.6060962677001953,
+      "rewards/margins": 13.63697338104248,
+      "rewards/rejected": -11.030878067016602,
+      "step": 925
+    },
+    {
+      "epoch": 0.6,
+      "grad_norm": 14.252922058105469,
+      "kl": 0.0,
+      "learning_rate": 2.251066856330014e-07,
+      "logps/chosen": -237.12985229492188,
+      "logps/rejected": -375.7847900390625,
+      "loss": 0.0968,
+      "rewards/chosen": 2.277179718017578,
+      "rewards/margins": 13.618906021118164,
+      "rewards/rejected": -11.341727256774902,
+      "step": 930
+    },
+    {
+      "epoch": 0.6,
+      "grad_norm": 14.224340438842773,
+      "kl": 0.0,
+      "learning_rate": 2.2332859174964438e-07,
+      "logps/chosen": -247.994384765625,
+      "logps/rejected": -384.1309509277344,
+      "loss": 0.0921,
+      "rewards/chosen": 2.5065815448760986,
+      "rewards/margins": 14.67822551727295,
+      "rewards/rejected": -12.17164421081543,
+      "step": 935
+    },
+    {
+      "epoch": 0.6,
+      "grad_norm": 15.533319473266602,
+      "kl": 0.0,
+      "learning_rate": 2.2155049786628733e-07,
+      "logps/chosen": -217.72982788085938,
+      "logps/rejected": -364.0068054199219,
+      "loss": 0.1044,
+      "rewards/chosen": 2.4342358112335205,
+      "rewards/margins": 13.825854301452637,
+      "rewards/rejected": -11.391618728637695,
+      "step": 940
+    },
+    {
+      "epoch": 0.6,
+      "grad_norm": 12.951379776000977,
+      "kl": 0.0,
+      "learning_rate": 2.1977240398293027e-07,
+      "logps/chosen": -216.9298858642578,
+      "logps/rejected": -391.88916015625,
+      "loss": 0.0959,
+      "rewards/chosen": 2.2248167991638184,
+      "rewards/margins": 14.502099990844727,
+      "rewards/rejected": -12.27728271484375,
+      "step": 945
+    },
+    {
+      "epoch": 0.61,
+      "grad_norm": 12.726017951965332,
+      "kl": 0.0,
+      "learning_rate": 2.1799431009957325e-07,
+      "logps/chosen": -250.4014129638672,
+      "logps/rejected": -367.2666931152344,
+      "loss": 0.099,
+      "rewards/chosen": 2.6326889991760254,
+      "rewards/margins": 14.203544616699219,
+      "rewards/rejected": -11.570856094360352,
+      "step": 950
+    },
+    {
+      "epoch": 0.61,
+      "grad_norm": 24.116321563720703,
+      "kl": 0.0,
+      "learning_rate": 2.1621621621621622e-07,
+      "logps/chosen": -256.0219421386719,
+      "logps/rejected": -359.89410400390625,
+      "loss": 0.1114,
+      "rewards/chosen": 2.291504383087158,
+      "rewards/margins": 13.204824447631836,
+      "rewards/rejected": -10.913320541381836,
+      "step": 955
+    },
+    {
+      "epoch": 0.61,
+      "grad_norm": 21.19695472717285,
+      "kl": 0.0,
+      "learning_rate": 2.1443812233285914e-07,
+      "logps/chosen": -253.9010009765625,
+      "logps/rejected": -348.592041015625,
+      "loss": 0.0887,
+      "rewards/chosen": 2.41105580329895,
+      "rewards/margins": 13.58563232421875,
+      "rewards/rejected": -11.174577713012695,
+      "step": 960
+    },
+    {
+      "epoch": 0.62,
+      "grad_norm": 19.317626953125,
+      "kl": 0.0,
+      "learning_rate": 2.1266002844950212e-07,
+      "logps/chosen": -245.3134765625,
+      "logps/rejected": -380.17327880859375,
+      "loss": 0.0998,
+      "rewards/chosen": 2.4610061645507812,
+      "rewards/margins": 13.953335762023926,
+      "rewards/rejected": -11.492330551147461,
+      "step": 965
+    },
+    {
+      "epoch": 0.62,
+      "grad_norm": 22.688295364379883,
+      "kl": 0.0,
+      "learning_rate": 2.108819345661451e-07,
+      "logps/chosen": -278.2052917480469,
+      "logps/rejected": -357.1881408691406,
+      "loss": 0.1063,
+      "rewards/chosen": 2.404897928237915,
+      "rewards/margins": 13.51880931854248,
+      "rewards/rejected": -11.113912582397461,
+      "step": 970
+    },
+    {
+      "epoch": 0.62,
+      "grad_norm": 14.956331253051758,
+      "kl": 0.0,
+      "learning_rate": 2.0910384068278806e-07,
+      "logps/chosen": -274.7865295410156,
+      "logps/rejected": -348.69024658203125,
+      "loss": 0.1178,
+      "rewards/chosen": 2.560181140899658,
+      "rewards/margins": 13.09942626953125,
+      "rewards/rejected": -10.539244651794434,
+      "step": 975
+    },
+    {
+      "epoch": 0.63,
+      "grad_norm": 14.130335807800293,
+      "kl": 0.0,
+      "learning_rate": 2.0732574679943098e-07,
+      "logps/chosen": -241.42257690429688,
+      "logps/rejected": -374.1187438964844,
+      "loss": 0.0728,
+      "rewards/chosen": 2.970536470413208,
+      "rewards/margins": 14.709304809570312,
+      "rewards/rejected": -11.738768577575684,
+      "step": 980
+    },
+    {
+      "epoch": 0.63,
+      "grad_norm": 20.794334411621094,
+      "kl": 0.0,
+      "learning_rate": 2.0554765291607396e-07,
+      "logps/chosen": -230.2430419921875,
+      "logps/rejected": -380.54608154296875,
+      "loss": 0.09,
+      "rewards/chosen": 2.673881769180298,
+      "rewards/margins": 14.4403076171875,
+      "rewards/rejected": -11.766425132751465,
+      "step": 985
+    },
+    {
+      "epoch": 0.63,
+      "grad_norm": 16.297340393066406,
+      "kl": 0.0,
+      "learning_rate": 2.0376955903271693e-07,
+      "logps/chosen": -264.10137939453125,
+      "logps/rejected": -372.11346435546875,
+      "loss": 0.0896,
+      "rewards/chosen": 2.7595813274383545,
+      "rewards/margins": 13.790555953979492,
+      "rewards/rejected": -11.030974388122559,
+      "step": 990
+    },
+    {
+      "epoch": 0.64,
+      "grad_norm": 20.304162979125977,
+      "kl": 0.0,
+      "learning_rate": 2.0199146514935988e-07,
+      "logps/chosen": -243.2149200439453,
+      "logps/rejected": -342.080322265625,
+      "loss": 0.1091,
+      "rewards/chosen": 2.3006443977355957,
+      "rewards/margins": 13.069559097290039,
+      "rewards/rejected": -10.768914222717285,
+      "step": 995
+    },
+    {
+      "epoch": 0.64,
+      "grad_norm": 19.80646514892578,
+      "kl": 0.0,
+      "learning_rate": 2.0021337126600283e-07,
+      "logps/chosen": -251.78524780273438,
+      "logps/rejected": -360.4002380371094,
+      "loss": 0.088,
+      "rewards/chosen": 2.7110652923583984,
+      "rewards/margins": 13.96354866027832,
+      "rewards/rejected": -11.252483367919922,
+      "step": 1000
+    },
+    {
+      "epoch": 0.64,
+      "grad_norm": 19.04472541809082,
+      "kl": 0.0,
+      "learning_rate": 1.984352773826458e-07,
+      "logps/chosen": -227.35873413085938,
+      "logps/rejected": -363.24713134765625,
+      "loss": 0.0952,
+      "rewards/chosen": 2.7294869422912598,
+      "rewards/margins": 14.287317276000977,
