{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.9994242947610823,
  "eval_steps": 100,
  "global_step": 868,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.0,
      "grad_norm": 146.8957421194674,
      "learning_rate": 5.747126436781609e-09,
      "logits/chosen": -1.8686045408248901,
      "logits/rejected": -1.7644572257995605,
      "logps/chosen": -235.48362731933594,
      "logps/rejected": -183.77415466308594,
      "loss": 0.6922,
      "rewards/accuracies": 0.0,
      "rewards/chosen": 0.0,
      "rewards/margins": 0.0,
      "rewards/rejected": 0.0,
      "step": 1
    },
    {
      "epoch": 0.01,
      "grad_norm": 237.04909600464902,
      "learning_rate": 5.747126436781609e-08,
      "logits/chosen": -1.9218311309814453,
      "logits/rejected": -1.8686226606369019,
      "logps/chosen": -240.50628662109375,
      "logps/rejected": -216.8230438232422,
      "loss": 0.6941,
      "rewards/accuracies": 0.4444444477558136,
      "rewards/chosen": -0.002868607407435775,
      "rewards/margins": -0.0005126786418259144,
      "rewards/rejected": -0.002355928998440504,
      "step": 10
    },
    {
      "epoch": 0.02,
      "grad_norm": 211.82620683349913,
      "learning_rate": 1.1494252873563217e-07,
      "logits/chosen": -2.010253429412842,
      "logits/rejected": -1.8642921447753906,
      "logps/chosen": -283.1783752441406,
      "logps/rejected": -215.68887329101562,
      "loss": 0.6864,
      "rewards/accuracies": 0.612500011920929,
      "rewards/chosen": 0.008166970685124397,
      "rewards/margins": 0.013039084151387215,
      "rewards/rejected": -0.004872114397585392,
      "step": 20
    },
    {
      "epoch": 0.03,
      "grad_norm": 112.97267570781544,
      "learning_rate": 1.7241379310344828e-07,
      "logits/chosen": -1.9509646892547607,
      "logits/rejected": -1.8835735321044922,
      "logps/chosen": -240.29074096679688,
      "logps/rejected": -221.15274047851562,
      "loss": 0.666,
      "rewards/accuracies": 0.668749988079071,
      "rewards/chosen": 0.06300461292266846,
      "rewards/margins": 0.06308884918689728,
      "rewards/rejected": -8.423496183240786e-05,
      "step": 30
    },
    {
      "epoch": 0.05,
      "grad_norm": 95.74241597637995,
      "learning_rate": 2.2988505747126435e-07,
      "logits/chosen": -1.9107061624526978,
      "logits/rejected": -1.8901317119598389,
      "logps/chosen": -237.59036254882812,
      "logps/rejected": -216.1823272705078,
      "loss": 0.6254,
      "rewards/accuracies": 0.7124999761581421,
      "rewards/chosen": 0.1618741899728775,
      "rewards/margins": 0.1895591914653778,
      "rewards/rejected": -0.027684981003403664,
      "step": 40
    },
    {
      "epoch": 0.06,
      "grad_norm": 99.41131976874209,
      "learning_rate": 2.873563218390804e-07,
      "logits/chosen": -1.9803855419158936,
      "logits/rejected": -1.9117343425750732,
      "logps/chosen": -222.8354949951172,
      "logps/rejected": -207.8356170654297,
      "loss": 0.587,
      "rewards/accuracies": 0.768750011920929,
      "rewards/chosen": 0.23140163719654083,
      "rewards/margins": 0.3971676230430603,
      "rewards/rejected": -0.16576598584651947,
      "step": 50
    },
    {
      "epoch": 0.07,
      "grad_norm": 85.607755330573,
      "learning_rate": 3.4482758620689656e-07,
      "logits/chosen": -2.0026185512542725,
      "logits/rejected": -1.9393552541732788,
      "logps/chosen": -283.61199951171875,
      "logps/rejected": -245.7741241455078,
      "loss": 0.5351,
      "rewards/accuracies": 0.793749988079071,
      "rewards/chosen": 0.2226170301437378,
      "rewards/margins": 0.6314491629600525,
      "rewards/rejected": -0.4088321626186371,
      "step": 60
    },
    {
      "epoch": 0.08,
      "grad_norm": 105.09596276450131,
      "learning_rate": 4.0229885057471266e-07,
      "logits/chosen": -1.9401233196258545,
      "logits/rejected": -1.9097837209701538,
      "logps/chosen": -223.41983032226562,
      "logps/rejected": -222.15283203125,
      "loss": 0.5181,
      "rewards/accuracies": 0.7749999761581421,
      "rewards/chosen": -0.10973484814167023,
      "rewards/margins": 0.7763268351554871,
      "rewards/rejected": -0.8860616683959961,
      "step": 70
    },
    {
      "epoch": 0.09,
      "grad_norm": 103.35591967882027,
      "learning_rate": 4.597701149425287e-07,
      "logits/chosen": -1.8168106079101562,
      "logits/rejected": -1.7877649068832397,
      "logps/chosen": -227.1862030029297,
      "logps/rejected": -217.2149200439453,
      "loss": 0.487,
      "rewards/accuracies": 0.824999988079071,
      "rewards/chosen": -0.1610104888677597,
      "rewards/margins": 0.8937110900878906,
      "rewards/rejected": -1.0547215938568115,
      "step": 80
    },
    {
      "epoch": 0.1,
      "grad_norm": 96.75574881093117,
      "learning_rate": 4.999817969178237e-07,
      "logits/chosen": -1.8949458599090576,
      "logits/rejected": -1.8469880819320679,
      "logps/chosen": -249.9053192138672,
      "logps/rejected": -233.94631958007812,
      "loss": 0.4415,
      "rewards/accuracies": 0.8062499761581421,
      "rewards/chosen": -0.24459132552146912,
      "rewards/margins": 1.046515703201294,
      "rewards/rejected": -1.291106939315796,
      "step": 90
    },
    {
      "epoch": 0.12,
      "grad_norm": 92.71145741103783,
      "learning_rate": 4.996582603056428e-07,
      "logits/chosen": -1.8467813730239868,
      "logits/rejected": -1.7593021392822266,
      "logps/chosen": -251.5709686279297,
      "logps/rejected": -255.50814819335938,
      "loss": 0.4898,
      "rewards/accuracies": 0.7749999761581421,
      "rewards/chosen": -0.3932832181453705,
      "rewards/margins": 1.1354423761367798,
      "rewards/rejected": -1.5287256240844727,
      "step": 100
    },
    {
      "epoch": 0.12,
      "eval_logits/chosen": -1.8469043970108032,
      "eval_logits/rejected": -1.7659016847610474,
      "eval_logps/chosen": -339.4445495605469,
      "eval_logps/rejected": -353.2087707519531,
      "eval_loss": 0.5505225658416748,
      "eval_rewards/accuracies": 0.6875,
      "eval_rewards/chosen": -0.19665074348449707,
      "eval_rewards/margins": 0.8084924817085266,
      "eval_rewards/rejected": -1.005143165588379,
