{
  "best_global_step": null,
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.8228034087569791,
  "eval_steps": 200,
  "global_step": 2100,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.019590557351356647,
      "grad_norm": 140.09786987304688,
      "learning_rate": 3.2666666666666663e-07,
      "logits/chosen": -0.9822076559066772,
      "logits/rejected": -0.6774621605873108,
      "logps/chosen": -352.589111328125,
      "logps/rejected": -284.4425354003906,
      "loss": 0.6901,
      "rewards/accuracies": 0.47999998927116394,
      "rewards/chosen": 0.0031729438342154026,
      "rewards/margins": 0.008149052038788795,
      "rewards/rejected": -0.004976108204573393,
      "step": 50
    },
    {
      "epoch": 0.03918111470271329,
      "grad_norm": 128.5372314453125,
      "learning_rate": 6.6e-07,
      "logits/chosen": -0.9334858059883118,
      "logits/rejected": -0.6409775018692017,
      "logps/chosen": -366.2322692871094,
      "logps/rejected": -287.4626770019531,
      "loss": 0.6758,
      "rewards/accuracies": 0.6150000095367432,
      "rewards/chosen": 0.02407059632241726,
      "rewards/margins": 0.04141266271471977,
      "rewards/rejected": -0.017342066392302513,
      "step": 100
    },
    {
      "epoch": 0.05877167205406994,
      "grad_norm": 103.95256805419922,
      "learning_rate": 9.933333333333333e-07,
      "logits/chosen": -0.8153188228607178,
      "logits/rejected": -0.5176287293434143,
      "logps/chosen": -373.8503112792969,
      "logps/rejected": -307.275634765625,
      "loss": 0.6225,
      "rewards/accuracies": 0.6775000095367432,
      "rewards/chosen": 0.07690317183732986,
      "rewards/margins": 0.18566985428333282,
      "rewards/rejected": -0.10876669734716415,
      "step": 150
    },
    {
      "epoch": 0.07836222940542659,
      "grad_norm": 130.03802490234375,
      "learning_rate": 9.99894936347371e-07,
      "logits/chosen": -0.8271468877792358,
      "logits/rejected": -0.5104550719261169,
      "logps/chosen": -350.3367614746094,
      "logps/rejected": -296.0983581542969,
      "loss": 0.6152,
      "rewards/accuracies": 0.6775000095367432,
      "rewards/chosen": -0.0007632786291651428,
      "rewards/margins": 0.25347909331321716,
      "rewards/rejected": -0.2542423605918884,
      "step": 200
    },
    {
      "epoch": 0.07836222940542659,
      "eval_logits/chosen": -0.8247441053390503,
      "eval_logits/rejected": -0.5691510438919067,
      "eval_logps/chosen": -352.2530212402344,
      "eval_logps/rejected": -288.40521240234375,
      "eval_loss": 0.6011638641357422,
      "eval_rewards/accuracies": 0.684013307094574,
      "eval_rewards/chosen": 0.006044471170753241,
      "eval_rewards/margins": 0.30330440402030945,
      "eval_rewards/rejected": -0.2972599267959595,
      "eval_runtime": 180.5524,
      "eval_samples_per_second": 13.304,
      "eval_steps_per_second": 6.652,
      "step": 200
    },
    {
      "epoch": 0.09795278675678323,
      "grad_norm": 127.07317352294922,
      "learning_rate": 9.995711712979657e-07,
      "logits/chosen": -0.7803874015808105,
      "logits/rejected": -0.5503089427947998,
      "logps/chosen": -354.3157043457031,
      "logps/rejected": -271.8356018066406,
      "loss": 0.5935,
      "rewards/accuracies": 0.6825000047683716,
      "rewards/chosen": -0.0023533145431429148,
      "rewards/margins": 0.3317827880382538,
      "rewards/rejected": -0.33413612842559814,
      "step": 250
    },
    {
      "epoch": 0.11754334410813988,
      "grad_norm": 107.00403594970703,
      "learning_rate": 9.990288027658056e-07,
