{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 1.0,
  "eval_steps": 50,
  "global_step": 436,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.022935779816513763,
      "grad_norm": 3.1641646757229265,
      "learning_rate": 1.1363636363636363e-07,
      "logits/chosen": -2.6194896697998047,
      "logits/rejected": -2.552656650543213,
      "logps/chosen": -265.41778564453125,
      "logps/rejected": -236.15280151367188,
      "loss": 0.0585,
      "rewards/accuracies": 0.39375001192092896,
      "rewards/chosen": 0.00028036616276949644,
      "rewards/margins": 0.0005085940356366336,
      "rewards/rejected": -0.00022822784376330674,
      "step": 10
    },
    {
      "epoch": 0.045871559633027525,
      "grad_norm": 2.9506602825308903,
      "learning_rate": 2.2727272727272726e-07,
      "logits/chosen": -2.6579580307006836,
      "logits/rejected": -2.575996160507202,
      "logps/chosen": -298.8293151855469,
      "logps/rejected": -274.3349304199219,
      "loss": 0.059,
      "rewards/accuracies": 0.643750011920929,
      "rewards/chosen": -0.0004406004154589027,
      "rewards/margins": 0.0018091611564159393,
      "rewards/rejected": -0.0022497617173939943,
      "step": 20
    },
    {
      "epoch": 0.06880733944954129,
      "grad_norm": 2.7648586784794067,
      "learning_rate": 3.4090909090909085e-07,
      "logits/chosen": -2.6760480403900146,
      "logits/rejected": -2.6023683547973633,
      "logps/chosen": -290.41986083984375,
      "logps/rejected": -234.40579223632812,
      "loss": 0.0571,
      "rewards/accuracies": 0.699999988079071,
      "rewards/chosen": 0.009129038080573082,
      "rewards/margins": 0.01336810551583767,
      "rewards/rejected": -0.004239068366587162,
      "step": 30
    },
    {
      "epoch": 0.09174311926605505,
      "grad_norm": 2.7634574887860825,
      "learning_rate": 4.545454545454545e-07,
      "logits/chosen": -2.6598825454711914,
      "logits/rejected": -2.610196352005005,
      "logps/chosen": -280.923828125,
      "logps/rejected": -267.7273864746094,
      "loss": 0.0529,
      "rewards/accuracies": 0.6875,
      "rewards/chosen": 0.04245365411043167,
      "rewards/margins": 0.04179743677377701,
      "rewards/rejected": 0.0006562232738360763,
      "step": 40
    },
    {
      "epoch": 0.11467889908256881,
      "grad_norm": 2.9044861722207567,
      "learning_rate": 4.997110275491701e-07,
      "logits/chosen": -2.620276927947998,
      "logits/rejected": -2.6136012077331543,
      "logps/chosen": -292.2350158691406,
      "logps/rejected": -302.38018798828125,
      "loss": 0.0487,
      "rewards/accuracies": 0.6937500238418579,
      "rewards/chosen": 0.04016200453042984,
      "rewards/margins": 0.07212186604738235,
      "rewards/rejected": -0.031959857791662216,
      "step": 50
    },
    {
      "epoch": 0.11467889908256881,
      "eval_logits/chosen": -2.5716989040374756,
      "eval_logits/rejected": -2.4927897453308105,
      "eval_logps/chosen": -282.0911865234375,
      "eval_logps/rejected": -254.96926879882812,
      "eval_loss": 0.04548575356602669,
      "eval_rewards/accuracies": 0.7025862336158752,
      "eval_rewards/chosen": 0.02998923696577549,
      "eval_rewards/margins": 0.11445755511522293,
      "eval_rewards/rejected": -0.08446833491325378,
      "eval_runtime": 95.7446,
      "eval_samples_per_second": 18.988,
      "eval_steps_per_second": 0.303,
      "step": 50
    },
    {
      "epoch": 0.13761467889908258,
      "grad_norm": 3.2642316994426857,
      "learning_rate": 4.979475034558115e-07,