      "eval_runtime": 98.3147,
      "eval_samples_per_second": 20.343,
      "eval_steps_per_second": 0.325,
      "step": 100
    },
    {
      "epoch": 0.13,
      "grad_norm": 108.17707201404545,
      "learning_rate": 4.989308132738126e-07,
      "logits/chosen": -1.8153671026229858,
      "logits/rejected": -1.6921818256378174,
      "logps/chosen": -244.46237182617188,
      "logps/rejected": -222.83145141601562,
      "loss": 0.4368,
      "rewards/accuracies": 0.7875000238418579,
      "rewards/chosen": -0.10259479284286499,
      "rewards/margins": 1.188932180404663,
      "rewards/rejected": -1.2915267944335938,
      "step": 110
    },
    {
      "epoch": 0.14,
      "grad_norm": 100.40422750465044,
      "learning_rate": 4.978006327248536e-07,
      "logits/chosen": -1.7503843307495117,
      "logits/rejected": -1.6768817901611328,
      "logps/chosen": -247.3853759765625,
      "logps/rejected": -240.05496215820312,
      "loss": 0.4237,
      "rewards/accuracies": 0.8125,
      "rewards/chosen": -0.34009289741516113,
      "rewards/margins": 1.1859456300735474,
      "rewards/rejected": -1.526038408279419,
      "step": 120
    },
    {
      "epoch": 0.15,
      "grad_norm": 91.1070676572417,
      "learning_rate": 4.962695471250032e-07,
      "logits/chosen": -1.654911756515503,
      "logits/rejected": -1.62875235080719,
      "logps/chosen": -254.4684600830078,
      "logps/rejected": -245.8209686279297,
      "loss": 0.4249,
      "rewards/accuracies": 0.8125,
      "rewards/chosen": -0.5070122480392456,
      "rewards/margins": 1.3276703357696533,
      "rewards/rejected": -1.8346824645996094,
      "step": 130
    },
    {
      "epoch": 0.16,
      "grad_norm": 97.14051346245,
      "learning_rate": 4.94340033546025e-07,
      "logits/chosen": -1.6183685064315796,
      "logits/rejected": -1.6242666244506836,
      "logps/chosen": -220.20852661132812,
      "logps/rejected": -242.1243133544922,
      "loss": 0.4218,
      "rewards/accuracies": 0.8062499761581421,
      "rewards/chosen": -0.6856725811958313,
      "rewards/margins": 1.2277655601501465,
      "rewards/rejected": -1.913438081741333,
      "step": 140
    },
    {
      "epoch": 0.17,
      "grad_norm": 94.8408241800107,
      "learning_rate": 4.920152136576705e-07,
      "logits/chosen": -1.4978981018066406,
      "logits/rejected": -1.5163871049880981,
      "logps/chosen": -251.6667938232422,
      "logps/rejected": -254.0281219482422,
      "loss": 0.4261,
      "rewards/accuracies": 0.824999988079071,
      "rewards/chosen": -0.6180266737937927,
      "rewards/margins": 1.381075143814087,
      "rewards/rejected": -1.9991016387939453,
      "step": 150
    },
    {
      "epoch": 0.18,
      "grad_norm": 108.14894780962342,
      "learning_rate": 4.892988486772756e-07,
      "logits/chosen": -1.579886794090271,
      "logits/rejected": -1.5718797445297241,
      "logps/chosen": -250.04495239257812,
      "logps/rejected": -262.3863830566406,
      "loss": 0.4246,
      "rewards/accuracies": 0.824999988079071,
      "rewards/chosen": -0.6856819987297058,
      "rewards/margins": 1.5999362468719482,
      "rewards/rejected": -2.285618305206299,
      "step": 160
    },
    {
      "epoch": 0.2,
      "grad_norm": 70.9507948612333,
      "learning_rate": 4.861953332846629e-07,
      "logits/chosen": -1.7066303491592407,
      "logits/rejected": -1.6715686321258545,
      "logps/chosen": -270.17132568359375,
      "logps/rejected": -260.03350830078125,
      "loss": 0.4316,
      "rewards/accuracies": 0.7875000238418579,
      "rewards/chosen": -0.8634172677993774,
      "rewards/margins": 1.3165854215621948,
      "rewards/rejected": -2.1800026893615723,
      "step": 170
    },
    {
      "epoch": 0.21,
      "grad_norm": 114.85174445864124,
      "learning_rate": 4.827096885121953e-07,
      "logits/chosen": -1.817678689956665,
      "logits/rejected": -1.7374699115753174,
      "logps/chosen": -286.0107116699219,
      "logps/rejected": -279.7050476074219,
      "loss": 0.4309,
      "rewards/accuracies": 0.75,
      "rewards/chosen": -1.0613857507705688,
      "rewards/margins": 1.1405597925186157,
      "rewards/rejected": -2.2019453048706055,
      "step": 180
    },
    {
      "epoch": 0.22,
      "grad_norm": 90.50728576001431,
      "learning_rate": 4.788475536214821e-07,
      "logits/chosen": -1.7829034328460693,
      "logits/rejected": -1.7400600910186768,
      "logps/chosen": -236.38650512695312,
      "logps/rejected": -245.64291381835938,
      "loss": 0.3873,
      "rewards/accuracies": 0.84375,
      "rewards/chosen": -0.8258365392684937,
      "rewards/margins": 1.4504865407943726,
      "rewards/rejected": -2.276322841644287,
      "step": 190
    },
    {
      "epoch": 0.23,
      "grad_norm": 100.41476084088063,
      "learning_rate": 4.746151769798818e-07,
      "logits/chosen": -1.7743650674819946,
      "logits/rejected": -1.7203725576400757,
      "logps/chosen": -279.8271789550781,
      "logps/rejected": -269.1463317871094,
      "loss": 0.4277,
      "rewards/accuracies": 0.831250011920929,
      "rewards/chosen": -0.8149574398994446,
      "rewards/margins": 1.5317128896713257,
      "rewards/rejected": -2.346670389175415,
      "step": 200
    },
    {
      "epoch": 0.23,
      "eval_logits/chosen": -1.8008995056152344,
      "eval_logits/rejected": -1.7248116731643677,
      "eval_logps/chosen": -345.1794738769531,
      "eval_logps/rejected": -370.77880859375,
      "eval_loss": 0.46549829840660095,
      "eval_rewards/accuracies": 0.73828125,
      "eval_rewards/chosen": -0.4833980202674866,
      "eval_rewards/margins": 1.4002480506896973,
      "eval_rewards/rejected": -1.8836462497711182,
      "eval_runtime": 98.0211,
      "eval_samples_per_second": 20.404,
      "eval_steps_per_second": 0.326,
      "step": 200
    },
    {
      "epoch": 0.24,
      "grad_norm": 99.20354803224151,
      "learning_rate": 4.7001940595156055e-07,
      "logits/chosen": -1.7606436014175415,
      "logits/rejected": -1.6915203332901,
      "logps/chosen": -232.93832397460938,
      "logps/rejected": -246.24072265625,
      "loss": 0.441,
      "rewards/accuracies": 0.8062499761581421,
