{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.9921671018276762,
  "eval_steps": 500,
  "global_step": 95,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.01,
      "grad_norm": 6.531803428424505,
      "learning_rate": 1.5000000000000002e-07,
      "logits/chosen": -2.851747512817383,
      "logits/rejected": -2.833996534347534,
      "logps/chosen": -165.70089721679688,
      "logps/rejected": -198.857666015625,
      "loss": 0.6931,
      "rewards/accuracies": 0.0,
      "rewards/chosen": 0.0,
      "rewards/margins": 0.0,
      "rewards/rejected": 0.0,
      "step": 1
    },
    {
      "epoch": 0.05,
      "grad_norm": 5.945146082388948,
      "learning_rate": 7.5e-07,
      "logits/chosen": -2.7703874111175537,
      "logits/rejected": -2.773144006729126,
      "logps/chosen": -171.2422332763672,
      "logps/rejected": -172.52732849121094,
      "loss": 0.6929,
      "rewards/accuracies": 0.3984375,
      "rewards/chosen": 0.001207518856972456,
      "rewards/margins": 0.00034020940074697137,
      "rewards/rejected": 0.0008673094562254846,
      "step": 5
    },
    {
      "epoch": 0.1,
      "grad_norm": 7.177620483727426,
      "learning_rate": 1.5e-06,
      "logits/chosen": -2.7866053581237793,
      "logits/rejected": -2.7954766750335693,
      "logps/chosen": -186.5630340576172,
      "logps/rejected": -192.17149353027344,
      "loss": 0.6878,
      "rewards/accuracies": 0.625,
      "rewards/chosen": 0.038089457899332047,
      "rewards/margins": 0.011779792606830597,
      "rewards/rejected": 0.02630966529250145,
      "step": 10
    },
    {
      "epoch": 0.16,
      "grad_norm": 8.417917120076636,
      "learning_rate": 1.4872298247629265e-06,
      "logits/chosen": -2.824795722961426,
      "logits/rejected": -2.8347983360290527,
      "logps/chosen": -197.54190063476562,
      "logps/rejected": -192.82992553710938,
      "loss": 0.6676,
      "rewards/accuracies": 0.6000000238418579,
      "rewards/chosen": 0.02421323023736477,
      "rewards/margins": 0.06345993280410767,
      "rewards/rejected": -0.03924670070409775,
      "step": 15
    },
    {
      "epoch": 0.21,
      "grad_norm": 9.10298055209343,
      "learning_rate": 1.4493541720532668e-06,
      "logits/chosen": -2.7853965759277344,
      "logits/rejected": -2.7954306602478027,
      "logps/chosen": -172.51913452148438,
      "logps/rejected": -199.7716522216797,
      "loss": 0.6292,
      "rewards/accuracies": 0.699999988079071,
      "rewards/chosen": -0.0377628356218338,
      "rewards/margins": 0.16324351727962494,
      "rewards/rejected": -0.20100633800029755,
      "step": 20
    },
    {
      "epoch": 0.26,
      "grad_norm": 10.051372999934031,
      "learning_rate": 1.3876628517972108e-06,
      "logits/chosen": -2.7528138160705566,
      "logits/rejected": -2.7506392002105713,
      "logps/chosen": -147.0410919189453,
      "logps/rejected": -184.7159423828125,
      "loss": 0.6228,
      "rewards/accuracies": 0.643750011920929,
      "rewards/chosen": -0.15569067001342773,
      "rewards/margins": 0.19696763157844543,
      "rewards/rejected": -0.3526582419872284,
      "step": 25
    },
    {
      "epoch": 0.31,
      "grad_norm": 11.47609664425213,
      "learning_rate": 1.3042566879154944e-06,
      "logits/chosen": -2.77292799949646,
      "logits/rejected": -2.770139217376709,
      "logps/chosen": -189.57742309570312,
      "logps/rejected": -228.9635467529297,
      "loss": 0.6066,
      "rewards/accuracies": 0.7437499761581421,
      "rewards/chosen": -0.3009790778160095,
      "rewards/margins": 0.33582720160484314,