      "logits/chosen": -2.5661163330078125,
      "logits/rejected": -2.50827956199646,
      "logps/chosen": -289.7759704589844,
      "logps/rejected": -267.379638671875,
      "loss": 0.0449,
      "rewards/accuracies": 0.706250011920929,
      "rewards/chosen": -0.0029130387119948864,
      "rewards/margins": 0.09531786292791367,
      "rewards/rejected": -0.09823091328144073,
      "step": 60
    },
    {
      "epoch": 0.16055045871559634,
      "grad_norm": 3.167167289379583,
      "learning_rate": 4.945923025551788e-07,
      "logits/chosen": -2.460839033126831,
      "logits/rejected": -2.4139201641082764,
      "logps/chosen": -324.13690185546875,
      "logps/rejected": -276.5367126464844,
      "loss": 0.0446,
      "rewards/accuracies": 0.706250011920929,
      "rewards/chosen": 0.034340955317020416,
      "rewards/margins": 0.1856319159269333,
      "rewards/rejected": -0.15129096806049347,
      "step": 70
    },
    {
      "epoch": 0.1834862385321101,
      "grad_norm": 3.00053019606861,
      "learning_rate": 4.896669632591651e-07,
      "logits/chosen": -2.356163740158081,
      "logits/rejected": -2.2467029094696045,
      "logps/chosen": -287.4019470214844,
      "logps/rejected": -272.3543701171875,
      "loss": 0.0416,
      "rewards/accuracies": 0.71875,
      "rewards/chosen": -0.015520763583481312,
      "rewards/margins": 0.18146219849586487,
      "rewards/rejected": -0.1969829499721527,
      "step": 80
    },
    {
      "epoch": 0.20642201834862386,
      "grad_norm": 3.936347429433164,
      "learning_rate": 4.832031033425662e-07,
      "logits/chosen": -2.0837090015411377,
      "logits/rejected": -2.0328478813171387,
      "logps/chosen": -287.70654296875,
      "logps/rejected": -268.2559509277344,
      "loss": 0.0405,
      "rewards/accuracies": 0.7124999761581421,
      "rewards/chosen": -0.001131159020587802,
      "rewards/margins": 0.2134155035018921,
      "rewards/rejected": -0.21454665064811707,
      "step": 90
    },
    {
      "epoch": 0.22935779816513763,
      "grad_norm": 3.652555898125068,
      "learning_rate": 4.752422169756047e-07,
      "logits/chosen": -1.5697656869888306,
      "logits/rejected": -1.561859130859375,
      "logps/chosen": -289.6307678222656,
      "logps/rejected": -307.66363525390625,
      "loss": 0.0408,
      "rewards/accuracies": 0.6625000238418579,
      "rewards/chosen": -0.17206048965454102,
      "rewards/margins": 0.18529178202152252,
      "rewards/rejected": -0.35735225677490234,
      "step": 100
    },
    {
      "epoch": 0.22935779816513763,
      "eval_logits/chosen": -1.8404057025909424,
      "eval_logits/rejected": -1.7416090965270996,
      "eval_logps/chosen": -295.8004455566406,
      "eval_logps/rejected": -277.408935546875,
      "eval_loss": 0.03911930322647095,
      "eval_rewards/accuracies": 0.6896551847457886,
      "eval_rewards/chosen": -0.10710364580154419,
      "eval_rewards/margins": 0.201760932803154,
      "eval_rewards/rejected": -0.308864563703537,
      "eval_runtime": 95.267,
      "eval_samples_per_second": 19.083,
      "eval_steps_per_second": 0.304,
      "step": 100
    },
    {
      "epoch": 0.25229357798165136,
      "grad_norm": 3.509676092042844,
      "learning_rate": 4.658354083558188e-07,
      "logits/chosen": -1.9262816905975342,
      "logits/rejected": -1.80355966091156,
      "logps/chosen": -279.3715515136719,
      "logps/rejected": -263.81549072265625,
      "loss": 0.0402,
      "rewards/accuracies": 0.6812499761581421,
      "rewards/chosen": -0.14183419942855835,