      "rewards/chosen": -1.0429778099060059,
      "rewards/margins": 1.6135696172714233,
      "rewards/rejected": -2.6565470695495605,
      "step": 210
    },
    {
      "epoch": 0.25,
      "grad_norm": 80.49166768353011,
      "learning_rate": 4.650676758194623e-07,
      "logits/chosen": -1.820059061050415,
      "logits/rejected": -1.741408348083496,
      "logps/chosen": -265.69769287109375,
      "logps/rejected": -259.860595703125,
      "loss": 0.4003,
      "rewards/accuracies": 0.793749988079071,
      "rewards/chosen": -0.915117621421814,
      "rewards/margins": 1.692671537399292,
      "rewards/rejected": -2.6077892780303955,
      "step": 220
    },
    {
      "epoch": 0.26,
      "grad_norm": 103.96889607439884,
      "learning_rate": 4.5976799775611215e-07,
      "logits/chosen": -1.84983229637146,
      "logits/rejected": -1.7714653015136719,
      "logps/chosen": -260.39678955078125,
      "logps/rejected": -254.94235229492188,
      "loss": 0.4118,
      "rewards/accuracies": 0.8812500238418579,
      "rewards/chosen": -1.0704491138458252,
      "rewards/margins": 1.982627511024475,
      "rewards/rejected": -3.05307674407959,
      "step": 230
    },
    {
      "epoch": 0.28,
      "grad_norm": 87.59607388393054,
      "learning_rate": 4.5412894586271543e-07,
      "logits/chosen": -1.8541374206542969,
      "logits/rejected": -1.7867300510406494,
      "logps/chosen": -259.6327819824219,
      "logps/rejected": -233.4642333984375,
      "loss": 0.3961,
      "rewards/accuracies": 0.831250011920929,
      "rewards/chosen": -0.8504000902175903,
      "rewards/margins": 1.6259987354278564,
      "rewards/rejected": -2.4763987064361572,
      "step": 240
    },
    {
      "epoch": 0.29,
      "grad_norm": 87.76098428515671,
      "learning_rate": 4.481596432975201e-07,
      "logits/chosen": -1.8068411350250244,
      "logits/rejected": -1.7720015048980713,
      "logps/chosen": -218.69546508789062,
      "logps/rejected": -231.09292602539062,
      "loss": 0.4236,
      "rewards/accuracies": 0.762499988079071,
      "rewards/chosen": -0.8816181421279907,
      "rewards/margins": 1.4680168628692627,
      "rewards/rejected": -2.349634885787964,
      "step": 250
    },
    {
      "epoch": 0.3,
      "grad_norm": 99.04429789846779,
      "learning_rate": 4.41869747515886e-07,
      "logits/chosen": -1.7596534490585327,
      "logits/rejected": -1.7420539855957031,
      "logps/chosen": -272.49273681640625,
      "logps/rejected": -301.8647766113281,
      "loss": 0.3824,
      "rewards/accuracies": 0.8125,
      "rewards/chosen": -1.0995886325836182,
      "rewards/margins": 1.7163059711456299,
      "rewards/rejected": -2.815894603729248,
      "step": 260
    },
    {
      "epoch": 0.31,
      "grad_norm": 91.17380836071179,
      "learning_rate": 4.352694346459396e-07,
      "logits/chosen": -1.7119308710098267,
      "logits/rejected": -1.713330626487732,
      "logps/chosen": -245.8417205810547,
      "logps/rejected": -257.9246826171875,
      "loss": 0.3928,
      "rewards/accuracies": 0.8062499761581421,
      "rewards/chosen": -0.6780133843421936,
      "rewards/margins": 1.5433403253555298,
      "rewards/rejected": -2.221353769302368,
      "step": 270
    },
    {
      "epoch": 0.32,
      "grad_norm": 85.89753055245001,
      "learning_rate": 4.2836938302509256e-07,
      "logits/chosen": -1.7531852722167969,
      "logits/rejected": -1.6354246139526367,
      "logps/chosen": -243.12454223632812,
      "logps/rejected": -249.15576171875,
      "loss": 0.4123,
      "rewards/accuracies": 0.8125,
      "rewards/chosen": -0.8161875605583191,
      "rewards/margins": 1.92093825340271,
      "rewards/rejected": -2.737125873565674,
      "step": 280
    },
    {
      "epoch": 0.33,
      "grad_norm": 93.29457537220665,
      "learning_rate": 4.2118075592405874e-07,
      "logits/chosen": -1.6761655807495117,
      "logits/rejected": -1.668341040611267,
      "logps/chosen": -263.28924560546875,
      "logps/rejected": -276.89190673828125,
      "loss": 0.3899,
      "rewards/accuracies": 0.8062499761581421,
      "rewards/chosen": -0.7706686854362488,
      "rewards/margins": 1.6867666244506836,
      "rewards/rejected": -2.457435131072998,
      "step": 290
    },
    {
      "epoch": 0.35,
      "grad_norm": 104.22663391151875,
      "learning_rate": 4.137151834863213e-07,
      "logits/chosen": -1.5810635089874268,
      "logits/rejected": -1.56969153881073,
      "logps/chosen": -215.5782012939453,
      "logps/rejected": -242.8097381591797,
      "loss": 0.4188,
      "rewards/accuracies": 0.7562500238418579,
      "rewards/chosen": -1.0694911479949951,
      "rewards/margins": 1.6282224655151367,
      "rewards/rejected": -2.6977133750915527,
      "step": 300
    },
    {
      "epoch": 0.35,
      "eval_logits/chosen": -1.6898695230484009,
      "eval_logits/rejected": -1.6142553091049194,
      "eval_logps/chosen": -336.9513244628906,
      "eval_logps/rejected": -373.6328430175781,
      "eval_loss": 0.39220622181892395,
      "eval_rewards/accuracies": 0.796875,
      "eval_rewards/chosen": -0.07198946177959442,
      "eval_rewards/margins": 1.9543578624725342,
      "eval_rewards/rejected": -2.0263473987579346,
      "eval_runtime": 98.0591,
      "eval_samples_per_second": 20.396,
      "eval_steps_per_second": 0.326,
      "step": 300
    },
    {
      "epoch": 0.36,
      "grad_norm": 83.12181184953562,
      "learning_rate": 4.059847439122671e-07,
      "logits/chosen": -1.7229493856430054,
      "logits/rejected": -1.590850591659546,
      "logps/chosen": -260.96539306640625,
      "logps/rejected": -262.43865966796875,
      "loss": 0.4106,
      "rewards/accuracies": 0.800000011920929,
      "rewards/chosen": -1.0447856187820435,
      "rewards/margins": 1.545778512954712,
      "rewards/rejected": -2.590564250946045,
      "step": 310
    },
    {
      "epoch": 0.37,
      "grad_norm": 101.57242075352875,
      "learning_rate": 3.98001943918432e-07,
      "logits/chosen": -1.6267013549804688,
      "logits/rejected": -1.5495989322662354,
      "logps/chosen": -246.096923828125,
      "logps/rejected": -268.87274169921875,
      "loss": 0.3937,
      "rewards/accuracies": 0.824999988079071,