+      "rewards/rejected": -11.557830810546875,
+      "step": 1005
+    },
+    {
+      "epoch": 0.65,
+      "grad_norm": 14.0511474609375,
+      "kl": 0.0,
+      "learning_rate": 1.9665718349928875e-07,
+      "logps/chosen": -248.2740478515625,
+      "logps/rejected": -359.29534912109375,
+      "loss": 0.0783,
+      "rewards/chosen": 2.7295162677764893,
+      "rewards/margins": 14.418438911437988,
+      "rewards/rejected": -11.688921928405762,
+      "step": 1010
+    },
+    {
+      "epoch": 0.65,
+      "grad_norm": 21.88315773010254,
+      "kl": 0.0,
+      "learning_rate": 1.9487908961593172e-07,
+      "logps/chosen": -228.99447631835938,
+      "logps/rejected": -387.5201721191406,
+      "loss": 0.0771,
+      "rewards/chosen": 2.3790504932403564,
+      "rewards/margins": 13.84051513671875,
+      "rewards/rejected": -11.461464881896973,
+      "step": 1015
+    },
+    {
+      "epoch": 0.65,
+      "grad_norm": 15.932575225830078,
+      "kl": 0.0,
+      "learning_rate": 1.931009957325747e-07,
+      "logps/chosen": -278.6766357421875,
+      "logps/rejected": -372.94793701171875,
+      "loss": 0.0826,
+      "rewards/chosen": 2.363306760787964,
+      "rewards/margins": 14.281936645507812,
+      "rewards/rejected": -11.918628692626953,
+      "step": 1020
+    },
+    {
+      "epoch": 0.66,
+      "grad_norm": 11.954800605773926,
+      "kl": 0.0,
+      "learning_rate": 1.9132290184921761e-07,
+      "logps/chosen": -223.6442108154297,
+      "logps/rejected": -371.5616149902344,
+      "loss": 0.0961,
+      "rewards/chosen": 2.332761526107788,
+      "rewards/margins": 13.71965217590332,
+      "rewards/rejected": -11.386890411376953,
+      "step": 1025
+    },
+    {
+      "epoch": 0.66,
+      "grad_norm": 14.154836654663086,
+      "kl": 0.0,
+      "learning_rate": 1.895448079658606e-07,
+      "logps/chosen": -223.6517791748047,
+      "logps/rejected": -386.1749267578125,
+      "loss": 0.0908,
+      "rewards/chosen": 2.7226128578186035,
+      "rewards/margins": 14.438260078430176,
+      "rewards/rejected": -11.71564769744873,
+      "step": 1030
+    },
+    {
+      "epoch": 0.66,
+      "grad_norm": 18.134370803833008,
+      "kl": 0.0,
+      "learning_rate": 1.8776671408250356e-07,
+      "logps/chosen": -260.0550842285156,
+      "logps/rejected": -346.24127197265625,
+      "loss": 0.1007,
+      "rewards/chosen": 2.5339343547821045,
+      "rewards/margins": 13.723528861999512,
+      "rewards/rejected": -11.189595222473145,
+      "step": 1035
+    },
+    {
+      "epoch": 0.67,
+      "grad_norm": 15.416353225708008,
+      "kl": 0.0,
+      "learning_rate": 1.859886201991465e-07,
+      "logps/chosen": -247.7535400390625,
+      "logps/rejected": -371.42034912109375,
+      "loss": 0.0912,
+      "rewards/chosen": 2.3352103233337402,
+      "rewards/margins": 13.56823444366455,
+      "rewards/rejected": -11.233022689819336,
+      "step": 1040
+    },
+    {
+      "epoch": 0.67,
+      "grad_norm": 18.23054313659668,
+      "kl": 0.0,
+      "learning_rate": 1.8421052631578946e-07,
+      "logps/chosen": -213.8344268798828,
+      "logps/rejected": -369.88702392578125,
+      "loss": 0.0758,
+      "rewards/chosen": 2.6615304946899414,
+      "rewards/margins": 14.028742790222168,
+      "rewards/rejected": -11.367212295532227,
+      "step": 1045
+    },
+    {
+      "epoch": 0.67,
+      "grad_norm": 18.42432975769043,
+      "kl": 0.0,
+      "learning_rate": 1.8243243243243243e-07,
+      "logps/chosen": -238.5988006591797,
+      "logps/rejected": -392.305419921875,
+      "loss": 0.0965,
+      "rewards/chosen": 2.5309016704559326,
+      "rewards/margins": 14.56842041015625,
+      "rewards/rejected": -12.037518501281738,
+      "step": 1050
+    },
+    {
+      "epoch": 0.67,
+      "grad_norm": 15.395295143127441,
+      "kl": 0.0,
+      "learning_rate": 1.8065433854907538e-07,
+      "logps/chosen": -242.9210968017578,
+      "logps/rejected": -375.423828125,
+      "loss": 0.0985,
+      "rewards/chosen": 2.644166946411133,
+      "rewards/margins": 13.905471801757812,
+      "rewards/rejected": -11.26130485534668,
+      "step": 1055
+    },
+    {
+      "epoch": 0.68,
+      "grad_norm": 16.343006134033203,
+      "kl": 0.0,
+      "learning_rate": 1.7887624466571835e-07,
+      "logps/chosen": -198.0528106689453,
+      "logps/rejected": -354.1863098144531,
+      "loss": 0.092,
+      "rewards/chosen": 2.4538371562957764,
+      "rewards/margins": 12.678072929382324,
+      "rewards/rejected": -10.224235534667969,
+      "step": 1060
+    },
+    {
+      "epoch": 0.68,
+      "grad_norm": 19.673669815063477,
+      "kl": 0.0,
+      "learning_rate": 1.770981507823613e-07,
+      "logps/chosen": -241.1757049560547,
+      "logps/rejected": -339.1280822753906,
+      "loss": 0.088,
+      "rewards/chosen": 2.550891876220703,
+      "rewards/margins": 13.831698417663574,
+      "rewards/rejected": -11.280807495117188,
+      "step": 1065
+    },
+    {
+      "epoch": 0.68,
+      "grad_norm": 15.624777793884277,
+      "kl": 0.0,
+      "learning_rate": 1.7532005689900424e-07,
+      "logps/chosen": -247.8914031982422,
+      "logps/rejected": -364.5996398925781,
+      "loss": 0.1038,
+      "rewards/chosen": 2.5451152324676514,
+      "rewards/margins": 13.162847518920898,
+      "rewards/rejected": -10.617732048034668,
+      "step": 1070
+    },
+    {
+      "epoch": 0.69,
+      "grad_norm": 16.0270938873291,
+      "kl": 0.0,
+      "learning_rate": 1.7354196301564722e-07,
+      "logps/chosen": -271.2415466308594,
+      "logps/rejected": -348.6362609863281,
+      "loss": 0.0832,
+      "rewards/chosen": 2.6901183128356934,
+      "rewards/margins": 13.692281723022461,
+      "rewards/rejected": -11.002164840698242,
+      "step": 1075
+    },
+    {
+      "epoch": 0.69,
+      "grad_norm": 16.198070526123047,
+      "kl": 0.0,
+      "learning_rate": 1.717638691322902e-07,
+      "logps/chosen": -228.4770965576172,
+      "logps/rejected": -371.384521484375,
+      "loss": 0.0879,
+      "rewards/chosen": 2.317864179611206,
+      "rewards/margins": 14.147501945495605,
+      "rewards/rejected": -11.82963752746582,
+      "step": 1080
+    },
+    {
+      "epoch": 0.69,
+      "grad_norm": 24.463367462158203,
+      "kl": 0.0,
+      "learning_rate": 1.6998577524893314e-07,
+      "logps/chosen": -271.0051574707031,
+      "logps/rejected": -403.8445129394531,
+      "loss": 0.0817,
+      "rewards/chosen": 2.82243275642395,
+      "rewards/margins": 14.68195629119873,