      "rewards/margins": 0.15898862481117249,
      "rewards/rejected": -0.30082279443740845,
      "step": 110
    },
    {
      "epoch": 0.27522935779816515,
      "grad_norm": 3.984377752564919,
      "learning_rate": 4.550430636492389e-07,
      "logits/chosen": -1.8143861293792725,
      "logits/rejected": -1.7607357501983643,
      "logps/chosen": -295.43243408203125,
      "logps/rejected": -289.3270568847656,
      "loss": 0.0395,
      "rewards/accuracies": 0.675000011920929,
      "rewards/chosen": -0.19613118469715118,
      "rewards/margins": 0.18250016868114471,
      "rewards/rejected": -0.3786313831806183,
      "step": 120
    },
    {
      "epoch": 0.2981651376146789,
      "grad_norm": 3.895538046894766,
      "learning_rate": 4.429344633468004e-07,
      "logits/chosen": -1.742222785949707,
      "logits/rejected": -1.6753323078155518,
      "logps/chosen": -268.340087890625,
      "logps/rejected": -275.31500244140625,
      "loss": 0.0396,
      "rewards/accuracies": 0.6875,
      "rewards/chosen": -0.12159997224807739,
      "rewards/margins": 0.20542438328266144,
      "rewards/rejected": -0.32702434062957764,
      "step": 130
    },
    {
      "epoch": 0.3211009174311927,
      "grad_norm": 4.968671690125824,
      "learning_rate": 4.2958733752443187e-07,
      "logits/chosen": -1.6684496402740479,
      "logits/rejected": -1.5686622858047485,
      "logps/chosen": -286.41973876953125,
      "logps/rejected": -254.42691040039062,
      "loss": 0.0381,
      "rewards/accuracies": 0.71875,
      "rewards/chosen": -0.10234608501195908,
      "rewards/margins": 0.19895747303962708,
      "rewards/rejected": -0.30130359530448914,
      "step": 140
    },
    {
      "epoch": 0.3440366972477064,
      "grad_norm": 3.5584001205811377,
      "learning_rate": 4.150873668617898e-07,
      "logits/chosen": -1.703791856765747,
      "logits/rejected": -1.5906720161437988,
      "logps/chosen": -280.5138854980469,
      "logps/rejected": -269.3014831542969,
      "loss": 0.0379,
      "rewards/accuracies": 0.7250000238418579,
      "rewards/chosen": -0.13970300555229187,
      "rewards/margins": 0.22143635153770447,
      "rewards/rejected": -0.36113935708999634,
      "step": 150
    },
    {
      "epoch": 0.3440366972477064,
      "eval_logits/chosen": -1.7903783321380615,
      "eval_logits/rejected": -1.635672688484192,
      "eval_logps/chosen": -299.1519470214844,
      "eval_logps/rejected": -287.88739013671875,
      "eval_loss": 0.03649381175637245,
      "eval_rewards/accuracies": 0.7155172228813171,
      "eval_rewards/chosen": -0.14061833918094635,
      "eval_rewards/margins": 0.27303099632263184,
      "eval_rewards/rejected": -0.413649320602417,
      "eval_runtime": 94.6259,
      "eval_samples_per_second": 19.212,
      "eval_steps_per_second": 0.306,
      "step": 150
    },
    {
      "epoch": 0.3669724770642202,
      "grad_norm": 5.2149351876346905,
      "learning_rate": 3.9952763262280397e-07,
      "logits/chosen": -1.6572643518447876,
      "logits/rejected": -1.5417451858520508,
      "logps/chosen": -309.5204162597656,
      "logps/rejected": -324.3457946777344,
      "loss": 0.0359,
      "rewards/accuracies": 0.731249988079071,
      "rewards/chosen": -0.12969312071800232,
      "rewards/margins": 0.254078209400177,
      "rewards/rejected": -0.3837713599205017,
      "step": 160
    },
    {
      "epoch": 0.38990825688073394,
      "grad_norm": 2.9551901080690395,
      "learning_rate": 3.8300801912883414e-07,