      "logits/chosen": -0.8059212565422058,
      "logits/rejected": -0.5893051028251648,
      "logps/chosen": -349.48760986328125,
      "logps/rejected": -285.4340515136719,
      "loss": 0.5528,
      "rewards/accuracies": 0.7599999904632568,
      "rewards/chosen": 0.04797028377652168,
      "rewards/margins": 0.4950464367866516,
      "rewards/rejected": -0.44707614183425903,
      "step": 300
    },
    {
      "epoch": 0.13713390145949653,
      "grad_norm": 118.62214660644531,
      "learning_rate": 9.982680680817391e-07,
      "logits/chosen": -0.9657462239265442,
      "logits/rejected": -0.647233784198761,
      "logps/chosen": -355.55096435546875,
      "logps/rejected": -308.36053466796875,
      "loss": 0.56,
      "rewards/accuracies": 0.7074999809265137,
      "rewards/chosen": -0.11357284337282181,
      "rewards/margins": 0.5093904137611389,
      "rewards/rejected": -0.6229632496833801,
      "step": 350
    },
    {
      "epoch": 0.15672445881085317,
      "grad_norm": 80.19744110107422,
      "learning_rate": 9.972893001297698e-07,
      "logits/chosen": -1.0753679275512695,
      "logits/rejected": -0.731191873550415,
      "logps/chosen": -374.8988952636719,
      "logps/rejected": -319.61199951171875,
      "loss": 0.5683,
      "rewards/accuracies": 0.7049999833106995,
      "rewards/chosen": -0.20595918595790863,
      "rewards/margins": 0.5294285416603088,
      "rewards/rejected": -0.7353877425193787,
      "step": 400
    },
    {
      "epoch": 0.15672445881085317,
      "eval_logits/chosen": -1.0395917892456055,
      "eval_logits/rejected": -0.7957448363304138,
      "eval_logps/chosen": -354.0768127441406,
      "eval_logps/rejected": -293.1625061035156,
      "eval_loss": 0.5456792116165161,
      "eval_rewards/accuracies": 0.7264779210090637,
      "eval_rewards/chosen": -0.17633533477783203,
      "eval_rewards/margins": 0.5966517329216003,
      "eval_rewards/rejected": -0.7729870080947876,
      "eval_runtime": 180.5278,
      "eval_samples_per_second": 13.305,
      "eval_steps_per_second": 6.653,
      "step": 400
    },
    {
      "epoch": 0.17631501616220982,
      "grad_norm": 97.5062026977539,
      "learning_rate": 9.96092927201391e-07,
      "logits/chosen": -1.0175271034240723,
      "logits/rejected": -0.752033531665802,
      "logps/chosen": -355.3705749511719,
      "logps/rejected": -302.32373046875,
      "loss": 0.5547,
      "rewards/accuracies": 0.7024999856948853,
      "rewards/chosen": -0.2221955806016922,
      "rewards/margins": 0.6058707237243652,
      "rewards/rejected": -0.828066349029541,
      "step": 450
    },
    {
      "epoch": 0.19590557351356647,
      "grad_norm": 109.52022552490234,
      "learning_rate": 9.946794728081737e-07,
      "logits/chosen": -1.2173391580581665,
      "logits/rejected": -0.8966682553291321,
      "logps/chosen": -345.02984619140625,
      "logps/rejected": -300.69268798828125,
      "loss": 0.5145,
      "rewards/accuracies": 0.7250000238418579,
      "rewards/chosen": -0.26011258363723755,
      "rewards/margins": 0.7154355049133301,
      "rewards/rejected": -0.9755480885505676,
      "step": 500
    },
    {
      "epoch": 0.2154961308649231,
      "grad_norm": 65.85977172851562,
      "learning_rate": 9.930495554526878e-07,
      "logits/chosen": -1.1911191940307617,
      "logits/rejected": -0.9048889875411987,
      "logps/chosen": -360.3658142089844,
      "logps/rejected": -297.0860900878906,
      "loss": 0.5025,
      "rewards/accuracies": 0.7275000214576721,
      "rewards/chosen": -0.28734278678894043,
      "rewards/margins": 0.8380499482154846,
