{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 1.0,
  "eval_steps": 100,
  "global_step": 385,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.0,
      "grad_norm": 78.5,
      "learning_rate": 1.282051282051282e-07,
      "logits/chosen": 88.18099975585938,
      "logits/rejected": 88.25153350830078,
      "logps/chosen": -29.073104858398438,
      "logps/rejected": -26.25731658935547,
      "loss": 1.0,
      "rewards/accuracies": 0.0,
      "rewards/chosen": 0.0,
      "rewards/margins": 0.0,
      "rewards/rejected": 0.0,
      "step": 1
    },
    {
      "epoch": 0.03,
      "grad_norm": 66.0,
      "learning_rate": 1.282051282051282e-06,
      "logits/chosen": 81.06854248046875,
      "logits/rejected": 80.77387237548828,
      "logps/chosen": -34.293479919433594,
      "logps/rejected": -32.99631118774414,
      "loss": 1.0126,
      "rewards/accuracies": 0.4305555522441864,
      "rewards/chosen": -0.03105219267308712,
      "rewards/margins": -0.008308152668178082,
      "rewards/rejected": -0.022744040936231613,
      "step": 10
    },
    {
      "epoch": 0.05,
      "grad_norm": 67.0,
      "learning_rate": 2.564102564102564e-06,
      "logits/chosen": 80.62992095947266,
      "logits/rejected": 80.52108001708984,
      "logps/chosen": -33.51418685913086,
      "logps/rejected": -30.840993881225586,
      "loss": 0.8773,
      "rewards/accuracies": 0.5625,
      "rewards/chosen": 0.10543360561132431,
      "rewards/margins": 0.14353612065315247,
      "rewards/rejected": -0.038102518767118454,
      "step": 20
    },
    {
      "epoch": 0.08,
      "grad_norm": 68.0,
      "learning_rate": 3.846153846153847e-06,
      "logits/chosen": 82.46273803710938,
      "logits/rejected": 82.4950180053711,
      "logps/chosen": -33.98906707763672,
      "logps/rejected": -31.265905380249023,
      "loss": 1.1079,
      "rewards/accuracies": 0.44999998807907104,
      "rewards/chosen": 0.014249947853386402,
      "rewards/margins": -0.08050690591335297,
      "rewards/rejected": 0.09475686401128769,
      "step": 30
    },
    {
      "epoch": 0.1,
      "grad_norm": 55.75,
      "learning_rate": 4.999896948438434e-06,
      "logits/chosen": 80.95271301269531,
      "logits/rejected": 80.94709777832031,
      "logps/chosen": -32.69623565673828,
      "logps/rejected": -33.20489501953125,
      "loss": 0.8832,
      "rewards/accuracies": 0.6000000238418579,
      "rewards/chosen": 0.2866959571838379,
      "rewards/margins": 0.1885182112455368,
      "rewards/rejected": 0.09817773848772049,
      "step": 40
    },
    {
      "epoch": 0.13,
      "grad_norm": 37.5,
      "learning_rate": 4.987541037542187e-06,
      "logits/chosen": 78.52803039550781,
      "logits/rejected": 78.53800964355469,
      "logps/chosen": -30.692174911499023,
      "logps/rejected": -30.75704002380371,
      "loss": 1.0095,
      "rewards/accuracies": 0.4375,
      "rewards/chosen": 0.30784496665000916,
      "rewards/margins": 0.1180974617600441,
      "rewards/rejected": 0.18974749743938446,
      "step": 50
    },
    {
      "epoch": 0.16,
      "grad_norm": 53.0,
      "learning_rate": 4.954691471941119e-06,
      "logits/chosen": 83.14569854736328,
      "logits/rejected": 83.19773864746094,
      "logps/chosen": -30.899333953857422,
      "logps/rejected": -29.450210571289062,
      "loss": 0.9563,
      "rewards/accuracies": 0.512499988079071,
      "rewards/chosen": 0.16549696028232574,
      "rewards/margins": 0.08130919188261032,
      "rewards/rejected": 0.08418775349855423,
      "step": 60
    },
    {
      "epoch": 0.18,
      "grad_norm": 69.5,
      "learning_rate": 4.901618883413549e-06,