      "logits/chosen": -1.6375774145126343,
      "logits/rejected": -1.5148413181304932,
      "logps/chosen": -277.994873046875,
      "logps/rejected": -301.9854431152344,
      "loss": 0.0339,
      "rewards/accuracies": 0.731249988079071,
      "rewards/chosen": -0.06393978744745255,
      "rewards/margins": 0.2552258372306824,
      "rewards/rejected": -0.3191656172275543,
      "step": 170
    },
    {
      "epoch": 0.41284403669724773,
      "grad_norm": 3.6761291533851423,
      "learning_rate": 3.6563457256020884e-07,
      "logits/chosen": -1.527038335800171,
      "logits/rejected": -1.2771427631378174,
      "logps/chosen": -323.6756896972656,
      "logps/rejected": -284.1974792480469,
      "loss": 0.0365,
      "rewards/accuracies": 0.7437499761581421,
      "rewards/chosen": -0.20317766070365906,
      "rewards/margins": 0.2538478374481201,
      "rewards/rejected": -0.4570254683494568,
      "step": 180
    },
    {
      "epoch": 0.43577981651376146,
      "grad_norm": 5.675419312165092,
      "learning_rate": 3.475188202022617e-07,
      "logits/chosen": -1.6745436191558838,
      "logits/rejected": -1.604264259338379,
      "logps/chosen": -268.5342102050781,
      "logps/rejected": -302.51898193359375,
      "loss": 0.0368,
      "rewards/accuracies": 0.7749999761581421,
      "rewards/chosen": -0.10419674217700958,
      "rewards/margins": 0.27754858136177063,
      "rewards/rejected": -0.38174527883529663,
      "step": 190
    },
    {
      "epoch": 0.45871559633027525,
      "grad_norm": 3.4589111733943283,
      "learning_rate": 3.287770545059052e-07,
      "logits/chosen": -1.8974136114120483,
      "logits/rejected": -1.7577918767929077,
      "logps/chosen": -291.4000549316406,
      "logps/rejected": -281.4195861816406,
      "loss": 0.0365,
      "rewards/accuracies": 0.7562500238418579,
      "rewards/chosen": -0.1311841905117035,
      "rewards/margins": 0.24445155262947083,
      "rewards/rejected": -0.3756357431411743,
      "step": 200
    },
    {
      "epoch": 0.45871559633027525,
      "eval_logits/chosen": -1.9178062677383423,
      "eval_logits/rejected": -1.7552225589752197,
      "eval_logps/chosen": -291.5865783691406,
      "eval_logps/rejected": -279.1630859375,
      "eval_loss": 0.03504867106676102,
      "eval_rewards/accuracies": 0.7543103694915771,
      "eval_rewards/chosen": -0.06496486067771912,
      "eval_rewards/margins": 0.2614416182041168,
      "eval_rewards/rejected": -0.32640647888183594,
      "eval_runtime": 95.5023,
      "eval_samples_per_second": 19.036,
      "eval_steps_per_second": 0.304,
      "step": 200
    },
    {
      "epoch": 0.481651376146789,
      "grad_norm": 2.8163994777137717,
      "learning_rate": 3.0952958655864954e-07,
      "logits/chosen": -1.794647455215454,
      "logits/rejected": -1.7260096073150635,
      "logps/chosen": -280.9212951660156,
      "logps/rejected": -287.47674560546875,
      "loss": 0.0349,
      "rewards/accuracies": 0.762499988079071,
      "rewards/chosen": -0.07616890966892242,
      "rewards/margins": 0.2551766633987427,
      "rewards/rejected": -0.3313455581665039,
      "step": 210
    },
    {
      "epoch": 0.5045871559633027,
      "grad_norm": 3.602353089915511,
      "learning_rate": 2.898999737583448e-07,
      "logits/chosen": -1.672721266746521,
      "logits/rejected": -1.4848198890686035,
      "logps/chosen": -344.14825439453125,
      "logps/rejected": -337.38470458984375,
      "loss": 0.0336,
      "rewards/accuracies": 0.8062499761581421,
      "rewards/chosen": -0.2146792709827423,