      "rewards/rejected": -1.1253927946090698,
      "step": 550
    },
    {
      "epoch": 0.23508668821627976,
      "grad_norm": 63.237606048583984,
      "learning_rate": 9.912038883578552e-07,
      "logits/chosen": -1.0660635232925415,
      "logits/rejected": -0.7923431992530823,
      "logps/chosen": -363.8102722167969,
      "logps/rejected": -295.3084716796875,
      "loss": 0.5179,
      "rewards/accuracies": 0.7300000190734863,
      "rewards/chosen": -0.4248141050338745,
      "rewards/margins": 0.8246171474456787,
      "rewards/rejected": -1.2494312524795532,
      "step": 600
    },
    {
      "epoch": 0.23508668821627976,
      "eval_logits/chosen": -1.0540363788604736,
      "eval_logits/rejected": -0.8268531560897827,
      "eval_logps/chosen": -355.8159484863281,
      "eval_logps/rejected": -297.4617919921875,
      "eval_loss": 0.5122258067131042,
      "eval_rewards/accuracies": 0.7373022437095642,
      "eval_rewards/chosen": -0.35024747252464294,
      "eval_rewards/margins": 0.8526709675788879,
      "eval_rewards/rejected": -1.2029184103012085,
      "eval_runtime": 181.1919,
      "eval_samples_per_second": 13.257,
      "eval_steps_per_second": 6.628,
      "step": 600
    },
    {
      "epoch": 0.2546772455676364,
      "grad_norm": 194.335693359375,
      "learning_rate": 9.891432791548562e-07,
      "logits/chosen": -1.0396273136138916,
      "logits/rejected": -0.8581287860870361,
      "logps/chosen": -375.4624938964844,
      "logps/rejected": -302.572998046875,
      "loss": 0.4782,
      "rewards/accuracies": 0.7799999713897705,
      "rewards/chosen": -0.3973628580570221,
      "rewards/margins": 0.9747829437255859,
      "rewards/rejected": -1.3721458911895752,
      "step": 650
    },
    {
      "epoch": 0.27426780291899305,
      "grad_norm": 65.69367980957031,
      "learning_rate": 9.868686295297258e-07,
      "logits/chosen": -1.070594072341919,
      "logits/rejected": -0.848624050617218,
      "logps/chosen": -358.2480773925781,
      "logps/rejected": -311.5008850097656,
      "loss": 0.5109,
      "rewards/accuracies": 0.7475000023841858,
      "rewards/chosen": -0.465026319026947,
      "rewards/margins": 0.9383654594421387,
      "rewards/rejected": -1.4033918380737305,
      "step": 700
    },
    {
      "epoch": 0.29385836027034967,
      "grad_norm": 116.7066650390625,
      "learning_rate": 9.8438093482879e-07,
      "logits/chosen": -1.0724475383758545,
      "logits/rejected": -0.8224292993545532,
      "logps/chosen": -373.66754150390625,
      "logps/rejected": -297.4761047363281,
      "loss": 0.483,
      "rewards/accuracies": 0.7699999809265137,
      "rewards/chosen": -0.39762890338897705,
      "rewards/margins": 1.163187026977539,
      "rewards/rejected": -1.5608159303665161,
      "step": 750
    },
    {
      "epoch": 0.31344891762170635,
      "grad_norm": 124.66069793701172,
      "learning_rate": 9.81681283623121e-07,
      "logits/chosen": -1.0539674758911133,
      "logits/rejected": -0.779601514339447,
      "logps/chosen": -347.1493225097656,
      "logps/rejected": -300.728515625,
      "loss": 0.5427,
      "rewards/accuracies": 0.7250000238418579,
      "rewards/chosen": -0.48545289039611816,
      "rewards/margins": 0.8515650033950806,
      "rewards/rejected": -1.3370177745819092,
      "step": 800
    },
    {
      "epoch": 0.31344891762170635,
      "eval_logits/chosen": -0.9799102544784546,
      "eval_logits/rejected": -0.7651399970054626,
      "eval_logps/chosen": -356.1210021972656,
      "eval_logps/rejected": -299.1585388183594,
      "eval_loss": 0.4938255250453949,