+      "rewards/rejected": -11.859524726867676,
+      "step": 1085
+    },
+    {
+      "epoch": 0.7,
+      "grad_norm": 21.81173324584961,
+      "kl": 0.0,
+      "learning_rate": 1.6820768136557609e-07,
+      "logps/chosen": -209.64572143554688,
+      "logps/rejected": -360.5612487792969,
+      "loss": 0.0849,
+      "rewards/chosen": 2.4915404319763184,
+      "rewards/margins": 13.439886093139648,
+      "rewards/rejected": -10.948348045349121,
+      "step": 1090
+    },
+    {
+      "epoch": 0.7,
+      "grad_norm": 20.325061798095703,
+      "kl": 0.0,
+      "learning_rate": 1.6642958748221906e-07,
+      "logps/chosen": -244.2580108642578,
+      "logps/rejected": -348.1650390625,
+      "loss": 0.0804,
+      "rewards/chosen": 2.744297742843628,
+      "rewards/margins": 13.572134017944336,
+      "rewards/rejected": -10.827836990356445,
+      "step": 1095
+    },
+    {
+      "epoch": 0.7,
+      "grad_norm": 24.380229949951172,
+      "kl": 0.0,
+      "learning_rate": 1.64651493598862e-07,
+      "logps/chosen": -253.1858673095703,
+      "logps/rejected": -348.04144287109375,
+      "loss": 0.0846,
+      "rewards/chosen": 2.656752824783325,
+      "rewards/margins": 13.392297744750977,
+      "rewards/rejected": -10.735544204711914,
+      "step": 1100
+    },
+    {
+      "epoch": 0.71,
+      "grad_norm": 12.965998649597168,
+      "kl": 0.0,
+      "learning_rate": 1.6287339971550498e-07,
+      "logps/chosen": -262.8606262207031,
+      "logps/rejected": -356.92242431640625,
+      "loss": 0.0783,
+      "rewards/chosen": 2.8926401138305664,
+      "rewards/margins": 14.1354341506958,
+      "rewards/rejected": -11.242793083190918,
+      "step": 1105
+    },
+    {
+      "epoch": 0.71,
+      "grad_norm": 16.383338928222656,
+      "kl": 0.0,
+      "learning_rate": 1.6109530583214793e-07,
+      "logps/chosen": -300.94384765625,
+      "logps/rejected": -358.71588134765625,
+      "loss": 0.0951,
+      "rewards/chosen": 2.539355993270874,
+      "rewards/margins": 13.336338996887207,
+      "rewards/rejected": -10.79698371887207,
+      "step": 1110
+    },
+    {
+      "epoch": 0.71,
+      "grad_norm": 15.756513595581055,
+      "kl": 0.0,
+      "learning_rate": 1.5931721194879087e-07,
+      "logps/chosen": -261.84637451171875,
+      "logps/rejected": -374.3198547363281,
+      "loss": 0.0791,
+      "rewards/chosen": 2.755915880203247,
+      "rewards/margins": 14.211552619934082,
+      "rewards/rejected": -11.455635070800781,
+      "step": 1115
+    },
+    {
+      "epoch": 0.72,
+      "grad_norm": 18.087120056152344,
+      "kl": 0.0,
+      "learning_rate": 1.5753911806543385e-07,
+      "logps/chosen": -274.73370361328125,
+      "logps/rejected": -326.76123046875,
+      "loss": 0.0935,
+      "rewards/chosen": 2.786874294281006,
+      "rewards/margins": 12.87488079071045,
+      "rewards/rejected": -10.088006973266602,
+      "step": 1120
+    },
+    {
+      "epoch": 0.72,
+      "grad_norm": 15.65069580078125,
+      "kl": 0.0,
+      "learning_rate": 1.5576102418207682e-07,
+      "logps/chosen": -235.80062866210938,
+      "logps/rejected": -340.6771545410156,
+      "loss": 0.1005,
+      "rewards/chosen": 2.691729784011841,
+      "rewards/margins": 13.200857162475586,
+      "rewards/rejected": -10.509127616882324,
+      "step": 1125
+    },
+    {
+      "epoch": 0.72,
+      "grad_norm": 25.325843811035156,
+      "kl": 0.0,
+      "learning_rate": 1.5398293029871974e-07,
+      "logps/chosen": -244.3973388671875,
+      "logps/rejected": -343.04412841796875,
+      "loss": 0.0883,
+      "rewards/chosen": 2.485719680786133,
+      "rewards/margins": 13.161664962768555,
+      "rewards/rejected": -10.675946235656738,
+      "step": 1130
+    },
+    {
+      "epoch": 0.73,
+      "grad_norm": 17.044322967529297,
+      "kl": 0.0,
+      "learning_rate": 1.5220483641536272e-07,
+      "logps/chosen": -244.70559692382812,
+      "logps/rejected": -358.86920166015625,
+      "loss": 0.0987,
+      "rewards/chosen": 2.5949511528015137,
+      "rewards/margins": 13.485272407531738,
+      "rewards/rejected": -10.890320777893066,
+      "step": 1135
+    },
+    {
+      "epoch": 0.73,
+      "grad_norm": 27.13024139404297,
+      "kl": 0.0,
+      "learning_rate": 1.504267425320057e-07,
+      "logps/chosen": -247.91641235351562,
+      "logps/rejected": -343.4930419921875,
+      "loss": 0.0899,
+      "rewards/chosen": 2.4175989627838135,
+      "rewards/margins": 13.832735061645508,
+      "rewards/rejected": -11.415135383605957,
+      "step": 1140
+    },
+    {
+      "epoch": 0.73,
+      "grad_norm": 14.320551872253418,
+      "kl": 0.0,
+      "learning_rate": 1.4864864864864866e-07,
+      "logps/chosen": -246.96670532226562,
+      "logps/rejected": -370.0877990722656,
+      "loss": 0.0758,
+      "rewards/chosen": 2.691598653793335,
+      "rewards/margins": 13.812780380249023,
+      "rewards/rejected": -11.121182441711426,
+      "step": 1145
+    },
+    {
+      "epoch": 0.74,
+      "grad_norm": 16.582712173461914,
+      "kl": 0.0,
+      "learning_rate": 1.4687055476529158e-07,
+      "logps/chosen": -224.07785034179688,
+      "logps/rejected": -349.31927490234375,
+      "loss": 0.082,
+      "rewards/chosen": 2.6934258937835693,
+      "rewards/margins": 13.236696243286133,
+      "rewards/rejected": -10.543269157409668,
+      "step": 1150
+    },
+    {
+      "epoch": 0.74,
+      "grad_norm": 17.989919662475586,
+      "kl": 0.0,
+      "learning_rate": 1.4509246088193456e-07,
+      "logps/chosen": -243.8923797607422,
+      "logps/rejected": -367.2593688964844,
+      "loss": 0.0846,
+      "rewards/chosen": 2.7338271141052246,
+      "rewards/margins": 14.203822135925293,
+      "rewards/rejected": -11.469995498657227,
+      "step": 1155
+    },
+    {
+      "epoch": 0.74,
+      "grad_norm": 17.582975387573242,
+      "kl": 0.0,
+      "learning_rate": 1.4331436699857753e-07,
+      "logps/chosen": -227.1583709716797,
+      "logps/rejected": -378.5196228027344,
+      "loss": 0.1011,
+      "rewards/chosen": 2.531038284301758,
+      "rewards/margins": 13.578967094421387,
+      "rewards/rejected": -11.047929763793945,
+      "step": 1160
+    },
+    {
+      "epoch": 0.75,
+      "grad_norm": 8.821680068969727,
+      "kl": 0.0,
+      "learning_rate": 1.4153627311522048e-07,
+      "logps/chosen": -277.82684326171875,
+      "logps/rejected": -364.6512145996094,
+      "loss": 0.0742,
+      "rewards/chosen": 2.9555764198303223,
+      "rewards/margins": 14.14686393737793,
+      "rewards/rejected": -11.19128704071045,