      "rewards/margins": 0.3094654977321625,
      "rewards/rejected": -0.5241447687149048,
      "step": 220
    },
    {
      "epoch": 0.5275229357798165,
      "grad_norm": 3.2287234258499633,
      "learning_rate": 2.7001422664752333e-07,
      "logits/chosen": -1.4607375860214233,
      "logits/rejected": -1.3907452821731567,
      "logps/chosen": -288.92095947265625,
      "logps/rejected": -309.4497985839844,
      "loss": 0.0354,
      "rewards/accuracies": 0.731249988079071,
      "rewards/chosen": -0.20977428555488586,
      "rewards/margins": 0.22852472960948944,
      "rewards/rejected": -0.4382990002632141,
      "step": 230
    },
    {
      "epoch": 0.5504587155963303,
      "grad_norm": 3.200697411595364,
      "learning_rate": 2.5e-07,
      "logits/chosen": -1.6722291707992554,
      "logits/rejected": -1.496265172958374,
      "logps/chosen": -304.0669250488281,
      "logps/rejected": -301.3495788574219,
      "loss": 0.0352,
      "rewards/accuracies": 0.75,
      "rewards/chosen": -0.16535088419914246,
      "rewards/margins": 0.23464509844779968,
      "rewards/rejected": -0.39999598264694214,
      "step": 240
    },
    {
      "epoch": 0.573394495412844,
      "grad_norm": 3.068087094061713,
      "learning_rate": 2.2998577335247667e-07,
      "logits/chosen": -1.8118059635162354,
      "logits/rejected": -1.58005952835083,
      "logps/chosen": -316.8459777832031,
      "logps/rejected": -301.0585632324219,
      "loss": 0.0346,
      "rewards/accuracies": 0.7749999761581421,
      "rewards/chosen": -0.1305214762687683,
      "rewards/margins": 0.27562573552131653,
      "rewards/rejected": -0.40614718198776245,
      "step": 250
    },
    {
      "epoch": 0.573394495412844,
      "eval_logits/chosen": -1.719161868095398,
      "eval_logits/rejected": -1.4870593547821045,
      "eval_logps/chosen": -298.28277587890625,
      "eval_logps/rejected": -291.9156494140625,
      "eval_loss": 0.033715467900037766,
      "eval_rewards/accuracies": 0.7543103694915771,
      "eval_rewards/chosen": -0.1319267898797989,
      "eval_rewards/margins": 0.3220053017139435,
      "eval_rewards/rejected": -0.4539320766925812,
      "eval_runtime": 94.8489,
      "eval_samples_per_second": 19.167,
      "eval_steps_per_second": 0.306,
      "step": 250
    },
    {
      "epoch": 0.5963302752293578,
      "grad_norm": 2.973513185196363,
      "learning_rate": 2.1010002624165524e-07,
      "logits/chosen": -1.625998854637146,
      "logits/rejected": -1.528102159500122,
      "logps/chosen": -289.2758483886719,
      "logps/rejected": -324.4759216308594,
      "loss": 0.0343,
      "rewards/accuracies": 0.800000011920929,
      "rewards/chosen": -0.18561328947544098,
      "rewards/margins": 0.3152090609073639,
      "rewards/rejected": -0.5008223652839661,
      "step": 260
    },
    {
      "epoch": 0.6192660550458715,
      "grad_norm": 3.3992650920748635,
      "learning_rate": 1.9047041344135043e-07,
      "logits/chosen": -1.468390703201294,
      "logits/rejected": -1.4237253665924072,
      "logps/chosen": -287.03558349609375,
      "logps/rejected": -301.84613037109375,
      "loss": 0.0339,
      "rewards/accuracies": 0.8125,
      "rewards/chosen": -0.209732323884964,
      "rewards/margins": 0.2902582287788391,
      "rewards/rejected": -0.4999905526638031,
      "step": 270
    },
    {
      "epoch": 0.6422018348623854,
      "grad_norm": 3.2930402827157863,
      "learning_rate": 1.7122294549409482e-07,
      "logits/chosen": -1.6038357019424438,