      "rewards/chosen": -1.1101688146591187,
      "rewards/margins": 1.6341445446014404,
      "rewards/rejected": -2.7443130016326904,
      "step": 320
    },
    {
      "epoch": 0.38,
      "grad_norm": 95.29775137930592,
      "learning_rate": 3.8977969850346866e-07,
      "logits/chosen": -1.5687649250030518,
      "logits/rejected": -1.5476423501968384,
      "logps/chosen": -262.67156982421875,
      "logps/rejected": -260.74725341796875,
      "loss": 0.3887,
      "rewards/accuracies": 0.768750011920929,
      "rewards/chosen": -0.9716947674751282,
      "rewards/margins": 1.6578747034072876,
      "rewards/rejected": -2.6295692920684814,
      "step": 330
    },
    {
      "epoch": 0.39,
      "grad_norm": 105.61150036189296,
      "learning_rate": 3.8133131005357465e-07,
      "logits/chosen": -1.662001371383667,
      "logits/rejected": -1.5881233215332031,
      "logps/chosen": -249.8246612548828,
      "logps/rejected": -272.1719970703125,
      "loss": 0.38,
      "rewards/accuracies": 0.862500011920929,
      "rewards/chosen": -0.9675191044807434,
      "rewards/margins": 2.143432140350342,
      "rewards/rejected": -3.1109511852264404,
      "step": 340
    },
    {
      "epoch": 0.4,
      "grad_norm": 90.41304599394526,
      "learning_rate": 3.7267044682118435e-07,
      "logits/chosen": -1.6518446207046509,
      "logits/rejected": -1.6180702447891235,
      "logps/chosen": -237.4104461669922,
      "logps/rejected": -247.13473510742188,
      "loss": 0.3749,
      "rewards/accuracies": 0.824999988079071,
      "rewards/chosen": -1.1348434686660767,
      "rewards/margins": 1.6873624324798584,
      "rewards/rejected": -2.8222060203552246,
      "step": 350
    },
    {
      "epoch": 0.41,
      "grad_norm": 95.16076600849088,
      "learning_rate": 3.638111208117425e-07,
      "logits/chosen": -1.7075703144073486,
      "logits/rejected": -1.6682260036468506,
      "logps/chosen": -247.7568817138672,
      "logps/rejected": -263.2861633300781,
      "loss": 0.3914,
      "rewards/accuracies": 0.7250000238418579,
      "rewards/chosen": -1.0359086990356445,
      "rewards/margins": 1.2954949140548706,
      "rewards/rejected": -2.3314034938812256,
      "step": 360
    },
    {
      "epoch": 0.43,
      "grad_norm": 96.37188736705814,
      "learning_rate": 3.5476766511433605e-07,
      "logits/chosen": -1.7550392150878906,
      "logits/rejected": -1.6907631158828735,
      "logps/chosen": -281.13836669921875,
      "logps/rejected": -266.0453796386719,
      "loss": 0.4097,
      "rewards/accuracies": 0.7875000238418579,
      "rewards/chosen": -1.0189838409423828,
      "rewards/margins": 1.6438014507293701,
      "rewards/rejected": -2.662785291671753,
      "step": 370
    },
    {
      "epoch": 0.44,
      "grad_norm": 101.71059089383448,
      "learning_rate": 3.455547107128602e-07,
      "logits/chosen": -1.7938178777694702,
      "logits/rejected": -1.787418007850647,
      "logps/chosen": -298.5751647949219,
      "logps/rejected": -281.03118896484375,
      "loss": 0.3611,
      "rewards/accuracies": 0.831250011920929,
      "rewards/chosen": -1.0122178792953491,
      "rewards/margins": 2.0528714656829834,
      "rewards/rejected": -3.065088987350464,
      "step": 380
    },
    {
      "epoch": 0.45,
      "grad_norm": 109.30188938122785,
      "learning_rate": 3.361871628152338e-07,
      "logits/chosen": -1.8471622467041016,
      "logits/rejected": -1.820481300354004,
      "logps/chosen": -251.8716278076172,
      "logps/rejected": -283.156982421875,
      "loss": 0.4006,
      "rewards/accuracies": 0.800000011920929,
      "rewards/chosen": -1.0065103769302368,
      "rewards/margins": 1.774526834487915,
      "rewards/rejected": -2.7810370922088623,
      "step": 390
    },
    {
      "epoch": 0.46,
      "grad_norm": 89.15033820155391,
      "learning_rate": 3.2668017673896077e-07,
      "logits/chosen": -1.8275989294052124,
      "logits/rejected": -1.8090099096298218,
      "logps/chosen": -244.0500946044922,
      "logps/rejected": -245.302490234375,
      "loss": 0.3506,
      "rewards/accuracies": 0.800000011920929,
      "rewards/chosen": -0.62747722864151,
      "rewards/margins": 2.005857467651367,
      "rewards/rejected": -2.6333346366882324,
      "step": 400
    },
    {
      "epoch": 0.46,
      "eval_logits/chosen": -2.029554843902588,
      "eval_logits/rejected": -1.9793704748153687,
      "eval_logps/chosen": -331.169189453125,
      "eval_logps/rejected": -374.0495300292969,
      "eval_loss": 0.3456653654575348,
      "eval_rewards/accuracies": 0.8203125,
      "eval_rewards/chosen": 0.217118039727211,
      "eval_rewards/margins": 2.2643015384674072,
      "eval_rewards/rejected": -2.0471832752227783,
      "eval_runtime": 97.9966,
      "eval_samples_per_second": 20.409,
      "eval_steps_per_second": 0.327,
      "step": 400
    },
    {
      "epoch": 0.47,
      "grad_norm": 95.54284470696203,
      "learning_rate": 3.1704913339205103e-07,
      "logits/chosen": -1.9682222604751587,
      "logits/rejected": -1.9323310852050781,
      "logps/chosen": -251.9891357421875,
      "logps/rejected": -276.3076477050781,
      "loss": 0.3841,
      "rewards/accuracies": 0.831250011920929,
      "rewards/chosen": -0.7251254320144653,
      "rewards/margins": 1.8693323135375977,
      "rewards/rejected": -2.5944573879241943,
      "step": 410
    },
    {
      "epoch": 0.48,
      "grad_norm": 84.89194807368709,
      "learning_rate": 3.0730961438896885e-07,
      "logits/chosen": -1.973587989807129,
      "logits/rejected": -1.9758260250091553,
      "logps/chosen": -326.0084533691406,
      "logps/rejected": -316.46380615234375,
      "loss": 0.3676,
      "rewards/accuracies": 0.8187500238418579,
      "rewards/chosen": -1.2638908624649048,
      "rewards/margins": 1.681646704673767,
      "rewards/rejected": -2.945537805557251,
      "step": 420
    },
    {
      "epoch": 0.5,
      "grad_norm": 79.45330669787697,
      "learning_rate": 2.9747737684186795e-07,
      "logits/chosen": -1.9824804067611694,
      "logits/rejected": -2.0262296199798584,
      "logps/chosen": -253.5684814453125,
      "logps/rejected": -258.55352783203125,
      "loss": 0.3861,