      "logits/chosen": 83.82180786132812,
      "logits/rejected": 83.84942626953125,
      "logps/chosen": -30.46747398376465,
      "logps/rejected": -33.01859664916992,
      "loss": 0.9812,
      "rewards/accuracies": 0.48750001192092896,
      "rewards/chosen": 0.09693069756031036,
      "rewards/margins": 0.08546572178602219,
      "rewards/rejected": 0.011464978568255901,
      "step": 70
    },
    {
      "epoch": 0.21,
      "grad_norm": 56.25,
      "learning_rate": 4.828760511501322e-06,
      "logits/chosen": 81.54592895507812,
      "logits/rejected": 81.53177642822266,
      "logps/chosen": -31.3655948638916,
      "logps/rejected": -30.929000854492188,
      "loss": 0.9098,
      "rewards/accuracies": 0.5375000238418579,
      "rewards/chosen": 0.0836077481508255,
      "rewards/margins": 0.1864190548658371,
      "rewards/rejected": -0.102811299264431,
      "step": 80
    },
    {
      "epoch": 0.23,
      "grad_norm": 71.5,
      "learning_rate": 4.7367166013034295e-06,
      "logits/chosen": 78.40837097167969,
      "logits/rejected": 78.37776947021484,
      "logps/chosen": -32.40131378173828,
      "logps/rejected": -30.98993492126465,
      "loss": 0.913,
      "rewards/accuracies": 0.574999988079071,
      "rewards/chosen": 0.14211884140968323,
      "rewards/margins": 0.16519510746002197,
      "rewards/rejected": -0.023076282814145088,
      "step": 90
    },
    {
      "epoch": 0.26,
      "grad_norm": 69.0,
      "learning_rate": 4.626245458345211e-06,
      "logits/chosen": 83.61566162109375,
      "logits/rejected": 83.63754272460938,
      "logps/chosen": -33.954978942871094,
      "logps/rejected": -31.89337158203125,
      "loss": 0.8073,
      "rewards/accuracies": 0.612500011920929,
      "rewards/chosen": 0.21000584959983826,
      "rewards/margins": 0.29570215940475464,
      "rewards/rejected": -0.08569628745317459,
      "step": 100
    },
    {
      "epoch": 0.26,
      "eval_logits/chosen": 98.75115966796875,
      "eval_logits/rejected": 98.7406005859375,
      "eval_logps/chosen": -32.40181350708008,
      "eval_logps/rejected": -35.920406341552734,
      "eval_loss": 1.0338481664657593,
      "eval_rewards/accuracies": 0.4966777563095093,
      "eval_rewards/chosen": 0.02481812983751297,
      "eval_rewards/margins": -0.002855697413906455,
      "eval_rewards/rejected": 0.02767382375895977,
      "eval_runtime": 104.4142,
      "eval_samples_per_second": 3.285,
      "eval_steps_per_second": 0.412,
      "step": 100
    },
    {
      "epoch": 0.29,
      "grad_norm": 79.0,
      "learning_rate": 4.498257201263691e-06,
      "logits/chosen": 83.83598327636719,
      "logits/rejected": 83.71295928955078,
      "logps/chosen": -32.41645431518555,
      "logps/rejected": -32.721168518066406,
      "loss": 0.8304,
      "rewards/accuracies": 0.612500011920929,
      "rewards/chosen": 0.3638074994087219,
      "rewards/margins": 0.4029727876186371,
      "rewards/rejected": -0.03916532173752785,
      "step": 110
    },
    {
      "epoch": 0.31,
      "grad_norm": 80.5,
      "learning_rate": 4.353806263777678e-06,
      "logits/chosen": 83.92488098144531,
      "logits/rejected": 84.05035400390625,
      "logps/chosen": -28.292224884033203,
      "logps/rejected": -35.42224884033203,
      "loss": 0.7915,
      "rewards/accuracies": 0.637499988079071,
      "rewards/chosen": 0.3824160099029541,
      "rewards/margins": 0.36027655005455017,
      "rewards/rejected": 0.022139430046081543,
      "step": 120
    },
    {
      "epoch": 0.34,
      "grad_norm": 48.5,
      "learning_rate": 4.1940827077152755e-06,