      "eval_rewards/accuracies": 0.7493755221366882,
      "eval_rewards/chosen": -0.38075584173202515,
      "eval_rewards/margins": 0.9918379783630371,
      "eval_rewards/rejected": -1.372593641281128,
      "eval_runtime": 181.2826,
      "eval_samples_per_second": 13.25,
      "eval_steps_per_second": 6.625,
      "step": 800
    },
    {
      "epoch": 0.33303947497306297,
      "grad_norm": 120.1912612915039,
      "learning_rate": 9.787708572321983e-07,
      "logits/chosen": -1.034041404724121,
      "logits/rejected": -0.810564398765564,
      "logps/chosen": -367.707763671875,
      "logps/rejected": -309.762939453125,
      "loss": 0.4326,
      "rewards/accuracies": 0.8125,
      "rewards/chosen": -0.3861949145793915,
      "rewards/margins": 1.226701021194458,
      "rewards/rejected": -1.6128960847854614,
      "step": 850
    },
    {
      "epoch": 0.35263003232441964,
      "grad_norm": 79.74503326416016,
      "learning_rate": 9.756509292069825e-07,
      "logits/chosen": -1.125342845916748,
      "logits/rejected": -0.875031590461731,
      "logps/chosen": -357.8549499511719,
      "logps/rejected": -289.9258117675781,
      "loss": 0.4864,
      "rewards/accuracies": 0.762499988079071,
      "rewards/chosen": -0.5919533967971802,
      "rewards/margins": 1.2158567905426025,
      "rewards/rejected": -1.8078101873397827,
      "step": 900
    },
    {
      "epoch": 0.37222058967577626,
      "grad_norm": 136.9542236328125,
      "learning_rate": 9.72322864772634e-07,
      "logits/chosen": -1.1560072898864746,
      "logits/rejected": -0.886911928653717,
      "logps/chosen": -364.5694274902344,
      "logps/rejected": -293.49981689453125,
      "loss": 0.51,
      "rewards/accuracies": 0.7400000095367432,
      "rewards/chosen": -0.5705673098564148,
      "rewards/margins": 1.0413060188293457,
      "rewards/rejected": -1.6118732690811157,
      "step": 950
    },
    {
      "epoch": 0.39181114702713293,
      "grad_norm": 116.3714370727539,
      "learning_rate": 9.687881202311132e-07,
      "logits/chosen": -1.0079857110977173,
      "logits/rejected": -0.7752091884613037,
      "logps/chosen": -349.3457946777344,
      "logps/rejected": -286.99993896484375,
      "loss": 0.4722,
      "rewards/accuracies": 0.7749999761581421,
      "rewards/chosen": -0.4944779872894287,
      "rewards/margins": 1.1931304931640625,
      "rewards/rejected": -1.6876084804534912,
      "step": 1000
    },
    {
      "epoch": 0.39181114702713293,
      "eval_logits/chosen": -1.0761741399765015,
      "eval_logits/rejected": -0.8803514838218689,
      "eval_logps/chosen": -357.8018798828125,
      "eval_logps/rejected": -302.2410888671875,
      "eval_loss": 0.4866333603858948,
      "eval_rewards/accuracies": 0.7510408163070679,
      "eval_rewards/chosen": -0.5488451719284058,
      "eval_rewards/margins": 1.1320006847381592,
      "eval_rewards/rejected": -1.680845856666565,
      "eval_runtime": 181.5986,
      "eval_samples_per_second": 13.227,
      "eval_steps_per_second": 6.613,
      "step": 1000
    },
    {
      "epoch": 0.41140170437848955,
      "grad_norm": 169.1030731201172,
      "learning_rate": 9.65048242323929e-07,
      "logits/chosen": -1.0233768224716187,
      "logits/rejected": -0.740478515625,
      "logps/chosen": -380.3099365234375,
      "logps/rejected": -323.1108703613281,
      "loss": 0.4805,
      "rewards/accuracies": 0.7774999737739563,
      "rewards/chosen": -0.6764897704124451,
      "rewards/margins": 1.2643505334854126,
      "rewards/rejected": -1.940840482711792,
      "step": 1050
    },
    {
      "epoch": 0.4309922617298462,