      "rewards/accuracies": 0.8125,
      "rewards/chosen": -0.8729881048202515,
      "rewards/margins": 1.7805559635162354,
      "rewards/rejected": -2.6535439491271973,
      "step": 430
    },
    {
      "epoch": 0.51,
      "grad_norm": 92.0142521630137,
      "learning_rate": 2.8756832786789663e-07,
      "logits/chosen": -2.008882761001587,
      "logits/rejected": -1.9781932830810547,
      "logps/chosen": -269.37042236328125,
      "logps/rejected": -264.4306335449219,
      "loss": 0.382,
      "rewards/accuracies": 0.8187500238418579,
      "rewards/chosen": -0.5206736326217651,
      "rewards/margins": 1.916638970375061,
      "rewards/rejected": -2.437312364578247,
      "step": 440
    },
    {
      "epoch": 0.52,
      "grad_norm": 81.53843237959488,
      "learning_rate": 2.7759849885381747e-07,
      "logits/chosen": -1.951061487197876,
      "logits/rejected": -1.9269872903823853,
      "logps/chosen": -282.7742614746094,
      "logps/rejected": -261.77130126953125,
      "loss": 0.3613,
      "rewards/accuracies": 0.824999988079071,
      "rewards/chosen": -0.7152455449104309,
      "rewards/margins": 2.0386595726013184,
      "rewards/rejected": -2.7539050579071045,
      "step": 450
    },
    {
      "epoch": 0.53,
      "grad_norm": 81.29719600574327,
      "learning_rate": 2.675840195195762e-07,
      "logits/chosen": -1.935346007347107,
      "logits/rejected": -1.8713289499282837,
      "logps/chosen": -237.9385528564453,
      "logps/rejected": -262.90496826171875,
      "loss": 0.3768,
      "rewards/accuracies": 0.84375,
      "rewards/chosen": -0.7526956796646118,
      "rewards/margins": 2.036323070526123,
      "rewards/rejected": -2.7890188694000244,
      "step": 460
    },
    {
      "epoch": 0.54,
      "grad_norm": 79.41069847067702,
      "learning_rate": 2.575410918227829e-07,
      "logits/chosen": -1.9017337560653687,
      "logits/rejected": -1.8612467050552368,
      "logps/chosen": -286.1165771484375,
      "logps/rejected": -286.47576904296875,
      "loss": 0.3801,
      "rewards/accuracies": 0.7875000238418579,
      "rewards/chosen": -0.7328876256942749,
      "rewards/margins": 1.7589473724365234,
      "rewards/rejected": -2.491835117340088,
      "step": 470
    },
    {
      "epoch": 0.55,
      "grad_norm": 84.31202005756423,
      "learning_rate": 2.474859637463226e-07,
      "logits/chosen": -1.7930266857147217,
      "logits/rejected": -1.6920406818389893,
      "logps/chosen": -261.0568542480469,
      "logps/rejected": -249.1828155517578,
      "loss": 0.373,
      "rewards/accuracies": 0.8125,
      "rewards/chosen": -0.6825451254844666,
      "rewards/margins": 2.1065242290496826,
      "rewards/rejected": -2.789069414138794,
      "step": 480
    },
    {
      "epoch": 0.56,
      "grad_norm": 85.77456711320595,
      "learning_rate": 2.3743490301150355e-07,
      "logits/chosen": -1.8167390823364258,
      "logits/rejected": -1.6751165390014648,
      "logps/chosen": -254.97879028320312,
      "logps/rejected": -258.82794189453125,
      "loss": 0.3847,
      "rewards/accuracies": 0.8374999761581421,
      "rewards/chosen": -0.421345055103302,
      "rewards/margins": 2.164630174636841,
      "rewards/rejected": -2.585975408554077,
      "step": 490
    },
    {
      "epoch": 0.58,
      "grad_norm": 80.63419673971578,
      "learning_rate": 2.274041707592724e-07,
      "logits/chosen": -1.7511818408966064,
      "logits/rejected": -1.744807481765747,
      "logps/chosen": -238.7528839111328,
      "logps/rejected": -277.8020935058594,
      "loss": 0.3611,
      "rewards/accuracies": 0.831250011920929,
      "rewards/chosen": -1.161329984664917,
      "rewards/margins": 2.127523422241211,
      "rewards/rejected": -3.288853406906128,
      "step": 500
    },
    {
      "epoch": 0.58,
      "eval_logits/chosen": -1.8592296838760376,
      "eval_logits/rejected": -1.818342685699463,
      "eval_logps/chosen": -330.5163879394531,
      "eval_logps/rejected": -381.7996520996094,
      "eval_loss": 0.29585352540016174,
      "eval_rewards/accuracies": 0.8515625,
      "eval_rewards/chosen": 0.24975742399692535,
      "eval_rewards/margins": 2.68444561958313,
      "eval_rewards/rejected": -2.4346883296966553,
      "eval_runtime": 97.917,
      "eval_samples_per_second": 20.425,
      "eval_steps_per_second": 0.327,
      "step": 500
    },
    {
      "epoch": 0.59,
      "grad_norm": 93.97823403064545,
      "learning_rate": 2.17409995242075e-07,
      "logits/chosen": -1.8619616031646729,
      "logits/rejected": -1.8197988271713257,
      "logps/chosen": -246.95120239257812,
      "logps/rejected": -251.1341552734375,
      "loss": 0.3787,
      "rewards/accuracies": 0.831250011920929,
      "rewards/chosen": -0.9560788869857788,
      "rewards/margins": 2.1973986625671387,
      "rewards/rejected": -3.153477668762207,
      "step": 510
    },
    {
      "epoch": 0.6,
      "grad_norm": 81.29931903371055,
      "learning_rate": 2.0746854556892544e-07,
      "logits/chosen": -1.8456933498382568,
      "logits/rejected": -1.8626302480697632,
      "logps/chosen": -228.947265625,
      "logps/rejected": -255.99887084960938,
      "loss": 0.383,
      "rewards/accuracies": 0.7562500238418579,
      "rewards/chosen": -0.850503146648407,
      "rewards/margins": 1.7254650592803955,
      "rewards/rejected": -2.5759682655334473,
      "step": 520
    },
    {
      "epoch": 0.61,
      "grad_norm": 82.34962963997694,
      "learning_rate": 1.9759590554616173e-07,
      "logits/chosen": -1.9397910833358765,
      "logits/rejected": -1.9602453708648682,
      "logps/chosen": -259.236328125,
      "logps/rejected": -268.33233642578125,
      "loss": 0.3812,
      "rewards/accuracies": 0.84375,
      "rewards/chosen": -0.5990003347396851,
      "rewards/margins": 1.7735519409179688,
      "rewards/rejected": -2.3725523948669434,
      "step": 530
    },
    {
      "epoch": 0.62,
      "grad_norm": 106.09742448074628,
      "learning_rate": 1.8780804765620746e-07,
      "logits/chosen": -1.8515198230743408,
      "logits/rejected": -1.8524954319000244,
      "logps/chosen": -268.37164306640625,
      "logps/rejected": -304.6097106933594,
      "loss": 0.3726,