      "logits/rejected": -1.4040790796279907,
      "logps/chosen": -287.1599426269531,
      "logps/rejected": -311.6736755371094,
      "loss": 0.0349,
      "rewards/accuracies": 0.7875000238418579,
      "rewards/chosen": -0.14161059260368347,
      "rewards/margins": 0.3371681869029999,
      "rewards/rejected": -0.4787788391113281,
      "step": 280
    },
    {
      "epoch": 0.6651376146788991,
      "grad_norm": 3.539437987763887,
      "learning_rate": 1.524811797977383e-07,
      "logits/chosen": -1.6025069952011108,
      "logits/rejected": -1.4434142112731934,
      "logps/chosen": -300.49578857421875,
      "logps/rejected": -295.3720397949219,
      "loss": 0.0334,
      "rewards/accuracies": 0.7437499761581421,
      "rewards/chosen": -0.12764978408813477,
      "rewards/margins": 0.2869029641151428,
      "rewards/rejected": -0.4145527780056,
      "step": 290
    },
    {
      "epoch": 0.6880733944954128,
      "grad_norm": 3.11392798700287,
      "learning_rate": 1.3436542743979125e-07,
      "logits/chosen": -1.6770799160003662,
      "logits/rejected": -1.6160328388214111,
      "logps/chosen": -319.42236328125,
      "logps/rejected": -295.0827331542969,
      "loss": 0.036,
      "rewards/accuracies": 0.762499988079071,
      "rewards/chosen": -0.1347302943468094,
      "rewards/margins": 0.22749857604503632,
      "rewards/rejected": -0.3622289299964905,
      "step": 300
    },
    {
      "epoch": 0.6880733944954128,
      "eval_logits/chosen": -1.6834876537322998,
      "eval_logits/rejected": -1.4841814041137695,
      "eval_logps/chosen": -298.45037841796875,
      "eval_logps/rejected": -289.4286193847656,
      "eval_loss": 0.03310078755021095,
      "eval_rewards/accuracies": 0.75,
      "eval_rewards/chosen": -0.13360273838043213,
      "eval_rewards/margins": 0.29545897245407104,
      "eval_rewards/rejected": -0.4290617108345032,
      "eval_runtime": 94.9394,
      "eval_samples_per_second": 19.149,
      "eval_steps_per_second": 0.305,
      "step": 300
    },
    {
      "epoch": 0.7110091743119266,
      "grad_norm": 5.19818595642257,
      "learning_rate": 1.1699198087116588e-07,
      "logits/chosen": -1.741039514541626,
      "logits/rejected": -1.6022865772247314,
      "logps/chosen": -294.25433349609375,
      "logps/rejected": -306.18817138671875,
      "loss": 0.0357,
      "rewards/accuracies": 0.71875,
      "rewards/chosen": -0.16289182007312775,
      "rewards/margins": 0.23449257016181946,
      "rewards/rejected": -0.397384375333786,
      "step": 310
    },
    {
      "epoch": 0.7339449541284404,
      "grad_norm": 3.4771448226315957,
      "learning_rate": 1.00472367377196e-07,
      "logits/chosen": -1.6833099126815796,
      "logits/rejected": -1.5138328075408936,
      "logps/chosen": -290.2877197265625,
      "logps/rejected": -281.9583435058594,
      "loss": 0.0353,
      "rewards/accuracies": 0.7562500238418579,
      "rewards/chosen": -0.16055694222450256,
      "rewards/margins": 0.3141244947910309,
      "rewards/rejected": -0.47468137741088867,
      "step": 320
    },
    {
      "epoch": 0.7568807339449541,
      "grad_norm": 3.5214688030455066,
      "learning_rate": 8.49126331382102e-08,
      "logits/chosen": -1.6057231426239014,
      "logits/rejected": -1.4786289930343628,
      "logps/chosen": -291.7789306640625,
      "logps/rejected": -288.677001953125,
      "loss": 0.0348,
      "rewards/accuracies": 0.706250011920929,
      "rewards/chosen": -0.17256753146648407,
      "rewards/margins": 0.24320188164710999,