+      "step": 1165
+    },
+    {
+      "epoch": 0.75,
+      "grad_norm": 15.817394256591797,
+      "kl": 0.0,
+      "learning_rate": 1.3975817923186345e-07,
+      "logps/chosen": -271.718017578125,
+      "logps/rejected": -352.4165954589844,
+      "loss": 0.0848,
+      "rewards/chosen": 2.5184245109558105,
+      "rewards/margins": 13.12096881866455,
+      "rewards/rejected": -10.602544784545898,
+      "step": 1170
+    },
+    {
+      "epoch": 0.75,
+      "grad_norm": 18.2441463470459,
+      "kl": 0.0,
+      "learning_rate": 1.379800853485064e-07,
+      "logps/chosen": -232.19467163085938,
+      "logps/rejected": -379.3485412597656,
+      "loss": 0.0909,
+      "rewards/chosen": 2.5335030555725098,
+      "rewards/margins": 14.489652633666992,
+      "rewards/rejected": -11.956149101257324,
+      "step": 1175
+    },
+    {
+      "epoch": 0.75,
+      "grad_norm": 30.594867706298828,
+      "kl": 0.0,
+      "learning_rate": 1.3620199146514935e-07,
+      "logps/chosen": -265.6886291503906,
+      "logps/rejected": -361.285400390625,
+      "loss": 0.0964,
+      "rewards/chosen": 2.450854778289795,
+      "rewards/margins": 13.442533493041992,
+      "rewards/rejected": -10.991681098937988,
+      "step": 1180
+    },
+    {
+      "epoch": 0.76,
+      "grad_norm": 12.528332710266113,
+      "kl": 0.0,
+      "learning_rate": 1.3442389758179232e-07,
+      "logps/chosen": -288.8167724609375,
+      "logps/rejected": -352.4346618652344,
+      "loss": 0.0741,
+      "rewards/chosen": 2.7889742851257324,
+      "rewards/margins": 14.034700393676758,
+      "rewards/rejected": -11.245725631713867,
+      "step": 1185
+    },
+    {
+      "epoch": 0.76,
+      "grad_norm": 13.565973281860352,
+      "kl": 0.0,
+      "learning_rate": 1.326458036984353e-07,
+      "logps/chosen": -263.8364562988281,
+      "logps/rejected": -346.26019287109375,
+      "loss": 0.0843,
+      "rewards/chosen": 2.638777256011963,
+      "rewards/margins": 13.351663589477539,
+      "rewards/rejected": -10.712886810302734,
+      "step": 1190
+    },
+    {
+      "epoch": 0.76,
+      "grad_norm": 21.124759674072266,
+      "kl": 0.0,
+      "learning_rate": 1.3086770981507821e-07,
+      "logps/chosen": -245.15576171875,
+      "logps/rejected": -372.0450134277344,
+      "loss": 0.0898,
+      "rewards/chosen": 2.657459259033203,
+      "rewards/margins": 14.276753425598145,
+      "rewards/rejected": -11.619293212890625,
+      "step": 1195
+    },
+    {
+      "epoch": 0.77,
+      "grad_norm": 15.396940231323242,
+      "kl": 0.0,
+      "learning_rate": 1.290896159317212e-07,
+      "logps/chosen": -247.9556121826172,
+      "logps/rejected": -371.3470764160156,
+      "loss": 0.0828,
+      "rewards/chosen": 2.4249672889709473,
+      "rewards/margins": 14.25316333770752,
+      "rewards/rejected": -11.828195571899414,
+      "step": 1200
+    },
+    {
+      "epoch": 0.77,
+      "grad_norm": 14.409557342529297,
+      "kl": 0.0,
+      "learning_rate": 1.2731152204836416e-07,
+      "logps/chosen": -252.07241821289062,
+      "logps/rejected": -361.1800842285156,
+      "loss": 0.085,
+      "rewards/chosen": 2.680985689163208,
+      "rewards/margins": 14.30175495147705,
+      "rewards/rejected": -11.620769500732422,
+      "step": 1205
+    },
+    {
+      "epoch": 0.77,
+      "grad_norm": 13.202485084533691,
+      "kl": 0.0,
+      "learning_rate": 1.255334281650071e-07,
+      "logps/chosen": -237.88565063476562,
+      "logps/rejected": -350.7021179199219,
+      "loss": 0.0874,
+      "rewards/chosen": 2.6545567512512207,
+      "rewards/margins": 13.95142936706543,
+      "rewards/rejected": -11.29687213897705,
+      "step": 1210
+    },
+    {
+      "epoch": 0.78,
+      "grad_norm": 28.25654411315918,
+      "kl": 0.0,
+      "learning_rate": 1.2375533428165005e-07,
+      "logps/chosen": -282.26031494140625,
+      "logps/rejected": -377.5342712402344,
+      "loss": 0.0771,
+      "rewards/chosen": 2.6503818035125732,
+      "rewards/margins": 14.911274909973145,
+      "rewards/rejected": -12.260892868041992,
+      "step": 1215
+    },
+    {
+      "epoch": 0.78,
+      "grad_norm": 20.115049362182617,
+      "kl": 0.0,
+      "learning_rate": 1.2197724039829303e-07,
+      "logps/chosen": -234.886962890625,
+      "logps/rejected": -347.6160583496094,
+      "loss": 0.0879,
+      "rewards/chosen": 2.644916296005249,
+      "rewards/margins": 14.264989852905273,
+      "rewards/rejected": -11.620074272155762,
+      "step": 1220
+    },
+    {
+      "epoch": 0.78,
+      "grad_norm": 17.797800064086914,
+      "kl": 0.0,
+      "learning_rate": 1.2019914651493598e-07,
+      "logps/chosen": -246.7455291748047,
+      "logps/rejected": -392.910400390625,
+      "loss": 0.0791,
+      "rewards/chosen": 2.706545352935791,
+      "rewards/margins": 14.843210220336914,
+      "rewards/rejected": -12.136663436889648,
+      "step": 1225
+    },
+    {
+      "epoch": 0.79,
+      "grad_norm": 12.662943840026855,
+      "kl": 0.0,
+      "learning_rate": 1.1842105263157894e-07,
+      "logps/chosen": -225.7235870361328,
+      "logps/rejected": -402.0716247558594,
+      "loss": 0.0925,
+      "rewards/chosen": 2.4518020153045654,
+      "rewards/margins": 14.207118034362793,
+      "rewards/rejected": -11.755315780639648,
+      "step": 1230
+    },
+    {
+      "epoch": 0.79,
+      "grad_norm": 16.056779861450195,
+      "kl": 0.0,
+      "learning_rate": 1.166429587482219e-07,
+      "logps/chosen": -215.66708374023438,
+      "logps/rejected": -375.41717529296875,
+      "loss": 0.0844,
+      "rewards/chosen": 2.730950355529785,
+      "rewards/margins": 14.23332691192627,
+      "rewards/rejected": -11.502375602722168,
+      "step": 1235
+    },
+    {
+      "epoch": 0.79,
+      "grad_norm": 16.644628524780273,
+      "kl": 0.0,
+      "learning_rate": 1.1486486486486487e-07,
+      "logps/chosen": -266.43878173828125,
+      "logps/rejected": -358.8226318359375,
+      "loss": 0.0835,
+      "rewards/chosen": 2.870779514312744,
+      "rewards/margins": 14.435707092285156,
+      "rewards/rejected": -11.56492805480957,
+      "step": 1240
+    },
+    {
+      "epoch": 0.8,
+      "grad_norm": 8.560157775878906,
+      "kl": 0.0,
+      "learning_rate": 1.1308677098150782e-07,
+      "logps/chosen": -238.55319213867188,
+      "logps/rejected": -390.98358154296875,
+      "loss": 0.0804,
+      "rewards/chosen": 2.759446620941162,
+      "rewards/margins": 14.843562126159668,
+      "rewards/rejected": -12.084115028381348,