      "grad_norm": 107.12033081054688,
      "learning_rate": 9.611048675553127e-07,
      "logits/chosen": -1.0799275636672974,
      "logits/rejected": -0.8886963129043579,
      "logps/chosen": -387.2933349609375,
      "logps/rejected": -321.2319030761719,
      "loss": 0.4268,
      "rewards/accuracies": 0.7774999737739563,
      "rewards/chosen": -0.5030026435852051,
      "rewards/margins": 1.4225863218307495,
      "rewards/rejected": -1.925588846206665,
      "step": 1100
    },
    {
      "epoch": 0.45058281908120285,
      "grad_norm": 124.23180389404297,
      "learning_rate": 9.569597214761124e-07,
      "logits/chosen": -1.0482264757156372,
      "logits/rejected": -0.8745830059051514,
      "logps/chosen": -352.1131591796875,
      "logps/rejected": -303.13592529296875,
      "loss": 0.4846,
      "rewards/accuracies": 0.7674999833106995,
      "rewards/chosen": -0.689760684967041,
      "rewards/margins": 1.3204187154769897,
      "rewards/rejected": -2.0101795196533203,
      "step": 1150
    },
    {
      "epoch": 0.4701733764325595,
      "grad_norm": 68.6223373413086,
      "learning_rate": 9.526146179287222e-07,
      "logits/chosen": -1.097909927368164,
      "logits/rejected": -0.8678469657897949,
      "logps/chosen": -358.7846984863281,
      "logps/rejected": -314.0942077636719,
      "loss": 0.4829,
      "rewards/accuracies": 0.7649999856948853,
      "rewards/chosen": -0.6658291816711426,
      "rewards/margins": 1.3254380226135254,
      "rewards/rejected": -1.991267204284668,
      "step": 1200
    },
    {
      "epoch": 0.4701733764325595,
      "eval_logits/chosen": -1.1058708429336548,
      "eval_logits/rejected": -0.9207807183265686,
      "eval_logps/chosen": -358.45928955078125,
      "eval_logps/rejected": -304.12744140625,
      "eval_loss": 0.4782937467098236,
      "eval_rewards/accuracies": 0.7585345506668091,
      "eval_rewards/chosen": -0.6145861148834229,
      "eval_rewards/margins": 1.2548983097076416,
      "eval_rewards/rejected": -1.8694841861724854,
      "eval_runtime": 181.5576,
      "eval_samples_per_second": 13.23,
      "eval_steps_per_second": 6.615,
      "step": 1200
    },
    {
      "epoch": 0.48976393378391614,
      "grad_norm": 182.0519561767578,
      "learning_rate": 9.480714582533773e-07,
      "logits/chosen": -1.0906615257263184,
      "logits/rejected": -0.8988884091377258,
      "logps/chosen": -363.1963195800781,
      "logps/rejected": -310.1281433105469,
      "loss": 0.5144,
      "rewards/accuracies": 0.7200000286102295,
      "rewards/chosen": -0.7843419909477234,
      "rewards/margins": 1.230003833770752,
      "rewards/rejected": -2.014345645904541,
      "step": 1250
    },
    {
      "epoch": 0.5093544911352728,
      "grad_norm": 124.72360229492188,
      "learning_rate": 9.433322304561614e-07,
      "logits/chosen": -1.069496750831604,
      "logits/rejected": -0.8835853338241577,
      "logps/chosen": -361.09356689453125,
      "logps/rejected": -313.1136779785156,
      "loss": 0.463,
      "rewards/accuracies": 0.7825000286102295,
      "rewards/chosen": -0.662202000617981,
      "rewards/margins": 1.3414545059204102,
      "rewards/rejected": -2.0036566257476807,
      "step": 1300
    },
    {
      "epoch": 0.5289450484866295,
      "grad_norm": 124.38642120361328,
      "learning_rate": 9.383990083390903e-07,
      "logits/chosen": -1.080662488937378,
      "logits/rejected": -0.8675041794776917,
      "logps/chosen": -355.84405517578125,
      "logps/rejected": -322.5790710449219,
      "loss": 0.4625,
      "rewards/accuracies": 0.7674999833106995,