      "rewards/rejected": -0.41576939821243286,
      "step": 330
    },
    {
      "epoch": 0.7798165137614679,
      "grad_norm": 3.5006792226854495,
      "learning_rate": 7.041266247556812e-08,
      "logits/chosen": -1.7440093755722046,
      "logits/rejected": -1.6453059911727905,
      "logps/chosen": -304.0636901855469,
      "logps/rejected": -299.057861328125,
      "loss": 0.0337,
      "rewards/accuracies": 0.78125,
      "rewards/chosen": -0.16071084141731262,
      "rewards/margins": 0.23002786934375763,
      "rewards/rejected": -0.39073869585990906,
      "step": 340
    },
    {
      "epoch": 0.8027522935779816,
      "grad_norm": 3.5127255839066094,
      "learning_rate": 5.706553665319955e-08,
      "logits/chosen": -1.7355537414550781,
      "logits/rejected": -1.5236848592758179,
      "logps/chosen": -292.27850341796875,
      "logps/rejected": -280.9353332519531,
      "loss": 0.0359,
      "rewards/accuracies": 0.7562500238418579,
      "rewards/chosen": -0.17787429690361023,
      "rewards/margins": 0.25486475229263306,
      "rewards/rejected": -0.4327389597892761,
      "step": 350
    },
    {
      "epoch": 0.8027522935779816,
      "eval_logits/chosen": -1.6786376237869263,
      "eval_logits/rejected": -1.465800166130066,
      "eval_logps/chosen": -298.86663818359375,
      "eval_logps/rejected": -291.2398681640625,
      "eval_loss": 0.0327293798327446,
      "eval_rewards/accuracies": 0.7586206793785095,
      "eval_rewards/chosen": -0.1377655267715454,
      "eval_rewards/margins": 0.30940860509872437,
      "eval_rewards/rejected": -0.44717416167259216,
      "eval_runtime": 93.7918,
      "eval_samples_per_second": 19.383,
      "eval_steps_per_second": 0.309,
      "step": 350
    },
    {
      "epoch": 0.8256880733944955,
      "grad_norm": 2.7935239013392237,
      "learning_rate": 4.4956936350761005e-08,
      "logits/chosen": -1.6336042881011963,
      "logits/rejected": -1.5328700542449951,
      "logps/chosen": -265.51251220703125,
      "logps/rejected": -297.154541015625,
      "loss": 0.0328,
      "rewards/accuracies": 0.737500011920929,
      "rewards/chosen": -0.1319248378276825,
      "rewards/margins": 0.2542495131492615,
      "rewards/rejected": -0.38617435097694397,
      "step": 360
    },
    {
      "epoch": 0.8486238532110092,
      "grad_norm": 3.203845878056458,
      "learning_rate": 3.416459164418123e-08,
      "logits/chosen": -1.746138572692871,
      "logits/rejected": -1.5625007152557373,
      "logps/chosen": -313.5556640625,
      "logps/rejected": -302.9591369628906,
      "loss": 0.0326,
      "rewards/accuracies": 0.737500011920929,
      "rewards/chosen": -0.13971158862113953,
      "rewards/margins": 0.2641337513923645,
      "rewards/rejected": -0.4038453996181488,
      "step": 370
    },
    {
      "epoch": 0.8715596330275229,
      "grad_norm": 3.9723478814606876,
      "learning_rate": 2.475778302439524e-08,
      "logits/chosen": -1.700884461402893,
      "logits/rejected": -1.4876903295516968,
      "logps/chosen": -308.3489074707031,
      "logps/rejected": -299.65850830078125,
      "loss": 0.0342,
      "rewards/accuracies": 0.78125,
      "rewards/chosen": -0.11970899254083633,
      "rewards/margins": 0.3003136217594147,
      "rewards/rejected": -0.4200226366519928,
      "step": 380
    },
    {
      "epoch": 0.8944954128440367,
      "grad_norm": 3.433718178917216,
      "learning_rate": 1.6796896657433805e-08,
      "logits/chosen": -1.6044851541519165,
      "logits/rejected": -1.4463411569595337,