+      "step": 1245
+    },
+    {
+      "epoch": 0.8,
+      "grad_norm": 16.91299819946289,
+      "kl": 0.0,
+      "learning_rate": 1.1130867709815078e-07,
+      "logps/chosen": -234.5216064453125,
+      "logps/rejected": -338.272705078125,
+      "loss": 0.093,
+      "rewards/chosen": 2.591742992401123,
+      "rewards/margins": 14.04884147644043,
+      "rewards/rejected": -11.457098007202148,
+      "step": 1250
+    },
+    {
+      "epoch": 0.8,
+      "grad_norm": 13.1578369140625,
+      "kl": 0.0,
+      "learning_rate": 1.0953058321479374e-07,
+      "logps/chosen": -275.4759826660156,
+      "logps/rejected": -379.9882507324219,
+      "loss": 0.0692,
+      "rewards/chosen": 2.984293222427368,
+      "rewards/margins": 15.574743270874023,
+      "rewards/rejected": -12.59045124053955,
+      "step": 1255
+    },
+    {
+      "epoch": 0.81,
+      "grad_norm": 17.815027236938477,
+      "kl": 0.0,
+      "learning_rate": 1.077524893314367e-07,
+      "logps/chosen": -267.3926696777344,
+      "logps/rejected": -353.8580017089844,
+      "loss": 0.0725,
+      "rewards/chosen": 2.5867929458618164,
+      "rewards/margins": 14.201835632324219,
+      "rewards/rejected": -11.615041732788086,
+      "step": 1260
+    },
+    {
+      "epoch": 0.81,
+      "grad_norm": 11.758225440979004,
+      "kl": 0.0,
+      "learning_rate": 1.0597439544807964e-07,
+      "logps/chosen": -249.6155242919922,
+      "logps/rejected": -364.9371337890625,
+      "loss": 0.0769,
+      "rewards/chosen": 2.7065796852111816,
+      "rewards/margins": 14.05157470703125,
+      "rewards/rejected": -11.344995498657227,
+      "step": 1265
+    },
+    {
+      "epoch": 0.81,
+      "grad_norm": 16.8466739654541,
+      "kl": 0.0,
+      "learning_rate": 1.0419630156472262e-07,
+      "logps/chosen": -238.05673217773438,
+      "logps/rejected": -373.03546142578125,
+      "loss": 0.0748,
+      "rewards/chosen": 2.9012789726257324,
+      "rewards/margins": 14.46104621887207,
+      "rewards/rejected": -11.55976676940918,
+      "step": 1270
+    },
+    {
+      "epoch": 0.82,
+      "grad_norm": 14.22836685180664,
+      "kl": 0.0,
+      "learning_rate": 1.0241820768136557e-07,
+      "logps/chosen": -224.74887084960938,
+      "logps/rejected": -332.44415283203125,
+      "loss": 0.084,
+      "rewards/chosen": 2.65814208984375,
+      "rewards/margins": 13.293182373046875,
+      "rewards/rejected": -10.635040283203125,
+      "step": 1275
+    },
+    {
+      "epoch": 0.82,
+      "grad_norm": 11.591891288757324,
+      "kl": 0.0,
+      "learning_rate": 1.0064011379800854e-07,
+      "logps/chosen": -280.44427490234375,
+      "logps/rejected": -365.83392333984375,
+      "loss": 0.0801,
+      "rewards/chosen": 2.5924744606018066,
+      "rewards/margins": 14.31025505065918,
+      "rewards/rejected": -11.717779159545898,
+      "step": 1280
+    },
+    {
+      "epoch": 0.82,
+      "grad_norm": 14.504942893981934,
+      "kl": 0.0,
+      "learning_rate": 9.886201991465149e-08,
+      "logps/chosen": -277.6451416015625,
+      "logps/rejected": -350.4718322753906,
+      "loss": 0.0661,
+      "rewards/chosen": 2.9364330768585205,
+      "rewards/margins": 14.681096076965332,
+      "rewards/rejected": -11.744662284851074,
+      "step": 1285
+    },
+    {
+      "epoch": 0.83,
+      "grad_norm": 16.065649032592773,
+      "kl": 0.0,
+      "learning_rate": 9.708392603129445e-08,
+      "logps/chosen": -262.8613586425781,
+      "logps/rejected": -391.71649169921875,
+      "loss": 0.0723,
+      "rewards/chosen": 2.7925260066986084,
+      "rewards/margins": 15.433810234069824,
+      "rewards/rejected": -12.641283988952637,
+      "step": 1290
+    },
+    {
+      "epoch": 0.83,
+      "grad_norm": 21.038543701171875,
+      "kl": 0.0,
+      "learning_rate": 9.530583214793741e-08,
+      "logps/chosen": -262.80584716796875,
+      "logps/rejected": -380.09625244140625,
+      "loss": 0.0801,
+      "rewards/chosen": 2.761460781097412,
+      "rewards/margins": 15.412538528442383,
+      "rewards/rejected": -12.651077270507812,
+      "step": 1295
+    },
+    {
+      "epoch": 0.83,
+      "grad_norm": 18.872102737426758,
+      "kl": 0.0,
+      "learning_rate": 9.352773826458037e-08,
+      "logps/chosen": -285.86712646484375,
+      "logps/rejected": -342.8421325683594,
+      "loss": 0.0819,
+      "rewards/chosen": 2.6993701457977295,
+      "rewards/margins": 14.52336311340332,
+      "rewards/rejected": -11.823991775512695,
+      "step": 1300
+    },
+    {
+      "epoch": 0.83,
+      "grad_norm": 15.029848098754883,
+      "kl": 0.0,
+      "learning_rate": 9.174964438122331e-08,
+      "logps/chosen": -245.3412322998047,
+      "logps/rejected": -356.58172607421875,
+      "loss": 0.094,
+      "rewards/chosen": 2.818472385406494,
+      "rewards/margins": 14.51873779296875,
+      "rewards/rejected": -11.700265884399414,
+      "step": 1305
+    },
+    {
+      "epoch": 0.84,
+      "grad_norm": 17.082958221435547,
+      "kl": 0.0,
+      "learning_rate": 8.997155049786629e-08,
+      "logps/chosen": -286.61383056640625,
+      "logps/rejected": -395.6304626464844,
+      "loss": 0.0716,
+      "rewards/chosen": 2.8008289337158203,
+      "rewards/margins": 15.597787857055664,
+      "rewards/rejected": -12.796960830688477,
+      "step": 1310
+    },
+    {
+      "epoch": 0.84,
+      "grad_norm": 22.768911361694336,
+      "kl": 0.0,
+      "learning_rate": 8.819345661450925e-08,
+      "logps/chosen": -250.6887664794922,
+      "logps/rejected": -358.8474426269531,
+      "loss": 0.0929,
+      "rewards/chosen": 2.7452361583709717,
+      "rewards/margins": 14.326945304870605,
+      "rewards/rejected": -11.581708908081055,
+      "step": 1315
+    },
+    {
+      "epoch": 0.84,
+      "grad_norm": 21.478580474853516,
+      "kl": 0.0,
+      "learning_rate": 8.64153627311522e-08,
+      "logps/chosen": -267.79974365234375,
+      "logps/rejected": -369.04815673828125,
+      "loss": 0.0838,
+      "rewards/chosen": 2.744828224182129,
+      "rewards/margins": 14.584070205688477,
+      "rewards/rejected": -11.839241027832031,
+      "step": 1320
+    },
+    {
+      "epoch": 0.85,
+      "grad_norm": 12.477672576904297,
+      "kl": 0.0,
+      "learning_rate": 8.463726884779517e-08,
+      "logps/chosen": -239.75399780273438,
+      "logps/rejected": -383.63812255859375,
+      "loss": 0.0833,
+      "rewards/chosen": 2.6762642860412598,
+      "rewards/margins": 14.272786140441895,
+      "rewards/rejected": -11.596521377563477,
+      "step": 1325