      "rewards/chosen": -0.5591977834701538,
      "rewards/margins": 1.3175528049468994,
      "rewards/rejected": -1.8767504692077637,
      "step": 1350
    },
    {
      "epoch": 0.5485356058379861,
      "grad_norm": 110.81146240234375,
      "learning_rate": 9.332739505926528e-07,
      "logits/chosen": -1.1147152185440063,
      "logits/rejected": -0.9014944434165955,
      "logps/chosen": -360.80523681640625,
      "logps/rejected": -295.3592834472656,
      "loss": 0.4819,
      "rewards/accuracies": 0.737500011920929,
      "rewards/chosen": -0.6334275603294373,
      "rewards/margins": 1.295997142791748,
      "rewards/rejected": -1.92942476272583,
      "step": 1400
    },
    {
      "epoch": 0.5485356058379861,
      "eval_logits/chosen": -1.0615795850753784,
      "eval_logits/rejected": -0.8742749094963074,
      "eval_logps/chosen": -358.2158508300781,
      "eval_logps/rejected": -304.3434753417969,
      "eval_loss": 0.4777006208896637,
      "eval_rewards/accuracies": 0.7635303735733032,
      "eval_rewards/chosen": -0.590242862701416,
      "eval_rewards/margins": 1.3008415699005127,
      "eval_rewards/rejected": -1.8910844326019287,
      "eval_runtime": 182.5694,
      "eval_samples_per_second": 13.157,
      "eval_steps_per_second": 6.578,
      "step": 1400
    },
    {
      "epoch": 0.5681261631893427,
      "grad_norm": 99.77544403076172,
      "learning_rate": 9.279592998512064e-07,
      "logits/chosen": -1.0060768127441406,
      "logits/rejected": -0.828403115272522,
      "logps/chosen": -386.8955078125,
      "logps/rejected": -324.961669921875,
      "loss": 0.4593,
      "rewards/accuracies": 0.7749999761581421,
      "rewards/chosen": -0.6105374693870544,
      "rewards/margins": 1.4274816513061523,
      "rewards/rejected": -2.0380191802978516,
      "step": 1450
    },
    {
      "epoch": 0.5877167205406993,
      "grad_norm": 113.63347625732422,
      "learning_rate": 9.224573817116396e-07,
      "logits/chosen": -1.1257575750350952,
      "logits/rejected": -0.8088225722312927,
      "logps/chosen": -356.369140625,
      "logps/rejected": -299.9069519042969,
      "loss": 0.4751,
      "rewards/accuracies": 0.7674999833106995,
      "rewards/chosen": -0.7516566514968872,
      "rewards/margins": 1.3119940757751465,
      "rewards/rejected": -2.063650608062744,
      "step": 1500
    },
    {
      "epoch": 0.6073072778920561,
      "grad_norm": 71.0758285522461,
      "learning_rate": 9.16770603715733e-07,
      "logits/chosen": -1.0219570398330688,
      "logits/rejected": -0.8258044719696045,
      "logps/chosen": -365.3489074707031,
      "logps/rejected": -317.51776123046875,
      "loss": 0.4755,
      "rewards/accuracies": 0.7724999785423279,
      "rewards/chosen": -0.722920298576355,
      "rewards/margins": 1.3333240747451782,
      "rewards/rejected": -2.0562446117401123,
      "step": 1550
    },
    {
      "epoch": 0.6268978352434127,
      "grad_norm": 63.271759033203125,
      "learning_rate": 9.109014542966609e-07,
      "logits/chosen": -1.0383367538452148,
      "logits/rejected": -0.8199602365493774,
      "logps/chosen": -361.697509765625,
      "logps/rejected": -319.4376525878906,
      "loss": 0.4271,
      "rewards/accuracies": 0.8100000023841858,
      "rewards/chosen": -0.8095757961273193,
      "rewards/margins": 1.4505212306976318,
      "rewards/rejected": -2.260097026824951,
      "step": 1600
    },
    {
      "epoch": 0.6268978352434127,
      "eval_logits/chosen": -1.0107686519622803,
      "eval_logits/rejected": -0.826732873916626,
      "eval_logps/chosen": -360.0285339355469,