      "logits/chosen": 81.21788024902344,
      "logits/rejected": 81.23727416992188,
      "logps/chosen": -30.410778045654297,
      "logps/rejected": -32.07198715209961,
      "loss": 0.828,
      "rewards/accuracies": 0.612500011920929,
      "rewards/chosen": 0.2979464530944824,
      "rewards/margins": 0.38216376304626465,
      "rewards/rejected": -0.08421732485294342,
      "step": 130
    },
    {
      "epoch": 0.36,
      "grad_norm": 62.5,
      "learning_rate": 4.0204024186666215e-06,
      "logits/chosen": 82.4339828491211,
      "logits/rejected": 82.45420837402344,
      "logps/chosen": -27.09214210510254,
      "logps/rejected": -33.011024475097656,
      "loss": 0.6935,
      "rewards/accuracies": 0.699999988079071,
      "rewards/chosen": 0.213111013174057,
      "rewards/margins": 0.5656993985176086,
      "rewards/rejected": -0.35258835554122925,
      "step": 140
    },
    {
      "epoch": 0.39,
      "grad_norm": 54.75,
      "learning_rate": 3.834196265035119e-06,
      "logits/chosen": 81.10945129394531,
      "logits/rejected": 81.082763671875,
      "logps/chosen": -28.871135711669922,
      "logps/rejected": -32.9850959777832,
      "loss": 0.6411,
      "rewards/accuracies": 0.7124999761581421,
      "rewards/chosen": 0.31078284978866577,
      "rewards/margins": 0.5619392395019531,
      "rewards/rejected": -0.25115638971328735,
      "step": 150
    },
    {
      "epoch": 0.42,
      "grad_norm": 63.25,
      "learning_rate": 3.636998309800573e-06,
      "logits/chosen": 83.00135803222656,
      "logits/rejected": 82.99880981445312,
      "logps/chosen": -33.499019622802734,
      "logps/rejected": -30.40065574645996,
      "loss": 0.6874,
      "rewards/accuracies": 0.7250000238418579,
      "rewards/chosen": 0.3826376497745514,
      "rewards/margins": 0.6515066027641296,
      "rewards/rejected": -0.26886898279190063,
      "step": 160
    },
    {
      "epoch": 0.44,
      "grad_norm": 68.0,
      "learning_rate": 3.4304331721118078e-06,
      "logits/chosen": 83.82991790771484,
      "logits/rejected": 83.77062225341797,
      "logps/chosen": -30.814950942993164,
      "logps/rejected": -32.44989776611328,
      "loss": 0.7545,
      "rewards/accuracies": 0.6875,
      "rewards/chosen": 0.2666035592556,
      "rewards/margins": 0.5245082974433899,
      "rewards/rejected": -0.2579047381877899,
      "step": 170
    },
    {
      "epoch": 0.47,
      "grad_norm": 50.25,
      "learning_rate": 3.2162026428305436e-06,
      "logits/chosen": 81.48609924316406,
      "logits/rejected": 81.4673843383789,
      "logps/chosen": -30.380239486694336,
      "logps/rejected": -31.43511962890625,
      "loss": 0.6492,
      "rewards/accuracies": 0.699999988079071,
      "rewards/chosen": 0.49014702439308167,
      "rewards/margins": 0.6568924188613892,
      "rewards/rejected": -0.16674548387527466,
      "step": 180
    },
    {
      "epoch": 0.49,
      "grad_norm": 33.5,
      "learning_rate": 2.996071664294641e-06,
      "logits/chosen": 83.15840148925781,
      "logits/rejected": 83.14761352539062,
      "logps/chosen": -30.205703735351562,
      "logps/rejected": -30.463912963867188,
      "loss": 0.9036,
      "rewards/accuracies": 0.6000000238418579,
      "rewards/chosen": 0.33642834424972534,
      "rewards/margins": 0.33396121859550476,
      "rewards/rejected": 0.0024670884013175964,
      "step": 190
    },
    {
      "epoch": 0.52,
      "grad_norm": 54.75,
      "learning_rate": 2.7718537898066833e-06,
      "logits/chosen": 78.7307357788086,
      "logits/rejected": 78.67295837402344,
      "logps/chosen": -33.62441635131836,