+    },
+    {
+      "epoch": 0.85,
+      "grad_norm": 19.415245056152344,
+      "kl": 0.0,
+      "learning_rate": 8.285917496443812e-08,
+      "logps/chosen": -258.97406005859375,
+      "logps/rejected": -372.96173095703125,
+      "loss": 0.0874,
+      "rewards/chosen": 2.7613649368286133,
+      "rewards/margins": 14.08574390411377,
+      "rewards/rejected": -11.324378967285156,
+      "step": 1330
+    },
+    {
+      "epoch": 0.85,
+      "grad_norm": 13.463936805725098,
+      "kl": 0.0,
+      "learning_rate": 8.108108108108108e-08,
+      "logps/chosen": -228.6225128173828,
+      "logps/rejected": -344.24688720703125,
+      "loss": 0.0777,
+      "rewards/chosen": 2.942438840866089,
+      "rewards/margins": 14.587237358093262,
+      "rewards/rejected": -11.644798278808594,
+      "step": 1335
+    },
+    {
+      "epoch": 0.86,
+      "grad_norm": 13.836956977844238,
+      "kl": 0.0,
+      "learning_rate": 7.930298719772404e-08,
+      "logps/chosen": -247.45742797851562,
+      "logps/rejected": -371.2969055175781,
+      "loss": 0.0721,
+      "rewards/chosen": 3.017321825027466,
+      "rewards/margins": 14.967196464538574,
+      "rewards/rejected": -11.949874877929688,
+      "step": 1340
+    },
+    {
+      "epoch": 0.86,
+      "grad_norm": 16.551321029663086,
+      "kl": 0.0,
+      "learning_rate": 7.7524893314367e-08,
+      "logps/chosen": -230.99227905273438,
+      "logps/rejected": -384.6443786621094,
+      "loss": 0.0871,
+      "rewards/chosen": 2.6670827865600586,
+      "rewards/margins": 14.841870307922363,
+      "rewards/rejected": -12.174787521362305,
+      "step": 1345
+    },
+    {
+      "epoch": 0.86,
+      "grad_norm": 15.87748908996582,
+      "kl": 0.0,
+      "learning_rate": 7.574679943100994e-08,
+      "logps/chosen": -266.7972717285156,
+      "logps/rejected": -335.5400390625,
+      "loss": 0.0976,
+      "rewards/chosen": 2.702658176422119,
+      "rewards/margins": 13.561399459838867,
+      "rewards/rejected": -10.858741760253906,
+      "step": 1350
+    },
+    {
+      "epoch": 0.87,
+      "grad_norm": 17.662395477294922,
+      "kl": 0.0,
+      "learning_rate": 7.396870554765292e-08,
+      "logps/chosen": -269.71026611328125,
+      "logps/rejected": -383.00628662109375,
+      "loss": 0.0921,
+      "rewards/chosen": 2.5814826488494873,
+      "rewards/margins": 14.34886646270752,
+      "rewards/rejected": -11.76738452911377,
+      "step": 1355
+    },
+    {
+      "epoch": 0.87,
+      "grad_norm": 16.09382438659668,
+      "kl": 0.0,
+      "learning_rate": 7.219061166429587e-08,
+      "logps/chosen": -269.3471984863281,
+      "logps/rejected": -375.32208251953125,
+      "loss": 0.0886,
+      "rewards/chosen": 2.8077282905578613,
+      "rewards/margins": 14.510396003723145,
+      "rewards/rejected": -11.702667236328125,
+      "step": 1360
+    },
+    {
+      "epoch": 0.87,
+      "grad_norm": 13.27104377746582,
+      "kl": 0.0,
+      "learning_rate": 7.041251778093883e-08,
+      "logps/chosen": -273.42950439453125,
+      "logps/rejected": -371.14837646484375,
+      "loss": 0.0685,
+      "rewards/chosen": 2.944091558456421,
+      "rewards/margins": 15.448992729187012,
+      "rewards/rejected": -12.504900932312012,
+      "step": 1365
+    },
+    {
+      "epoch": 0.88,
+      "grad_norm": 12.18476390838623,
+      "kl": 0.0,
+      "learning_rate": 6.863442389758179e-08,
+      "logps/chosen": -256.47161865234375,
+      "logps/rejected": -384.1866760253906,
+      "loss": 0.0742,
+      "rewards/chosen": 2.7861623764038086,
+      "rewards/margins": 15.34550952911377,
+      "rewards/rejected": -12.559347152709961,
+      "step": 1370
+    },
+    {
+      "epoch": 0.88,
+      "grad_norm": 23.00941276550293,
+      "kl": 0.0,
+      "learning_rate": 6.685633001422475e-08,
+      "logps/chosen": -242.3636474609375,
+      "logps/rejected": -404.00848388671875,
+      "loss": 0.0747,
+      "rewards/chosen": 2.5074126720428467,
+      "rewards/margins": 14.919418334960938,
+      "rewards/rejected": -12.412006378173828,
+      "step": 1375
+    },
+    {
+      "epoch": 0.88,
+      "grad_norm": 10.992572784423828,
+      "kl": 0.0,
+      "learning_rate": 6.507823613086771e-08,
+      "logps/chosen": -248.75967407226562,
+      "logps/rejected": -368.59063720703125,
+      "loss": 0.0867,
+      "rewards/chosen": 2.401717185974121,
+      "rewards/margins": 14.726252555847168,
+      "rewards/rejected": -12.32453441619873,
+      "step": 1380
+    },
+    {
+      "epoch": 0.89,
+      "grad_norm": 11.035951614379883,
+      "kl": 0.0,
+      "learning_rate": 6.330014224751067e-08,
+      "logps/chosen": -240.3590850830078,
+      "logps/rejected": -347.6131896972656,
+      "loss": 0.0921,
+      "rewards/chosen": 2.680974245071411,
+      "rewards/margins": 14.051698684692383,
+      "rewards/rejected": -11.370722770690918,
+      "step": 1385
+    },
+    {
+      "epoch": 0.89,
+      "grad_norm": 21.619415283203125,
+      "kl": 0.0,
+      "learning_rate": 6.152204836415363e-08,
+      "logps/chosen": -240.7654266357422,
+      "logps/rejected": -381.09307861328125,
+      "loss": 0.087,
+      "rewards/chosen": 2.646531105041504,
+      "rewards/margins": 14.633687019348145,
+      "rewards/rejected": -11.987154960632324,
+      "step": 1390
+    },
+    {
+      "epoch": 0.89,
+      "grad_norm": 20.996952056884766,
+      "kl": 0.0,
+      "learning_rate": 5.974395448079659e-08,
+      "logps/chosen": -228.12826538085938,
+      "logps/rejected": -373.8050231933594,
+      "loss": 0.0903,
+      "rewards/chosen": 2.5932085514068604,
+      "rewards/margins": 14.422918319702148,
+      "rewards/rejected": -11.829710960388184,
+      "step": 1395
+    },
+    {
+      "epoch": 0.9,
+      "grad_norm": 12.941699981689453,
+      "kl": 0.0,
+      "learning_rate": 5.796586059743954e-08,
+      "logps/chosen": -266.34759521484375,
+      "logps/rejected": -377.67572021484375,
+      "loss": 0.0733,
+      "rewards/chosen": 2.6969380378723145,
+      "rewards/margins": 14.922724723815918,
+      "rewards/rejected": -12.225786209106445,
+      "step": 1400
+    },
+    {
+      "epoch": 0.9,
+      "grad_norm": 18.563539505004883,
+      "kl": 0.0,
+      "learning_rate": 5.61877667140825e-08,
+      "logps/chosen": -233.05337524414062,
+      "logps/rejected": -351.65478515625,
+      "loss": 0.0819,
+      "rewards/chosen": 2.4380977153778076,
+      "rewards/margins": 14.38970947265625,
+      "rewards/rejected": -11.951611518859863,
+      "step": 1405
+    },
+    {
+      "epoch": 0.9,