      "rewards/accuracies": 0.78125,
      "rewards/chosen": -0.32395535707473755,
      "rewards/margins": 1.7947556972503662,
      "rewards/rejected": -2.118710994720459,
      "step": 540
    },
    {
      "epoch": 0.63,
      "grad_norm": 89.06639468347461,
      "learning_rate": 1.7812080721643973e-07,
      "logits/chosen": -1.8575937747955322,
      "logits/rejected": -1.761614441871643,
      "logps/chosen": -263.5863037109375,
      "logps/rejected": -246.20455932617188,
      "loss": 0.3972,
      "rewards/accuracies": 0.8500000238418579,
      "rewards/chosen": -0.7359617352485657,
      "rewards/margins": 1.8330894708633423,
      "rewards/rejected": -2.5690512657165527,
      "step": 550
    },
    {
      "epoch": 0.64,
      "grad_norm": 91.51640548835455,
      "learning_rate": 1.6854985675997063e-07,
      "logits/chosen": -1.8521220684051514,
      "logits/rejected": -1.8151109218597412,
      "logps/chosen": -261.0481872558594,
      "logps/rejected": -272.9272155761719,
      "loss": 0.3641,
      "rewards/accuracies": 0.78125,
      "rewards/chosen": -0.9162181615829468,
      "rewards/margins": 1.6873579025268555,
      "rewards/rejected": -2.603576183319092,
      "step": 560
    },
    {
      "epoch": 0.66,
      "grad_norm": 83.89007997657542,
      "learning_rate": 1.5911068067978818e-07,
      "logits/chosen": -1.7626062631607056,
      "logits/rejected": -1.7946131229400635,
      "logps/chosen": -240.85189819335938,
      "logps/rejected": -273.8741149902344,
      "loss": 0.3514,
      "rewards/accuracies": 0.768750011920929,
      "rewards/chosen": -0.9277374148368835,
      "rewards/margins": 1.9961919784545898,
      "rewards/rejected": -2.923929214477539,
      "step": 570
    },
    {
      "epoch": 0.67,
      "grad_norm": 103.64429212789244,
      "learning_rate": 1.4981855017728197e-07,
      "logits/chosen": -1.8487884998321533,
      "logits/rejected": -1.8991634845733643,
      "logps/chosen": -267.7337341308594,
      "logps/rejected": -304.248291015625,
      "loss": 0.37,
      "rewards/accuracies": 0.8125,
      "rewards/chosen": -1.1558853387832642,
      "rewards/margins": 1.9654242992401123,
      "rewards/rejected": -3.121309757232666,
      "step": 580
    },
    {
      "epoch": 0.68,
      "grad_norm": 87.97513457883551,
      "learning_rate": 1.406884985556804e-07,
      "logits/chosen": -1.8694698810577393,
      "logits/rejected": -1.844276785850525,
      "logps/chosen": -257.92413330078125,
      "logps/rejected": -268.24462890625,
      "loss": 0.3586,
      "rewards/accuracies": 0.8125,
      "rewards/chosen": -1.2031538486480713,
      "rewards/margins": 2.1714892387390137,
      "rewards/rejected": -3.374642848968506,
      "step": 590
    },
    {
      "epoch": 0.69,
      "grad_norm": 85.79715679551722,
      "learning_rate": 1.3173529689837354e-07,
      "logits/chosen": -1.8827781677246094,
      "logits/rejected": -1.8290605545043945,
      "logps/chosen": -256.6871032714844,
      "logps/rejected": -256.0238952636719,
      "loss": 0.3562,
      "rewards/accuracies": 0.800000011920929,
      "rewards/chosen": -1.1103084087371826,
      "rewards/margins": 1.9665857553482056,
      "rewards/rejected": -3.0768942832946777,
      "step": 600
    },
    {
      "epoch": 0.69,
      "eval_logits/chosen": -1.9736415147781372,
      "eval_logits/rejected": -1.9216868877410889,
      "eval_logps/chosen": -327.7752685546875,
      "eval_logps/rejected": -382.5696105957031,
      "eval_loss": 0.2513369023799896,
      "eval_rewards/accuracies": 0.87109375,
      "eval_rewards/chosen": 0.386812299489975,
      "eval_rewards/margins": 2.8599982261657715,
      "eval_rewards/rejected": -2.4731857776641846,
      "eval_runtime": 97.9427,
      "eval_samples_per_second": 20.42,
      "eval_steps_per_second": 0.327,
      "step": 600
    },
    {
      "epoch": 0.7,
      "grad_norm": 91.70310566548932,
      "learning_rate": 1.2297343017146726e-07,
      "logits/chosen": -1.8489186763763428,
      "logits/rejected": -1.8288800716400146,
      "logps/chosen": -259.6686096191406,
      "logps/rejected": -267.1070861816406,
      "loss": 0.3693,
      "rewards/accuracies": 0.8500000238418579,
      "rewards/chosen": -1.0914928913116455,
      "rewards/margins": 1.885284662246704,
      "rewards/rejected": -2.9767773151397705,
      "step": 610
    },
    {
      "epoch": 0.71,
      "grad_norm": 83.02991328479807,
      "learning_rate": 1.1441707378923474e-07,
      "logits/chosen": -1.950595498085022,
      "logits/rejected": -1.8974205255508423,
      "logps/chosen": -237.6828155517578,
      "logps/rejected": -257.06005859375,
      "loss": 0.3625,
      "rewards/accuracies": 0.862500011920929,
      "rewards/chosen": -0.4481216371059418,
      "rewards/margins": 2.152703046798706,
      "rewards/rejected": -2.600825071334839,
      "step": 620
    },
    {
      "epoch": 0.73,
      "grad_norm": 97.69736464926112,
      "learning_rate": 1.06080070680377e-07,
      "logits/chosen": -1.874447226524353,
      "logits/rejected": -1.8658256530761719,
      "logps/chosen": -269.42034912109375,
      "logps/rejected": -275.75946044921875,
      "loss": 0.3679,
      "rewards/accuracies": 0.862500011920929,
      "rewards/chosen": -0.6408880949020386,
      "rewards/margins": 2.110172986984253,
      "rewards/rejected": -2.751060962677002,
      "step": 630
    },
    {
      "epoch": 0.74,
      "grad_norm": 89.55100518815718,
      "learning_rate": 9.797590889219587e-08,
      "logits/chosen": -1.927781343460083,
      "logits/rejected": -1.915203332901001,
      "logps/chosen": -262.8731384277344,
      "logps/rejected": -265.70404052734375,
      "loss": 0.3725,
      "rewards/accuracies": 0.8374999761581421,
      "rewards/chosen": -0.5387133359909058,
      "rewards/margins": 2.1157500743865967,
      "rewards/rejected": -2.654463529586792,
      "step": 640
    },
    {
      "epoch": 0.75,
      "grad_norm": 88.47283263403602,
      "learning_rate": 9.011769976891367e-08,
      "logits/chosen": -1.9129142761230469,
      "logits/rejected": -1.9037758111953735,
      "logps/chosen": -253.7337188720703,
      "logps/rejected": -273.1281433105469,
      "loss": 0.3654,
