{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 2.957345971563981,
  "eval_steps": 100,
  "global_step": 156,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.02,
      "grad_norm": 132.97353908293687,
      "learning_rate": 3.125e-08,
      "logits/chosen": 123.11854553222656,
      "logits/rejected": 97.00198364257812,
      "logps/chosen": -425.18585205078125,
      "logps/rejected": -424.1869201660156,
      "loss": 0.6931,
      "rewards/accuracies": 0.0,
      "rewards/chosen": 0.0,
      "rewards/margins": 0.0,
      "rewards/rejected": 0.0,
      "step": 1
    },
    {
      "epoch": 0.19,
      "grad_norm": 206.0883100010928,
      "learning_rate": 3.1249999999999997e-07,
      "logits/chosen": 117.39097595214844,
      "logits/rejected": 136.3163299560547,
      "logps/chosen": -442.6399230957031,
      "logps/rejected": -524.91015625,
      "loss": 0.7186,
      "rewards/accuracies": 0.4930555522441864,
      "rewards/chosen": 0.037425246089696884,
      "rewards/margins": 0.07718456536531448,
      "rewards/rejected": -0.0397593155503273,
      "step": 10
    },
    {
      "epoch": 0.38,
      "grad_norm": 114.8435303205146,
      "learning_rate": 4.989935734988097e-07,
      "logits/chosen": 125.3319091796875,
      "logits/rejected": 132.9754638671875,
      "logps/chosen": -422.8042907714844,
      "logps/rejected": -491.63226318359375,
      "loss": 0.6164,
      "rewards/accuracies": 0.637499988079071,
      "rewards/chosen": 0.17301546037197113,
      "rewards/margins": 0.379099041223526,
      "rewards/rejected": -0.20608356595039368,
      "step": 20
    },
    {
      "epoch": 0.57,
      "grad_norm": 99.27143207986335,
      "learning_rate": 4.877641290737883e-07,
      "logits/chosen": 122.47686767578125,
      "logits/rejected": 125.91865539550781,
      "logps/chosen": -466.9618225097656,
      "logps/rejected": -540.3817138671875,
      "loss": 0.5813,
      "rewards/accuracies": 0.731249988079071,
      "rewards/chosen": -0.5920838117599487,
      "rewards/margins": 1.0662639141082764,
      "rewards/rejected": -1.658347725868225,
      "step": 30
    },
    {
      "epoch": 0.76,
      "grad_norm": 109.14521515462766,
      "learning_rate": 4.646121984004665e-07,
      "logits/chosen": 124.97059631347656,
      "logits/rejected": 119.9173583984375,
      "logps/chosen": -497.7147521972656,
      "logps/rejected": -527.3887939453125,
      "loss": 0.5426,
      "rewards/accuracies": 0.7562500238418579,
      "rewards/chosen": -1.2266700267791748,
      "rewards/margins": 0.9045358896255493,
      "rewards/rejected": -2.1312055587768555,
      "step": 40
    },
    {
      "epoch": 0.95,
      "grad_norm": 115.6113002085735,
      "learning_rate": 4.3069871595684787e-07,
      "logits/chosen": 132.8910369873047,
      "logits/rejected": 133.22190856933594,
      "logps/chosen": -520.63037109375,
      "logps/rejected": -549.1149291992188,
      "loss": 0.5202,
      "rewards/accuracies": 0.6812499761581421,
      "rewards/chosen": -1.8903814554214478,
      "rewards/margins": 1.0029468536376953,
      "rewards/rejected": -2.8933284282684326,
      "step": 50
    },
    {
      "epoch": 1.14,
      "grad_norm": 46.9650475313439,
      "learning_rate": 3.877242453630256e-07,
      "logits/chosen": 131.47854614257812,
      "logits/rejected": 134.71681213378906,
      "logps/chosen": -481.8072814941406,
      "logps/rejected": -534.0516357421875,
      "loss": 0.2837,
      "rewards/accuracies": 0.893750011920929,
      "rewards/chosen": -1.2520891427993774,
      "rewards/margins": 2.2355263233184814,
      "rewards/rejected": -3.4876155853271484,
      "step": 60
    },
    {
      "epoch": 1.33,
      "grad_norm": 47.26485069523079,
      "learning_rate": 3.378437060203357e-07,
      "logits/chosen": 126.1490707397461,
      "logits/rejected": 126.75111389160156,
      "logps/chosen": -452.6795349121094,
      "logps/rejected": -579.5133056640625,
      "loss": 0.1756,
      "rewards/accuracies": 0.9437500238418579,
      "rewards/chosen": -1.188291072845459,
      "rewards/margins": 2.6805100440979004,
      "rewards/rejected": -3.868800640106201,
      "step": 70
    },
    {
      "epoch": 1.52,
      "grad_norm": 46.43874254029814,
      "learning_rate": 2.8355831645441387e-07,
      "logits/chosen": 127.46858978271484,
      "logits/rejected": 128.4056396484375,
      "logps/chosen": -514.4637451171875,
      "logps/rejected": -621.2301635742188,
      "loss": 0.1711,
      "rewards/accuracies": 0.9624999761581421,
      "rewards/chosen": -1.268179178237915,
      "rewards/margins": 3.4565296173095703,