+      "grad_norm": 14.77774715423584,
+      "kl": 0.0,
+      "learning_rate": 5.4409672830725456e-08,
+      "logps/chosen": -272.6517028808594,
+      "logps/rejected": -359.89013671875,
+      "loss": 0.0762,
+      "rewards/chosen": 2.888676404953003,
+      "rewards/margins": 14.460576057434082,
+      "rewards/rejected": -11.571900367736816,
+      "step": 1410
+    },
+    {
+      "epoch": 0.91,
+      "grad_norm": 15.590229988098145,
+      "kl": 0.0,
+      "learning_rate": 5.2631578947368416e-08,
+      "logps/chosen": -256.3451843261719,
+      "logps/rejected": -375.3144226074219,
+      "loss": 0.0762,
+      "rewards/chosen": 2.667102336883545,
+      "rewards/margins": 15.10230827331543,
+      "rewards/rejected": -12.435206413269043,
+      "step": 1415
+    },
+    {
+      "epoch": 0.91,
+      "grad_norm": 10.8274507522583,
+      "kl": 0.0,
+      "learning_rate": 5.0853485064011376e-08,
+      "logps/chosen": -268.83636474609375,
+      "logps/rejected": -372.27838134765625,
+      "loss": 0.0648,
+      "rewards/chosen": 2.9176650047302246,
+      "rewards/margins": 14.840046882629395,
+      "rewards/rejected": -11.922381401062012,
+      "step": 1420
+    },
+    {
+      "epoch": 0.91,
+      "grad_norm": 15.774340629577637,
+      "kl": 0.0,
+      "learning_rate": 4.9075391180654337e-08,
+      "logps/chosen": -260.6099853515625,
+      "logps/rejected": -358.04241943359375,
+      "loss": 0.0585,
+      "rewards/chosen": 2.9233345985412598,
+      "rewards/margins": 14.653867721557617,
+      "rewards/rejected": -11.730535507202148,
+      "step": 1425
+    },
+    {
+      "epoch": 0.91,
+      "grad_norm": 15.554561614990234,
+      "kl": 0.0,
+      "learning_rate": 4.72972972972973e-08,
+      "logps/chosen": -238.23739624023438,
+      "logps/rejected": -386.2373962402344,
+      "loss": 0.0833,
+      "rewards/chosen": 2.658268451690674,
+      "rewards/margins": 14.15376091003418,
+      "rewards/rejected": -11.49549388885498,
+      "step": 1430
+    },
+    {
+      "epoch": 0.92,
+      "grad_norm": 17.29914093017578,
+      "kl": 0.0,
+      "learning_rate": 4.551920341394026e-08,
+      "logps/chosen": -248.63204956054688,
+      "logps/rejected": -371.215087890625,
+      "loss": 0.0872,
+      "rewards/chosen": 2.9267373085021973,
+      "rewards/margins": 14.192481994628906,
+      "rewards/rejected": -11.265745162963867,
+      "step": 1435
+    },
+    {
+      "epoch": 0.92,
+      "grad_norm": 14.301885604858398,
+      "kl": 0.0,
+      "learning_rate": 4.374110953058322e-08,
+      "logps/chosen": -240.0107421875,
+      "logps/rejected": -355.2042541503906,
+      "loss": 0.0846,
+      "rewards/chosen": 2.4140007495880127,
+      "rewards/margins": 13.864062309265137,
+      "rewards/rejected": -11.450060844421387,
+      "step": 1440
+    },
+    {
+      "epoch": 0.92,
+      "grad_norm": 13.32541275024414,
+      "kl": 0.0,
+      "learning_rate": 4.196301564722617e-08,
+      "logps/chosen": -247.22152709960938,
+      "logps/rejected": -360.07110595703125,
+      "loss": 0.0659,
+      "rewards/chosen": 2.7749335765838623,
+      "rewards/margins": 14.867749214172363,
+      "rewards/rejected": -12.092815399169922,
+      "step": 1445
+    },
+    {
+      "epoch": 0.93,
+      "grad_norm": 12.356550216674805,
+      "kl": 0.0,
+      "learning_rate": 4.018492176386913e-08,
+      "logps/chosen": -256.69683837890625,
+      "logps/rejected": -363.08489990234375,
+      "loss": 0.0799,
+      "rewards/chosen": 2.8861594200134277,
+      "rewards/margins": 14.505203247070312,
+      "rewards/rejected": -11.619044303894043,
+      "step": 1450
+    },
+    {
+      "epoch": 0.93,
+      "grad_norm": 14.969185829162598,
+      "kl": 0.0,
+      "learning_rate": 3.840682788051209e-08,
+      "logps/chosen": -254.7665252685547,
+      "logps/rejected": -354.30841064453125,
+      "loss": 0.0757,
+      "rewards/chosen": 2.8421378135681152,
+      "rewards/margins": 14.373845100402832,
+      "rewards/rejected": -11.531707763671875,
+      "step": 1455
+    },
+    {
+      "epoch": 0.93,
+      "grad_norm": 17.030691146850586,
+      "kl": 0.0,
+      "learning_rate": 3.6628733997155046e-08,
+      "logps/chosen": -263.8084411621094,
+      "logps/rejected": -395.4825439453125,
+      "loss": 0.0759,
+      "rewards/chosen": 2.9121432304382324,
+      "rewards/margins": 15.505640983581543,
+      "rewards/rejected": -12.593496322631836,
+      "step": 1460
+    },
+    {
+      "epoch": 0.94,
+      "grad_norm": 12.211559295654297,
+      "kl": 0.0,
+      "learning_rate": 3.4850640113798006e-08,
+      "logps/chosen": -236.26431274414062,
+      "logps/rejected": -357.1573791503906,
+      "loss": 0.0755,
+      "rewards/chosen": 2.843881130218506,
+      "rewards/margins": 14.507527351379395,
+      "rewards/rejected": -11.66364574432373,
+      "step": 1465
+    },
+    {
+      "epoch": 0.94,
+      "grad_norm": 17.574806213378906,
+      "kl": 0.0,
+      "learning_rate": 3.3072546230440967e-08,
+      "logps/chosen": -285.59942626953125,
+      "logps/rejected": -368.08245849609375,
+      "loss": 0.0676,
+      "rewards/chosen": 2.877531051635742,
+      "rewards/margins": 14.671765327453613,
+      "rewards/rejected": -11.794233322143555,
+      "step": 1470
+    },
+    {
+      "epoch": 0.94,
+      "grad_norm": 13.167614936828613,
+      "kl": 0.0,
+      "learning_rate": 3.129445234708392e-08,
+      "logps/chosen": -240.7215576171875,
+      "logps/rejected": -382.0995788574219,
+      "loss": 0.0651,
+      "rewards/chosen": 2.778978109359741,
+      "rewards/margins": 14.833264350891113,
+      "rewards/rejected": -12.054286003112793,
+      "step": 1475
+    },
+    {
+      "epoch": 0.95,
+      "grad_norm": 19.462112426757812,
+      "kl": 0.0,
+      "learning_rate": 2.9516358463726884e-08,
+      "logps/chosen": -243.0684356689453,
+      "logps/rejected": -366.4331359863281,
+      "loss": 0.0802,
+      "rewards/chosen": 2.8256452083587646,
+      "rewards/margins": 15.236946105957031,
+      "rewards/rejected": -12.411300659179688,
+      "step": 1480
+    },
+    {
+      "epoch": 0.95,
+      "grad_norm": 20.093833923339844,
+      "kl": 0.0,
+      "learning_rate": 2.7738264580369844e-08,
+      "logps/chosen": -209.17770385742188,
+      "logps/rejected": -359.9656677246094,
+      "loss": 0.0839,
+      "rewards/chosen": 2.901711940765381,
+      "rewards/margins": 14.679159164428711,
+      "rewards/rejected": -11.777449607849121,
+      "step": 1485
+    },
+    {
+      "epoch": 0.95,