      "eval_logps/rejected": -307.601318359375,
      "eval_loss": 0.4775010347366333,
      "eval_rewards/accuracies": 0.768109917640686,
      "eval_rewards/chosen": -0.7715086936950684,
      "eval_rewards/margins": 1.445361852645874,
      "eval_rewards/rejected": -2.2168707847595215,
      "eval_runtime": 181.1951,
      "eval_samples_per_second": 13.256,
      "eval_steps_per_second": 6.628,
      "step": 1600
    },
    {
      "epoch": 0.6464883925947693,
      "grad_norm": 69.98014831542969,
      "learning_rate": 9.04852501690097e-07,
      "logits/chosen": -1.0106360912322998,
      "logits/rejected": -0.769228458404541,
      "logps/chosen": -361.59783935546875,
      "logps/rejected": -298.4144287109375,
      "loss": 0.4807,
      "rewards/accuracies": 0.7749999761581421,
      "rewards/chosen": -0.6844636797904968,
      "rewards/margins": 1.4386271238327026,
      "rewards/rejected": -2.1230905055999756,
      "step": 1650
    },
    {
      "epoch": 0.6660789499461259,
      "grad_norm": 73.16764831542969,
      "learning_rate": 8.986263928104007e-07,
      "logits/chosen": -1.0747100114822388,
      "logits/rejected": -0.8522970676422119,
      "logps/chosen": -377.2361755371094,
      "logps/rejected": -314.54180908203125,
      "loss": 0.475,
      "rewards/accuracies": 0.7925000190734863,
      "rewards/chosen": -0.6853989958763123,
      "rewards/margins": 1.3294869661331177,
      "rewards/rejected": -2.014885902404785,
      "step": 1700
    },
    {
      "epoch": 0.6856695072974827,
      "grad_norm": 159.627685546875,
      "learning_rate": 8.922258520923739e-07,
      "logits/chosen": -1.1866579055786133,
      "logits/rejected": -0.9568431973457336,
      "logps/chosen": -363.7584228515625,
      "logps/rejected": -335.536376953125,
      "loss": 0.4701,
      "rewards/accuracies": 0.7549999952316284,
      "rewards/chosen": -0.6905403733253479,
      "rewards/margins": 1.4639062881469727,
      "rewards/rejected": -2.1544463634490967,
      "step": 1750
    },
    {
      "epoch": 0.7052600646488393,
      "grad_norm": 59.52585983276367,
      "learning_rate": 8.856536802990969e-07,
      "logits/chosen": -1.2136218547821045,
      "logits/rejected": -0.915833055973053,
      "logps/chosen": -368.9427795410156,
      "logps/rejected": -318.7424621582031,
      "loss": 0.4355,
      "rewards/accuracies": 0.7925000190734863,
      "rewards/chosen": -0.7404943704605103,
      "rewards/margins": 1.615696907043457,
      "rewards/rejected": -2.3561911582946777,
      "step": 1800
    },
    {
      "epoch": 0.7052600646488393,
      "eval_logits/chosen": -1.137018084526062,
      "eval_logits/rejected": -0.9567646384239197,
      "eval_logps/chosen": -360.75030517578125,
      "eval_logps/rejected": -307.82464599609375,
      "eval_loss": 0.4773857891559601,
      "eval_rewards/accuracies": 0.770191490650177,
      "eval_rewards/chosen": -0.8436892032623291,
      "eval_rewards/margins": 1.3955140113830566,
      "eval_rewards/rejected": -2.2392032146453857,
      "eval_runtime": 181.4174,
      "eval_samples_per_second": 13.24,
      "eval_steps_per_second": 6.62,
      "step": 1800
    },
    {
      "epoch": 0.7248506220001959,
      "grad_norm": 145.34722900390625,
      "learning_rate": 8.789127532963639e-07,
      "logits/chosen": -1.1216288805007935,
      "logits/rejected": -0.8633120656013489,
      "logps/chosen": -379.6500244140625,
      "logps/rejected": -340.3668212890625,
      "loss": 0.4937,
      "rewards/accuracies": 0.762499988079071,
      "rewards/chosen": -0.894076406955719,
      "rewards/margins": 1.335474967956543,