      "rewards/rejected": -4.724708557128906,
      "step": 80
    },
    {
      "epoch": 1.71,
      "grad_norm": 49.11808633093636,
      "learning_rate": 2.2759017277414164e-07,
      "logits/chosen": 112.5447998046875,
      "logits/rejected": 114.98893737792969,
      "logps/chosen": -497.70001220703125,
      "logps/rejected": -589.730224609375,
      "loss": 0.1524,
      "rewards/accuracies": 0.987500011920929,
      "rewards/chosen": -1.7766218185424805,
      "rewards/margins": 3.491291046142578,
      "rewards/rejected": -5.2679123878479,
      "step": 90
    },
    {
      "epoch": 1.9,
      "grad_norm": 47.47224806448749,
      "learning_rate": 1.7274575140626315e-07,
      "logits/chosen": 124.581787109375,
      "logits/rejected": 115.68563079833984,
      "logps/chosen": -516.1900634765625,
      "logps/rejected": -632.6817626953125,
      "loss": 0.1623,
      "rewards/accuracies": 0.987500011920929,
      "rewards/chosen": -1.8899990320205688,
      "rewards/margins": 3.6050896644592285,
      "rewards/rejected": -5.495089054107666,
      "step": 100
    },
    {
      "epoch": 1.9,
      "eval_logits/chosen": 93.8502197265625,
      "eval_logits/rejected": 87.7247543334961,
      "eval_logps/chosen": -512.8825073242188,
      "eval_logps/rejected": -541.5043334960938,
      "eval_loss": 0.48611319065093994,
      "eval_rewards/accuracies": 0.6770833134651184,
      "eval_rewards/chosen": -2.9739017486572266,
      "eval_rewards/margins": 1.5238369703292847,
      "eval_rewards/rejected": -4.497739315032959,
      "eval_runtime": 53.4905,
      "eval_samples_per_second": 14.021,
      "eval_steps_per_second": 0.449,
      "step": 100
    },
    {
      "epoch": 2.09,
      "grad_norm": 25.494387206609645,
      "learning_rate": 1.2177518064852348e-07,
      "logits/chosen": 102.986083984375,
      "logits/rejected": 116.60546875,
      "logps/chosen": -538.074951171875,
      "logps/rejected": -667.3218383789062,
      "loss": 0.1318,
      "rewards/accuracies": 0.96875,
      "rewards/chosen": -2.1493353843688965,
      "rewards/margins": 3.4991326332092285,
      "rewards/rejected": -5.648468017578125,
      "step": 110
    },
    {
      "epoch": 2.27,
      "grad_norm": 24.60483354043265,
      "learning_rate": 7.723433775328384e-08,
      "logits/chosen": 113.220703125,
      "logits/rejected": 114.29705810546875,
      "logps/chosen": -522.1823120117188,
      "logps/rejected": -628.0721435546875,
      "loss": 0.0837,
      "rewards/accuracies": 1.0,
      "rewards/chosen": -2.2452592849731445,
      "rewards/margins": 3.996518611907959,
      "rewards/rejected": -6.2417778968811035,
      "step": 120
    },
    {
      "epoch": 2.46,
      "grad_norm": 32.75955455536007,
      "learning_rate": 4.1356686569674335e-08,
      "logits/chosen": 115.95035552978516,
      "logits/rejected": 120.65645599365234,
      "logps/chosen": -537.8087158203125,
      "logps/rejected": -653.7862548828125,
      "loss": 0.0781,
      "rewards/accuracies": 0.987500011920929,
      "rewards/chosen": -2.2025058269500732,
      "rewards/margins": 4.200939178466797,
      "rewards/rejected": -6.403443813323975,
      "step": 130
    },
    {
      "epoch": 2.65,
      "grad_norm": 23.375967561613557,
      "learning_rate": 1.5941282340065697e-08,
      "logits/chosen": 101.51383972167969,
      "logits/rejected": 102.2659683227539,
      "logps/chosen": -499.16229248046875,
      "logps/rejected": -645.9388427734375,
      "loss": 0.0791,
      "rewards/accuracies": 0.981249988079071,
      "rewards/chosen": -2.6822938919067383,
      "rewards/margins": 4.259942054748535,
      "rewards/rejected": -6.942234992980957,
      "step": 140
    },
    {
      "epoch": 2.84,
      "grad_norm": 27.725216044545164,
      "learning_rate": 2.2625595580163247e-09,
      "logits/chosen": 108.08512878417969,
      "logits/rejected": 121.6434097290039,
      "logps/chosen": -524.5687866210938,
      "logps/rejected": -647.0615844726562,
      "loss": 0.079,
      "rewards/accuracies": 0.9750000238418579,
      "rewards/chosen": -2.486959457397461,
      "rewards/margins": 4.458805084228516,
      "rewards/rejected": -6.945765018463135,
      "step": 150
    },
    {
      "epoch": 2.96,
      "step": 156,
      "total_flos": 0.0,
      "train_loss": 0.28389122929328525,
      "train_runtime": 1811.0132,
      "train_samples_per_second": 11.182,
      "train_steps_per_second": 0.086
    }
  ],
  "logging_steps": 10,
  "max_steps": 156,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 3,
  "save_steps": 100,
  "total_flos": 0.0,
  "train_batch_size": 2,
  "trial_name": null,
  "trial_params": null
}