+      "grad_norm": 13.270203590393066,
+      "kl": 0.0,
+      "learning_rate": 2.59601706970128e-08,
+      "logps/chosen": -272.44610595703125,
+      "logps/rejected": -357.05975341796875,
+      "loss": 0.0868,
+      "rewards/chosen": 2.6972391605377197,
+      "rewards/margins": 15.144004821777344,
+      "rewards/rejected": -12.446764945983887,
+      "step": 1490
+    },
+    {
+      "epoch": 0.96,
+      "grad_norm": 14.922286987304688,
+      "kl": 0.0,
+      "learning_rate": 2.418207681365576e-08,
+      "logps/chosen": -238.18191528320312,
+      "logps/rejected": -367.4979553222656,
+      "loss": 0.0824,
+      "rewards/chosen": 2.5710983276367188,
+      "rewards/margins": 14.130340576171875,
+      "rewards/rejected": -11.559242248535156,
+      "step": 1495
+    },
+    {
+      "epoch": 0.96,
+      "grad_norm": 21.186229705810547,
+      "kl": 0.0,
+      "learning_rate": 2.240398293029872e-08,
+      "logps/chosen": -240.6565399169922,
+      "logps/rejected": -387.65936279296875,
+      "loss": 0.0783,
+      "rewards/chosen": 2.6365768909454346,
+      "rewards/margins": 15.131543159484863,
+      "rewards/rejected": -12.494966506958008,
+      "step": 1500
+    },
+    {
+      "epoch": 0.96,
+      "grad_norm": 14.05397891998291,
+      "kl": 0.0,
+      "learning_rate": 2.0625889046941676e-08,
+      "logps/chosen": -307.04583740234375,
+      "logps/rejected": -361.1058654785156,
+      "loss": 0.0903,
+      "rewards/chosen": 2.804368257522583,
+      "rewards/margins": 14.683168411254883,
+      "rewards/rejected": -11.878799438476562,
+      "step": 1505
+    },
+    {
+      "epoch": 0.97,
+      "grad_norm": 10.290091514587402,
+      "kl": 0.0,
+      "learning_rate": 1.8847795163584636e-08,
+      "logps/chosen": -280.72406005859375,
+      "logps/rejected": -393.7200622558594,
+      "loss": 0.0643,
+      "rewards/chosen": 2.76891827583313,
+      "rewards/margins": 15.010998725891113,
+      "rewards/rejected": -12.24207878112793,
+      "step": 1510
+    },
+    {
+      "epoch": 0.97,
+      "grad_norm": 13.187542915344238,
+      "kl": 0.0,
+      "learning_rate": 1.7069701280227596e-08,
+      "logps/chosen": -250.7726287841797,
+      "logps/rejected": -378.99993896484375,
+      "loss": 0.0697,
+      "rewards/chosen": 2.8748562335968018,
+      "rewards/margins": 15.209932327270508,
+      "rewards/rejected": -12.335077285766602,
+      "step": 1515
+    },
+    {
+      "epoch": 0.97,
+      "grad_norm": 13.64420223236084,
+      "kl": 0.0,
+      "learning_rate": 1.5291607396870554e-08,
+      "logps/chosen": -253.7127227783203,
+      "logps/rejected": -367.41558837890625,
+      "loss": 0.0824,
+      "rewards/chosen": 3.031802177429199,
+      "rewards/margins": 14.762832641601562,
+      "rewards/rejected": -11.73102855682373,
+      "step": 1520
+    },
+    {
+      "epoch": 0.98,
+      "grad_norm": 10.866116523742676,
+      "kl": 0.0,
+      "learning_rate": 1.3513513513513514e-08,
+      "logps/chosen": -250.94400024414062,
+      "logps/rejected": -381.5174865722656,
+      "loss": 0.0793,
+      "rewards/chosen": 2.8337435722351074,
+      "rewards/margins": 15.124834060668945,
+      "rewards/rejected": -12.29109001159668,
+      "step": 1525
+    },
+    {
+      "epoch": 0.98,
+      "grad_norm": 11.604872703552246,
+      "kl": 0.0,
+      "learning_rate": 1.1735419630156473e-08,
+      "logps/chosen": -233.5682830810547,
+      "logps/rejected": -377.261962890625,
+      "loss": 0.0687,
+      "rewards/chosen": 3.119915008544922,
+      "rewards/margins": 15.134057998657227,
+      "rewards/rejected": -12.014142990112305,
+      "step": 1530
+    },
+    {
+      "epoch": 0.98,
+      "grad_norm": 20.006731033325195,
+      "kl": 0.0,
+      "learning_rate": 9.95732574679943e-09,
+      "logps/chosen": -257.07525634765625,
+      "logps/rejected": -346.5194396972656,
+      "loss": 0.0907,
+      "rewards/chosen": 2.676340103149414,
+      "rewards/margins": 14.137725830078125,
+      "rewards/rejected": -11.461384773254395,
+      "step": 1535
+    },
+    {
+      "epoch": 0.99,
+      "grad_norm": 17.203874588012695,
+      "kl": 0.0,
+      "learning_rate": 8.179231863442388e-09,
+      "logps/chosen": -226.61843872070312,
+      "logps/rejected": -381.11029052734375,
+      "loss": 0.0683,
+      "rewards/chosen": 2.940342903137207,
+      "rewards/margins": 15.23707103729248,
+      "rewards/rejected": -12.296728134155273,
+      "step": 1540
+    },
+    {
+      "epoch": 0.99,
+      "grad_norm": 14.477093696594238,
+      "kl": 0.0,
+      "learning_rate": 6.401137980085348e-09,
+      "logps/chosen": -240.14157104492188,
+      "logps/rejected": -388.9137268066406,
+      "loss": 0.0658,
+      "rewards/chosen": 2.875831127166748,
+      "rewards/margins": 15.294588088989258,
+      "rewards/rejected": -12.418756484985352,
+      "step": 1545
+    },
+    {
+      "epoch": 0.99,
+      "grad_norm": 10.444808959960938,
+      "kl": 0.0,
+      "learning_rate": 4.623044096728307e-09,
+      "logps/chosen": -273.6938781738281,
+      "logps/rejected": -361.58782958984375,
+      "loss": 0.0592,
+      "rewards/chosen": 3.0903258323669434,
+      "rewards/margins": 15.089597702026367,
+      "rewards/rejected": -11.999273300170898,
+      "step": 1550
+    },
+    {
+      "epoch": 0.99,
+      "grad_norm": 13.01125717163086,
+      "kl": 0.0,
+      "learning_rate": 2.844950213371266e-09,
+      "logps/chosen": -263.2748107910156,
+      "logps/rejected": -382.49481201171875,
+      "loss": 0.0724,
+      "rewards/chosen": 2.801457166671753,
+      "rewards/margins": 14.89887809753418,
+      "rewards/rejected": -12.097421646118164,
+      "step": 1555
+    },
+    {
+      "epoch": 1.0,
+      "grad_norm": 13.397442817687988,
+      "kl": 0.0,
+      "learning_rate": 1.0668563300142248e-09,
+      "logps/chosen": -225.07894897460938,
+      "logps/rejected": -353.08856201171875,
+      "loss": 0.0824,
+      "rewards/chosen": 2.47590970993042,
+      "rewards/margins": 14.8174409866333,
+      "rewards/rejected": -12.341531753540039,
+      "step": 1560
+    },
+    {
+      "epoch": 1.0,
+      "step": 1563,
+      "total_flos": 0.0,
+      "train_loss": 0.11715243643320149,
+      "train_runtime": 11089.6574,
+      "train_samples_per_second": 9.017,
+      "train_steps_per_second": 0.141
+    }
+  ],
+  "logging_steps": 5,
+  "max_steps": 1563,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 1,
+  "save_steps": 100,
+  "total_flos": 0.0,
+  "train_batch_size": 8,
+  "trial_name": null,
+  "trial_params": null
+}