      "logps/rejected": -32.52568054199219,
      "loss": 0.7227,
      "rewards/accuracies": 0.637499988079071,
      "rewards/chosen": 0.6769934892654419,
      "rewards/margins": 0.6963596343994141,
      "rewards/rejected": -0.019366104155778885,
      "step": 200
    },
    {
      "epoch": 0.52,
      "eval_logits/chosen": 98.77264404296875,
      "eval_logits/rejected": 98.75276947021484,
      "eval_logps/chosen": -32.62995529174805,
      "eval_logps/rejected": -36.19297790527344,
      "eval_loss": 1.0227326154708862,
      "eval_rewards/accuracies": 0.49501657485961914,
      "eval_rewards/chosen": -0.11206817626953125,
      "eval_rewards/margins": 0.023798702284693718,
      "eval_rewards/rejected": -0.13586686551570892,
      "eval_runtime": 104.1915,
      "eval_samples_per_second": 3.292,
      "eval_steps_per_second": 0.413,
      "step": 200
    },
    {
      "epoch": 0.55,
      "grad_norm": 65.5,
      "learning_rate": 2.5453962426402006e-06,
      "logits/chosen": 81.24237823486328,
      "logits/rejected": 81.15150451660156,
      "logps/chosen": -33.17626953125,
      "logps/rejected": -35.2652587890625,
      "loss": 0.6447,
      "rewards/accuracies": 0.737500011920929,
      "rewards/chosen": 0.46167102456092834,
      "rewards/margins": 0.6291700601577759,
      "rewards/rejected": -0.1674990952014923,
      "step": 210
    },
    {
      "epoch": 0.57,
      "grad_norm": 48.5,
      "learning_rate": 2.3185646976551794e-06,
      "logits/chosen": 83.28003692626953,
      "logits/rejected": 83.37066650390625,
      "logps/chosen": -31.01590347290039,
      "logps/rejected": -31.09220314025879,
      "loss": 0.5863,
      "rewards/accuracies": 0.737500011920929,
      "rewards/chosen": 0.5598315000534058,
      "rewards/margins": 0.7777799367904663,
      "rewards/rejected": -0.21794843673706055,
      "step": 220
    },
    {
      "epoch": 0.6,
      "grad_norm": 62.75,
      "learning_rate": 2.0932279108998323e-06,
      "logits/chosen": 80.46641540527344,
      "logits/rejected": 80.52613067626953,
      "logps/chosen": -32.29460525512695,
      "logps/rejected": -34.29331588745117,
      "loss": 0.8345,
      "rewards/accuracies": 0.637499988079071,
      "rewards/chosen": 0.30668336153030396,
      "rewards/margins": 0.4753844141960144,
      "rewards/rejected": -0.16870108246803284,
      "step": 230
    },
    {
      "epoch": 0.62,
      "grad_norm": 74.0,
      "learning_rate": 1.8712423238279358e-06,
      "logits/chosen": 82.80165100097656,
      "logits/rejected": 83.07733917236328,
      "logps/chosen": -30.73067283630371,
      "logps/rejected": -31.75331687927246,
      "loss": 0.5911,
      "rewards/accuracies": 0.762499988079071,
      "rewards/chosen": 0.5744358897209167,
      "rewards/margins": 0.7072005271911621,
      "rewards/rejected": -0.13276457786560059,
      "step": 240
    },
    {
      "epoch": 0.65,
      "grad_norm": 69.0,
      "learning_rate": 1.6544367689701824e-06,
      "logits/chosen": 81.53695678710938,
      "logits/rejected": 81.60115051269531,
      "logps/chosen": -26.971792221069336,
      "logps/rejected": -30.23346519470215,
      "loss": 0.7449,
      "rewards/accuracies": 0.6499999761581421,
      "rewards/chosen": 0.3744624853134155,
      "rewards/margins": 0.5740996599197388,
      "rewards/rejected": -0.19963720440864563,
      "step": 250
    },
    {
      "epoch": 0.68,
      "grad_norm": 45.5,
      "learning_rate": 1.4445974030621963e-06,
      "logits/chosen": 78.7826156616211,
      "logits/rejected": 78.90965270996094,
      "logps/chosen": -30.57162857055664,
      "logps/rejected": -36.598594665527344,