      "logps/chosen": -265.7919006347656,
      "logps/rejected": -267.8514099121094,
      "loss": 0.0356,
      "rewards/accuracies": 0.7250000238418579,
      "rewards/chosen": -0.13834409415721893,
      "rewards/margins": 0.2471076250076294,
      "rewards/rejected": -0.3854517340660095,
      "step": 390
    },
    {
      "epoch": 0.9174311926605505,
      "grad_norm": 3.420165449234039,
      "learning_rate": 1.0333036740834855e-08,
      "logits/chosen": -1.5561296939849854,
      "logits/rejected": -1.4712246656417847,
      "logps/chosen": -236.0593719482422,
      "logps/rejected": -266.1527404785156,
      "loss": 0.0351,
      "rewards/accuracies": 0.731249988079071,
      "rewards/chosen": -0.11499743163585663,
      "rewards/margins": 0.2543741464614868,
      "rewards/rejected": -0.36937159299850464,
      "step": 400
    },
    {
      "epoch": 0.9174311926605505,
      "eval_logits/chosen": -1.6862857341766357,
      "eval_logits/rejected": -1.476194977760315,
      "eval_logps/chosen": -297.669189453125,
      "eval_logps/rejected": -289.0508728027344,
      "eval_loss": 0.0327373668551445,
      "eval_rewards/accuracies": 0.7543103694915771,
      "eval_rewards/chosen": -0.12579074501991272,
      "eval_rewards/margins": 0.2994934022426605,
      "eval_rewards/rejected": -0.42528414726257324,
      "eval_runtime": 95.429,
      "eval_samples_per_second": 19.051,
      "eval_steps_per_second": 0.304,
      "step": 400
    },
    {
      "epoch": 0.9403669724770642,
      "grad_norm": 3.5258604798563833,
      "learning_rate": 5.4076974448211685e-09,
      "logits/chosen": -1.5972687005996704,
      "logits/rejected": -1.4251827001571655,
      "logps/chosen": -284.83367919921875,
      "logps/rejected": -281.9726257324219,
      "loss": 0.0339,
      "rewards/accuracies": 0.7562500238418579,
      "rewards/chosen": -0.13881805539131165,
      "rewards/margins": 0.30044105648994446,
      "rewards/rejected": -0.4392591118812561,
      "step": 410
    },
    {
      "epoch": 0.963302752293578,
      "grad_norm": 5.008109192532265,
      "learning_rate": 2.052496544188487e-09,
      "logits/chosen": -1.5873286724090576,
      "logits/rejected": -1.4294344186782837,
      "logps/chosen": -275.04388427734375,
      "logps/rejected": -290.7890319824219,
      "loss": 0.0341,
      "rewards/accuracies": 0.7875000238418579,
      "rewards/chosen": -0.1560489982366562,
      "rewards/margins": 0.293067991733551,
      "rewards/rejected": -0.4491170048713684,
      "step": 420
    },
    {
      "epoch": 0.9862385321100917,
      "grad_norm": 3.385882880558537,
      "learning_rate": 2.889724508297886e-10,
      "logits/chosen": -1.6612422466278076,
      "logits/rejected": -1.3801202774047852,
      "logps/chosen": -316.4678039550781,
      "logps/rejected": -282.8833923339844,
      "loss": 0.0343,
      "rewards/accuracies": 0.768750011920929,
      "rewards/chosen": -0.13292355835437775,
      "rewards/margins": 0.25554487109184265,
      "rewards/rejected": -0.3884683847427368,
      "step": 430
    },
    {
      "epoch": 1.0,
      "step": 436,
      "total_flos": 0.0,
      "train_loss": 0.0384325077500912,
      "train_runtime": 11659.9634,
      "train_samples_per_second": 4.782,
      "train_steps_per_second": 0.037
    }
  ],
  "logging_steps": 10,
  "max_steps": 436,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 1,
  "save_steps": 100,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 0.0,
  "train_batch_size": 8,
  "trial_name": null,
  "trial_params": null
}