      "rewards/accuracies": 0.8187500238418579,
      "rewards/chosen": -0.8051580190658569,
      "rewards/margins": 1.8512630462646484,
      "rewards/rejected": -2.656421184539795,
      "step": 650
    },
    {
      "epoch": 0.76,
      "grad_norm": 101.63563998101947,
      "learning_rate": 8.251815673944218e-08,
      "logits/chosen": -1.8455785512924194,
      "logits/rejected": -1.862540602684021,
      "logps/chosen": -269.90655517578125,
      "logps/rejected": -266.5269470214844,
      "loss": 0.3636,
      "rewards/accuracies": 0.831250011920929,
      "rewards/chosen": -1.0431994199752808,
      "rewards/margins": 1.969745397567749,
      "rewards/rejected": -3.0129449367523193,
      "step": 660
    },
    {
      "epoch": 0.77,
      "grad_norm": 101.36570058059344,
      "learning_rate": 7.518957474892148e-08,
      "logits/chosen": -1.8844772577285767,
      "logits/rejected": -1.8293778896331787,
      "logps/chosen": -262.8396301269531,
      "logps/rejected": -270.7688293457031,
      "loss": 0.3566,
      "rewards/accuracies": 0.84375,
      "rewards/chosen": -1.0009379386901855,
      "rewards/margins": 2.288020133972168,
      "rewards/rejected": -3.2889580726623535,
      "step": 670
    },
    {
      "epoch": 0.78,
      "grad_norm": 108.29800746794625,
      "learning_rate": 6.814381036730274e-08,
      "logits/chosen": -1.9113209247589111,
      "logits/rejected": -1.8669430017471313,
      "logps/chosen": -248.9401092529297,
      "logps/rejected": -263.89459228515625,
      "loss": 0.3789,
      "rewards/accuracies": 0.8062499761581421,
      "rewards/chosen": -0.7384175062179565,
      "rewards/margins": 2.038583993911743,
      "rewards/rejected": -2.77700138092041,
      "step": 680
    },
    {
      "epoch": 0.79,
      "grad_norm": 81.30832233061167,
      "learning_rate": 6.139226260715872e-08,
      "logits/chosen": -1.936092734336853,
      "logits/rejected": -1.950042724609375,
      "logps/chosen": -261.84283447265625,
      "logps/rejected": -281.1523132324219,
      "loss": 0.3622,
      "rewards/accuracies": 0.8187500238418579,
      "rewards/chosen": -0.8801490068435669,
      "rewards/margins": 2.052794933319092,
      "rewards/rejected": -2.932943820953369,
      "step": 690
    },
    {
      "epoch": 0.81,
      "grad_norm": 81.87543968348014,
      "learning_rate": 5.4945854481754734e-08,
      "logits/chosen": -1.8324077129364014,
      "logits/rejected": -1.7888282537460327,
      "logps/chosen": -245.5817413330078,
      "logps/rejected": -259.8585510253906,
      "loss": 0.3624,
      "rewards/accuracies": 0.78125,
      "rewards/chosen": -0.9253555536270142,
      "rewards/margins": 2.031710624694824,
      "rewards/rejected": -2.957066059112549,
      "step": 700
    },
    {
      "epoch": 0.81,
      "eval_logits/chosen": -1.9716989994049072,
      "eval_logits/rejected": -1.930106282234192,
      "eval_logps/chosen": -322.60308837890625,
      "eval_logps/rejected": -380.21783447265625,
      "eval_loss": 0.21937939524650574,
      "eval_rewards/accuracies": 0.90625,
      "eval_rewards/chosen": 0.6454216241836548,
      "eval_rewards/margins": 3.001021385192871,
      "eval_rewards/rejected": -2.3555996417999268,
      "eval_runtime": 98.0586,
      "eval_samples_per_second": 20.396,
      "eval_steps_per_second": 0.326,
      "step": 700
    },
    {
      "epoch": 0.82,
      "grad_norm": 104.29775989655523,
      "learning_rate": 4.881501533321605e-08,
      "logits/chosen": -1.7864516973495483,
      "logits/rejected": -1.7958072423934937,
      "logps/chosen": -228.2520294189453,
      "logps/rejected": -255.38577270507812,
      "loss": 0.3399,
      "rewards/accuracies": 0.8125,
      "rewards/chosen": -0.952102780342102,
      "rewards/margins": 2.169015884399414,
      "rewards/rejected": -3.1211180686950684,
      "step": 710
    },
    {
      "epoch": 0.83,
      "grad_norm": 87.51056493700892,
      "learning_rate": 4.300966395938377e-08,
      "logits/chosen": -1.91278076171875,
      "logits/rejected": -1.887738823890686,
      "logps/chosen": -269.040283203125,
      "logps/rejected": -277.3271484375,
      "loss": 0.3737,
      "rewards/accuracies": 0.768750011920929,
      "rewards/chosen": -0.9400026202201843,
      "rewards/margins": 2.1387667655944824,
      "rewards/rejected": -3.0787696838378906,
      "step": 720
    },
    {
      "epoch": 0.84,
      "grad_norm": 82.85641235575972,
      "learning_rate": 3.7539192566655246e-08,
      "logits/chosen": -1.8706934452056885,
      "logits/rejected": -1.8632476329803467,
      "logps/chosen": -258.32257080078125,
      "logps/rejected": -262.11151123046875,
      "loss": 0.3615,
      "rewards/accuracies": 0.862500011920929,
      "rewards/chosen": -0.5498217344284058,
      "rewards/margins": 1.930101752281189,
      "rewards/rejected": -2.479923725128174,
      "step": 730
    },
    {
      "epoch": 0.85,
      "grad_norm": 86.13162915966026,
      "learning_rate": 3.24124515747731e-08,
      "logits/chosen": -1.827125906944275,
      "logits/rejected": -1.8333660364151,
      "logps/chosen": -245.1622314453125,
      "logps/rejected": -269.59857177734375,
      "loss": 0.376,
      "rewards/accuracies": 0.800000011920929,
      "rewards/chosen": -1.0357568264007568,
      "rewards/margins": 1.9602575302124023,
      "rewards/rejected": -2.996014356613159,
      "step": 740
    },
    {
      "epoch": 0.86,
      "grad_norm": 86.92913330390785,
      "learning_rate": 2.763773529814506e-08,
      "logits/chosen": -1.940159559249878,
      "logits/rejected": -1.9158337116241455,
      "logps/chosen": -282.51397705078125,
      "logps/rejected": -276.21331787109375,
      "loss": 0.372,
      "rewards/accuracies": 0.800000011920929,
      "rewards/chosen": -0.7603832483291626,
      "rewards/margins": 2.119266986846924,
      "rewards/rejected": -2.879650354385376,
      "step": 750
    },
    {
      "epoch": 0.88,
      "grad_norm": 95.56287400260709,
      "learning_rate": 2.3222768526860698e-08,
      "logits/chosen": -1.863567590713501,
      "logits/rejected": -1.788558006286621,
      "logps/chosen": -249.87838745117188,
      "logps/rejected": -264.7607727050781,
      "loss": 0.3676,