      "loss": 0.5112,
      "rewards/accuracies": 0.7749999761581421,
      "rewards/chosen": 0.5936219692230225,
      "rewards/margins": 0.8918012380599976,
      "rewards/rejected": -0.29817938804626465,
      "step": 260
    },
    {
      "epoch": 0.7,
      "grad_norm": 38.25,
      "learning_rate": 1.243452991757889e-06,
      "logits/chosen": 78.14024353027344,
      "logits/rejected": 78.16914367675781,
      "logps/chosen": -30.831195831298828,
      "logps/rejected": -31.901092529296875,
      "loss": 0.5884,
      "rewards/accuracies": 0.7250000238418579,
      "rewards/chosen": 0.56863933801651,
      "rewards/margins": 0.8060423135757446,
      "rewards/rejected": -0.23740296065807343,
      "step": 270
    },
    {
      "epoch": 0.73,
      "grad_norm": 65.0,
      "learning_rate": 1.0526606671603523e-06,
      "logits/chosen": 80.71138000488281,
      "logits/rejected": 80.48588562011719,
      "logps/chosen": -31.13303565979004,
      "logps/rejected": -29.891042709350586,
      "loss": 0.751,
      "rewards/accuracies": 0.7250000238418579,
      "rewards/chosen": 0.4061153829097748,
      "rewards/margins": 0.540542483329773,
      "rewards/rejected": -0.13442707061767578,
      "step": 280
    },
    {
      "epoch": 0.75,
      "grad_norm": 47.75,
      "learning_rate": 8.737922755071455e-07,
      "logits/chosen": 80.93512725830078,
      "logits/rejected": 80.84117126464844,
      "logps/chosen": -32.98380661010742,
      "logps/rejected": -32.5334587097168,
      "loss": 0.4856,
      "rewards/accuracies": 0.7875000238418579,
      "rewards/chosen": 0.672263503074646,
      "rewards/margins": 0.9822207689285278,
      "rewards/rejected": -0.30995720624923706,
      "step": 290
    },
    {
      "epoch": 0.78,
      "grad_norm": 51.75,
      "learning_rate": 7.08321427484816e-07,
      "logits/chosen": 76.60159301757812,
      "logits/rejected": 76.69679260253906,
      "logps/chosen": -31.996997833251953,
      "logps/rejected": -29.191226959228516,
      "loss": 0.6355,
      "rewards/accuracies": 0.7124999761581421,
      "rewards/chosen": 0.8452256321907043,
      "rewards/margins": 0.8987467885017395,
      "rewards/rejected": -0.05352109670639038,
      "step": 300
    },
    {
      "epoch": 0.78,
      "eval_logits/chosen": 98.76406860351562,
      "eval_logits/rejected": 98.7339859008789,
      "eval_logps/chosen": -32.608062744140625,
      "eval_logps/rejected": -36.16687774658203,
      "eval_loss": 1.0328983068466187,
      "eval_rewards/accuracies": 0.5456810593605042,
      "eval_rewards/chosen": -0.09893348813056946,
      "eval_rewards/margins": 0.02127554826438427,
      "eval_rewards/rejected": -0.12020901590585709,
      "eval_runtime": 103.9872,
      "eval_samples_per_second": 3.298,
      "eval_steps_per_second": 0.414,
      "step": 300
    },
    {
      "epoch": 0.81,
      "grad_norm": 53.25,
      "learning_rate": 5.576113578589035e-07,
      "logits/chosen": 83.55947875976562,
      "logits/rejected": 83.5940170288086,
      "logps/chosen": -29.987110137939453,
      "logps/rejected": -32.35546112060547,
      "loss": 0.6477,
      "rewards/accuracies": 0.7250000238418579,
      "rewards/chosen": 0.5258544683456421,
      "rewards/margins": 0.6192765831947327,
      "rewards/rejected": -0.09342211484909058,
      "step": 310
    },
    {
      "epoch": 0.83,
      "grad_norm": 46.5,
      "learning_rate": 4.229036944380913e-07,
      "logits/chosen": 81.14942932128906,
      "logits/rejected": 81.15238952636719,
      "logps/chosen": -30.662851333618164,
      "logps/rejected": -29.089075088500977,
      "loss": 0.6327,