      "rewards/rejected": -0.6368061900138855,
      "step": 30
    },
    {
      "epoch": 0.37,
      "grad_norm": 14.051983189813676,
      "learning_rate": 1.2019759772844424e-06,
      "logits/chosen": -2.8096296787261963,
      "logits/rejected": -2.8017215728759766,
      "logps/chosen": -245.0883026123047,
      "logps/rejected": -281.6932067871094,
      "loss": 0.602,
      "rewards/accuracies": 0.5874999761581421,
      "rewards/chosen": -0.5381699800491333,
      "rewards/margins": 0.27634209394454956,
      "rewards/rejected": -0.8145120739936829,
      "step": 35
    },
    {
      "epoch": 0.42,
      "grad_norm": 16.49665781047947,
      "learning_rate": 1.0843037668324039e-06,
      "logits/chosen": -2.690340042114258,
      "logits/rejected": -2.703282356262207,
      "logps/chosen": -204.72268676757812,
      "logps/rejected": -264.89361572265625,
      "loss": 0.5792,
      "rewards/accuracies": 0.65625,
      "rewards/chosen": -0.7186366319656372,
      "rewards/margins": 0.4523038864135742,
      "rewards/rejected": -1.1709405183792114,
      "step": 40
    },
    {
      "epoch": 0.47,
      "grad_norm": 18.566501994693994,
      "learning_rate": 9.552472425540623e-07,
      "logits/chosen": -2.699636936187744,
      "logits/rejected": -2.7163338661193848,
      "logps/chosen": -291.1601257324219,
      "logps/rejected": -333.49310302734375,
      "loss": 0.6088,
      "rewards/accuracies": 0.6937500238418579,
      "rewards/chosen": -1.0251235961914062,
      "rewards/margins": 0.493073046207428,
      "rewards/rejected": -1.518196702003479,
      "step": 45
    },
    {
      "epoch": 0.52,
      "grad_norm": 20.128869617238305,
      "learning_rate": 8.192012695974766e-07,
      "logits/chosen": -2.689870595932007,
      "logits/rejected": -2.6908297538757324,
      "logps/chosen": -252.3955078125,
      "logps/rejected": -331.952880859375,
      "loss": 0.577,
      "rewards/accuracies": 0.71875,
      "rewards/chosen": -0.780690610408783,
      "rewards/margins": 0.6397809982299805,
      "rewards/rejected": -1.4204715490341187,
      "step": 50
    },
    {
      "epoch": 0.57,
      "grad_norm": 17.347634723380718,
      "learning_rate": 6.807987304025236e-07,
      "logits/chosen": -2.582890748977661,
      "logits/rejected": -2.5895731449127197,
      "logps/chosen": -216.3740234375,
      "logps/rejected": -272.9044494628906,
      "loss": 0.5676,
      "rewards/accuracies": 0.643750011920929,
      "rewards/chosen": -0.6826119422912598,
      "rewards/margins": 0.5186464190483093,
      "rewards/rejected": -1.2012584209442139,
      "step": 55
    },
    {
      "epoch": 0.63,
      "grad_norm": 18.207607616337743,
      "learning_rate": 5.447527574459379e-07,
      "logits/chosen": -2.615662097930908,
      "logits/rejected": -2.612374782562256,
      "logps/chosen": -263.39892578125,
      "logps/rejected": -343.30780029296875,
      "loss": 0.556,
      "rewards/accuracies": 0.706250011920929,
      "rewards/chosen": -0.9144086837768555,
      "rewards/margins": 0.5602952241897583,
      "rewards/rejected": -1.4747039079666138,
      "step": 60
    },
    {
      "epoch": 0.68,
      "grad_norm": 17.920968395292732,
      "learning_rate": 4.1569623316759636e-07,
      "logits/chosen": -2.603685140609741,
      "logits/rejected": -2.6156582832336426,
      "logps/chosen": -278.8509216308594,
      "logps/rejected": -348.4021911621094,
      "loss": 0.5707,
      "rewards/accuracies": 0.7124999761581421,
      "rewards/chosen": -1.1677374839782715,
      "rewards/margins": 0.6181171536445618,