      "rewards/rejected": -2.2295515537261963,
      "step": 1850
    },
    {
      "epoch": 0.7444411793515525,
      "grad_norm": 85.14502716064453,
      "learning_rate": 8.720060207942547e-07,
      "logits/chosen": -1.1888411045074463,
      "logits/rejected": -0.8253584504127502,
      "logps/chosen": -358.1293029785156,
      "logps/rejected": -322.2793273925781,
      "loss": 0.453,
      "rewards/accuracies": 0.7774999737739563,
      "rewards/chosen": -0.7874532341957092,
      "rewards/margins": 1.4413965940475464,
      "rewards/rejected": -2.2288498878479004,
      "step": 1900
    },
    {
      "epoch": 0.7640317367029092,
      "grad_norm": 59.02031707763672,
      "learning_rate": 8.649365050563955e-07,
      "logits/chosen": -1.0339776277542114,
      "logits/rejected": -0.7359542846679688,
      "logps/chosen": -379.0245666503906,
      "logps/rejected": -335.61248779296875,
      "loss": 0.383,
      "rewards/accuracies": 0.824999988079071,
      "rewards/chosen": -0.9282081127166748,
      "rewards/margins": 1.724035382270813,
      "rewards/rejected": -2.6522433757781982,
      "step": 1950
    },
    {
      "epoch": 0.7836222940542659,
      "grad_norm": 100.15818786621094,
      "learning_rate": 8.577072995774679e-07,
      "logits/chosen": -1.0461410284042358,
      "logits/rejected": -0.7923344373703003,
      "logps/chosen": -358.7364196777344,
      "logps/rejected": -319.2044372558594,
      "loss": 0.5073,
      "rewards/accuracies": 0.7450000047683716,
      "rewards/chosen": -1.07187819480896,
      "rewards/margins": 1.472280740737915,
      "rewards/rejected": -2.544158697128296,
      "step": 2000
    },
    {
      "epoch": 0.7836222940542659,
      "eval_logits/chosen": -1.0298668146133423,
      "eval_logits/rejected": -0.857089638710022,
      "eval_logps/chosen": -361.0681457519531,
      "eval_logps/rejected": -309.28570556640625,
      "eval_loss": 0.4695436656475067,
      "eval_rewards/accuracies": 0.7756036520004272,
      "eval_rewards/chosen": -0.8754721283912659,
      "eval_rewards/margins": 1.5098369121551514,
      "eval_rewards/rejected": -2.3853089809417725,
      "eval_runtime": 181.4637,
      "eval_samples_per_second": 13.237,
      "eval_steps_per_second": 6.618,
      "step": 2000
    },
    {
      "epoch": 0.8032128514056225,
      "grad_norm": 109.68525695800781,
      "learning_rate": 8.503215677295522e-07,
      "logits/chosen": -1.1481313705444336,
      "logits/rejected": -0.8855399489402771,
      "logps/chosen": -377.3407897949219,
      "logps/rejected": -321.94903564453125,
      "loss": 0.4,
      "rewards/accuracies": 0.8075000047683716,
      "rewards/chosen": -0.8187921047210693,
      "rewards/margins": 1.6883336305618286,
      "rewards/rejected": -2.5071256160736084,
      "step": 2050
    },
    {
      "epoch": 0.8228034087569791,
      "grad_norm": 113.09908294677734,
      "learning_rate": 8.427825413778904e-07,
      "logits/chosen": -1.182418704032898,
      "logits/rejected": -1.0226842164993286,
      "logps/chosen": -370.88800048828125,
      "logps/rejected": -316.3885803222656,
      "loss": 0.4302,
      "rewards/accuracies": 0.7925000190734863,
      "rewards/chosen": -0.8482898473739624,
      "rewards/margins": 1.790774941444397,
      "rewards/rejected": -2.6390650272369385,
      "step": 2100
    }
  ],
  "logging_steps": 50,
  "max_steps": 7659,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 3,
  "save_steps": 300,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 0.0,
  "train_batch_size": 2,
  "trial_name": null,
  "trial_params": null
}