      "rewards/accuracies": 0.6875,
      "rewards/chosen": 0.5614250898361206,
      "rewards/margins": 0.6631743311882019,
      "rewards/rejected": -0.10174927860498428,
      "step": 320
    },
    {
      "epoch": 0.86,
      "grad_norm": 43.5,
      "learning_rate": 3.053082288996112e-07,
      "logits/chosen": 78.36723327636719,
      "logits/rejected": 78.41484832763672,
      "logps/chosen": -29.124780654907227,
      "logps/rejected": -32.84328079223633,
      "loss": 0.6057,
      "rewards/accuracies": 0.675000011920929,
      "rewards/chosen": 0.7366967797279358,
      "rewards/margins": 0.8617448806762695,
      "rewards/rejected": -0.12504807114601135,
      "step": 330
    },
    {
      "epoch": 0.88,
      "grad_norm": 65.5,
      "learning_rate": 2.0579377374915805e-07,
      "logits/chosen": 82.55430603027344,
      "logits/rejected": 82.59879302978516,
      "logps/chosen": -32.09717559814453,
      "logps/rejected": -33.71881103515625,
      "loss": 0.6103,
      "rewards/accuracies": 0.699999988079071,
      "rewards/chosen": 0.6689842343330383,
      "rewards/margins": 0.8790559768676758,
      "rewards/rejected": -0.2100718766450882,
      "step": 340
    },
    {
      "epoch": 0.91,
      "grad_norm": 45.25,
      "learning_rate": 1.2518018074041684e-07,
      "logits/chosen": 81.6875228881836,
      "logits/rejected": 81.70503234863281,
      "logps/chosen": -32.540645599365234,
      "logps/rejected": -33.16196823120117,
      "loss": 0.6663,
      "rewards/accuracies": 0.7250000238418579,
      "rewards/chosen": 0.6618569493293762,
      "rewards/margins": 0.7039581537246704,
      "rewards/rejected": -0.042101241648197174,
      "step": 350
    },
    {
      "epoch": 0.94,
      "grad_norm": 37.5,
      "learning_rate": 6.41315865106129e-08,
      "logits/chosen": 83.08977508544922,
      "logits/rejected": 83.10887145996094,
      "logps/chosen": -28.5622615814209,
      "logps/rejected": -31.776697158813477,
      "loss": 0.6199,
      "rewards/accuracies": 0.737500011920929,
      "rewards/chosen": 0.5939711332321167,
      "rewards/margins": 0.6700795292854309,
      "rewards/rejected": -0.07610837370157242,
      "step": 360
    },
    {
      "epoch": 0.96,
      "grad_norm": 69.0,
      "learning_rate": 2.3150941078050325e-08,
      "logits/chosen": 82.51141357421875,
      "logits/rejected": 82.53152465820312,
      "logps/chosen": -31.914348602294922,
      "logps/rejected": -35.504486083984375,
      "loss": 0.6338,
      "rewards/accuracies": 0.6499999761581421,
      "rewards/chosen": 0.5770877599716187,
      "rewards/margins": 0.8141579627990723,
      "rewards/rejected": -0.237070232629776,
      "step": 370
    },
    {
      "epoch": 0.99,
      "grad_norm": 52.75,
      "learning_rate": 2.575864278703266e-09,
      "logits/chosen": 76.49024963378906,
      "logits/rejected": 76.35295104980469,
      "logps/chosen": -29.67061996459961,
      "logps/rejected": -28.207141876220703,
      "loss": 0.6675,
      "rewards/accuracies": 0.7124999761581421,
      "rewards/chosen": 0.5433036088943481,
      "rewards/margins": 0.571243166923523,
      "rewards/rejected": -0.027939552441239357,
      "step": 380
    },
    {
      "epoch": 1.0,
      "step": 385,
      "total_flos": 0.0,
      "train_loss": 0.746250953302755,
      "train_runtime": 2556.1232,
      "train_samples_per_second": 1.205,
      "train_steps_per_second": 0.151
    }
  ],
  "logging_steps": 10,
  "max_steps": 385,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 1,
  "save_steps": 100,
  "total_flos": 0.0,
  "train_batch_size": 4,
  "trial_name": null,
  "trial_params": null
}