      "rewards/rejected": -1.7858545780181885,
      "step": 65
    },
    {
      "epoch": 0.73,
      "grad_norm": 21.60317231649874,
      "learning_rate": 2.980240227155578e-07,
      "logits/chosen": -2.530977725982666,
      "logits/rejected": -2.538461208343506,
      "logps/chosen": -298.65631103515625,
      "logps/rejected": -345.4963684082031,
      "loss": 0.5603,
      "rewards/accuracies": 0.6499999761581421,
      "rewards/chosen": -1.2151950597763062,
      "rewards/margins": 0.5020937323570251,
      "rewards/rejected": -1.7172887325286865,
      "step": 70
    },
    {
      "epoch": 0.78,
      "grad_norm": 23.519516498457133,
      "learning_rate": 1.9574331208450578e-07,
      "logits/chosen": -2.4647083282470703,
      "logits/rejected": -2.457724094390869,
      "logps/chosen": -297.4154052734375,
      "logps/rejected": -408.7651062011719,
      "loss": 0.5333,
      "rewards/accuracies": 0.71875,
      "rewards/chosen": -1.2555583715438843,
      "rewards/margins": 0.9959794282913208,
      "rewards/rejected": -2.251537799835205,
      "step": 75
    },
    {
      "epoch": 0.84,
      "grad_norm": 24.189874157051893,
      "learning_rate": 1.1233714820278951e-07,
      "logits/chosen": -2.415569305419922,
      "logits/rejected": -2.409623384475708,
      "logps/chosen": -286.6866455078125,
      "logps/rejected": -374.3118591308594,
      "loss": 0.5319,
      "rewards/accuracies": 0.737500011920929,
      "rewards/chosen": -1.1869741678237915,
      "rewards/margins": 0.8028388023376465,
      "rewards/rejected": -1.9898128509521484,
      "step": 80
    },
    {
      "epoch": 0.89,
      "grad_norm": 27.1421302316283,
      "learning_rate": 5.064582794673323e-08,
      "logits/chosen": -2.463069438934326,
      "logits/rejected": -2.4618234634399414,
      "logps/chosen": -291.75006103515625,
      "logps/rejected": -380.35186767578125,
      "loss": 0.5565,
      "rewards/accuracies": 0.6937500238418579,
      "rewards/chosen": -1.315664529800415,
      "rewards/margins": 0.7988798022270203,
      "rewards/rejected": -2.11454439163208,
      "step": 85
    },
    {
      "epoch": 0.94,
      "grad_norm": 24.52916506686646,
      "learning_rate": 1.2770175237073662e-08,
      "logits/chosen": -2.391339063644409,
      "logits/rejected": -2.3986663818359375,
      "logps/chosen": -315.0043640136719,
      "logps/rejected": -409.3547668457031,
      "loss": 0.552,
      "rewards/accuracies": 0.699999988079071,
      "rewards/chosen": -1.4606488943099976,
      "rewards/margins": 0.8770757913589478,
      "rewards/rejected": -2.337724447250366,
      "step": 90
    },
    {
      "epoch": 0.99,
      "grad_norm": 31.400140337371543,
      "learning_rate": 0.0,
      "logits/chosen": -2.4539380073547363,
      "logits/rejected": -2.4629311561584473,
      "logps/chosen": -328.31024169921875,
      "logps/rejected": -411.2120056152344,
      "loss": 0.5588,
      "rewards/accuracies": 0.6937500238418579,
      "rewards/chosen": -1.4270151853561401,
      "rewards/margins": 0.8573243021965027,
      "rewards/rejected": -2.284339666366577,
      "step": 95
    },
    {
      "epoch": 0.99,
      "step": 95,
      "total_flos": 0.0,
      "train_loss": 0.5926835913407175,
      "train_runtime": 2430.0992,
      "train_samples_per_second": 5.031,
      "train_steps_per_second": 0.039
    }
  ],
  "logging_steps": 5,
  "max_steps": 95,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 1,
  "save_steps": 500,
  "total_flos": 0.0,
  "train_batch_size": 8,
  "trial_name": null,
  "trial_params": null
}