      "rewards/accuracies": 0.824999988079071,
      "rewards/chosen": -0.6970678567886353,
      "rewards/margins": 2.2210423946380615,
      "rewards/rejected": -2.9181103706359863,
      "step": 760
    },
    {
      "epoch": 0.89,
      "grad_norm": 82.6081133840389,
      "learning_rate": 1.9174694029115146e-08,
      "logits/chosen": -1.9369175434112549,
      "logits/rejected": -1.9106756448745728,
      "logps/chosen": -287.55975341796875,
      "logps/rejected": -268.37493896484375,
      "loss": 0.3621,
      "rewards/accuracies": 0.800000011920929,
      "rewards/chosen": -0.7068324685096741,
      "rewards/margins": 1.9774510860443115,
      "rewards/rejected": -2.684283494949341,
      "step": 770
    },
    {
      "epoch": 0.9,
      "grad_norm": 84.83905388303022,
      "learning_rate": 1.5500060995258134e-08,
      "logits/chosen": -1.8579909801483154,
      "logits/rejected": -1.8068830966949463,
      "logps/chosen": -258.0039367675781,
      "logps/rejected": -259.26202392578125,
      "loss": 0.3456,
      "rewards/accuracies": 0.875,
      "rewards/chosen": -0.8524407148361206,
      "rewards/margins": 2.122145414352417,
      "rewards/rejected": -2.974586248397827,
      "step": 780
    },
    {
      "epoch": 0.91,
      "grad_norm": 94.72717106858794,
      "learning_rate": 1.2204814442165812e-08,
      "logits/chosen": -1.8840267658233643,
      "logits/rejected": -1.8291581869125366,
      "logps/chosen": -252.4437713623047,
      "logps/rejected": -252.0703582763672,
      "loss": 0.3741,
      "rewards/accuracies": 0.824999988079071,
      "rewards/chosen": -0.8425480127334595,
      "rewards/margins": 2.2477831840515137,
      "rewards/rejected": -3.0903308391571045,
      "step": 790
    },
    {
      "epoch": 0.92,
      "grad_norm": 102.52909854244074,
      "learning_rate": 9.294285595075669e-09,
      "logits/chosen": -1.9619266986846924,
      "logits/rejected": -1.904088020324707,
      "logps/chosen": -277.6653747558594,
      "logps/rejected": -272.3161315917969,
      "loss": 0.4069,
      "rewards/accuracies": 0.78125,
      "rewards/chosen": -0.8278266787528992,
      "rewards/margins": 2.0403804779052734,
      "rewards/rejected": -2.868206739425659,
      "step": 800
    },
    {
      "epoch": 0.92,
      "eval_logits/chosen": -1.9591288566589355,
      "eval_logits/rejected": -1.9204463958740234,
      "eval_logps/chosen": -322.0539245605469,
      "eval_logps/rejected": -380.2658386230469,
      "eval_loss": 0.20271854102611542,
      "eval_rewards/accuracies": 0.9140625,
      "eval_rewards/chosen": 0.6728801131248474,
      "eval_rewards/margins": 3.0308780670166016,
      "eval_rewards/rejected": -2.3579981327056885,
      "eval_runtime": 97.802,
      "eval_samples_per_second": 20.449,
      "eval_steps_per_second": 0.327,
      "step": 800
    },
    {
      "epoch": 0.93,
      "grad_norm": 85.0418535767566,
      "learning_rate": 6.773183262446914e-09,
      "logits/chosen": -1.8587850332260132,
      "logits/rejected": -1.7795337438583374,
      "logps/chosen": -248.3004913330078,
      "logps/rejected": -261.2222595214844,
      "loss": 0.3824,
      "rewards/accuracies": 0.8125,
      "rewards/chosen": -0.7576299905776978,
      "rewards/margins": 2.073674440383911,
      "rewards/rejected": -2.8313040733337402,
      "step": 810
    },
    {
      "epoch": 0.94,
      "grad_norm": 88.3275931540908,
      "learning_rate": 4.645586217799452e-09,
      "logits/chosen": -1.8911056518554688,
      "logits/rejected": -1.940203309059143,
      "logps/chosen": -265.9781799316406,
      "logps/rejected": -290.08770751953125,
      "loss": 0.3902,
      "rewards/accuracies": 0.8125,
      "rewards/chosen": -0.839026927947998,
      "rewards/margins": 2.1166014671325684,
      "rewards/rejected": -2.9556286334991455,
      "step": 820
    },
    {
      "epoch": 0.96,
      "grad_norm": 108.33767167817186,
      "learning_rate": 2.9149366008568987e-09,
      "logits/chosen": -1.8930469751358032,
      "logits/rejected": -1.9079952239990234,
      "logps/chosen": -263.65264892578125,
      "logps/rejected": -278.3803405761719,
      "loss": 0.3825,
      "rewards/accuracies": 0.793749988079071,
      "rewards/chosen": -0.6759049296379089,
      "rewards/margins": 2.146210193634033,
      "rewards/rejected": -2.822114944458008,
      "step": 830
    },
    {
      "epoch": 0.97,
      "grad_norm": 86.51713108380295,
      "learning_rate": 1.5840343486700215e-09,
      "logits/chosen": -1.9469448328018188,
      "logits/rejected": -1.939854383468628,
      "logps/chosen": -281.2010498046875,
      "logps/rejected": -275.5614318847656,
      "loss": 0.367,
      "rewards/accuracies": 0.8374999761581421,
      "rewards/chosen": -0.5235612392425537,
      "rewards/margins": 2.1837284564971924,
      "rewards/rejected": -2.707289457321167,
      "step": 840
    },
    {
      "epoch": 0.98,
      "grad_norm": 82.94960696998083,
      "learning_rate": 6.550326657293881e-10,
      "logits/chosen": -1.9341312646865845,
      "logits/rejected": -1.901523232460022,
      "logps/chosen": -257.5312805175781,
      "logps/rejected": -268.84619140625,
      "loss": 0.346,
      "rewards/accuracies": 0.8812500238418579,
      "rewards/chosen": -0.7536182999610901,
      "rewards/margins": 2.499413251876831,
      "rewards/rejected": -3.2530312538146973,
      "step": 850
    },
    {
      "epoch": 0.99,
      "grad_norm": 100.59142277641878,
      "learning_rate": 1.2943454039654467e-10,
      "logits/chosen": -1.821735143661499,
      "logits/rejected": -1.835100531578064,
      "logps/chosen": -244.0499267578125,
      "logps/rejected": -261.69805908203125,
      "loss": 0.3624,
      "rewards/accuracies": 0.800000011920929,
      "rewards/chosen": -1.0147624015808105,
      "rewards/margins": 1.7580705881118774,
      "rewards/rejected": -2.7728328704833984,
      "step": 860
    },
    {
      "epoch": 1.0,
      "step": 868,
      "total_flos": 0.0,
      "train_loss": 0.40559157427005504,
      "train_runtime": 13777.3263,
      "train_samples_per_second": 8.066,
      "train_steps_per_second": 0.063
    }
  ],
  "logging_steps": 10,
  "max_steps": 868,
  "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
}