{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 1.971563981042654,
  "eval_steps": 128,
  "global_step": 104,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.018957345971563982,
      "grad_norm": 176.7677007020494,
      "learning_rate": 4.545454545454545e-08,
      "logits/chosen": 117.53560638427734,
      "logits/rejected": 126.8960952758789,
      "logps/chosen": -335.40118408203125,
      "logps/rejected": -439.16552734375,
      "loss": 0.8447,
      "rewards/accuracies": 0.0,
      "rewards/chosen": 0.0,
      "rewards/margins": 0.0,
      "rewards/rejected": 0.0,
      "step": 1
    },
    {
      "epoch": 0.1895734597156398,
      "grad_norm": 193.24952000915732,
      "learning_rate": 4.545454545454545e-07,
      "logits/chosen": 135.0142822265625,
      "logits/rejected": 138.36141967773438,
      "logps/chosen": -395.7818603515625,
      "logps/rejected": -438.9291076660156,
      "loss": 0.8213,
      "rewards/accuracies": 0.4513888955116272,
      "rewards/chosen": 0.010734621435403824,
      "rewards/margins": -0.009185939095914364,
      "rewards/rejected": 0.019920555874705315,
      "step": 10
    },
    {
      "epoch": 0.3791469194312796,
      "grad_norm": 150.94677892744545,
      "learning_rate": 4.885348141000122e-07,
      "logits/chosen": 121.2293930053711,
      "logits/rejected": 124.93943786621094,
      "logps/chosen": -367.0641784667969,
      "logps/rejected": -420.8922424316406,
      "loss": 0.6725,
      "rewards/accuracies": 0.625,
      "rewards/chosen": 0.2673913538455963,
      "rewards/margins": 0.3124334216117859,
      "rewards/rejected": -0.04504210874438286,
      "step": 20
    },
    {
      "epoch": 0.5687203791469194,
      "grad_norm": 134.59541393391044,
      "learning_rate": 4.5025027361734613e-07,
      "logits/chosen": 142.46116638183594,
      "logits/rejected": 136.04672241210938,
      "logps/chosen": -417.72314453125,
      "logps/rejected": -463.00872802734375,
      "loss": 0.608,
      "rewards/accuracies": 0.699999988079071,
      "rewards/chosen": -1.2628967761993408,
      "rewards/margins": 0.8874266743659973,
      "rewards/rejected": -2.1503233909606934,
      "step": 30
    },
    {
      "epoch": 0.7582938388625592,
      "grad_norm": 133.59435318340778,
      "learning_rate": 3.893311157806091e-07,
      "logits/chosen": 126.9672622680664,
      "logits/rejected": 115.38719177246094,
      "logps/chosen": -408.3771057128906,
      "logps/rejected": -433.11651611328125,
      "loss": 0.5729,
      "rewards/accuracies": 0.6937500238418579,
      "rewards/chosen": -2.709144353866577,
      "rewards/margins": 1.0529193878173828,
      "rewards/rejected": -3.762063980102539,
      "step": 40
    },
    {
      "epoch": 0.9478672985781991,
      "grad_norm": 138.18876497797683,
      "learning_rate": 3.126631330646801e-07,
      "logits/chosen": 141.5520477294922,
      "logits/rejected": 145.66412353515625,
      "logps/chosen": -458.2511291503906,
      "logps/rejected": -540.9998779296875,
      "loss": 0.5054,
      "rewards/accuracies": 0.737500011920929,
      "rewards/chosen": -1.955364465713501,
      "rewards/margins": 1.278308629989624,
      "rewards/rejected": -3.233673572540283,
      "step": 50
    },
    {
      "epoch": 1.1374407582938388,
      "grad_norm": 76.7561439392463,
      "learning_rate": 2.2891223348923882e-07,
      "logits/chosen": 133.77720642089844,
      "logits/rejected": 137.45901489257812,
      "logps/chosen": -436.24725341796875,
      "logps/rejected": -518.0458374023438,
      "loss": 0.3092,
      "rewards/accuracies": 0.875,
      "rewards/chosen": -1.91208016872406,
      "rewards/margins": 2.171934127807617,
      "rewards/rejected": -4.084014415740967,
      "step": 60
    },
    {
      "epoch": 1.3270142180094786,
      "grad_norm": 87.35928232615755,
      "learning_rate": 1.4754491880085317e-07,
      "logits/chosen": 127.97108459472656,
      "logits/rejected": 129.74716186523438,
      "logps/chosen": -410.78009033203125,
      "logps/rejected": -508.24462890625,
      "loss": 0.1927,
      "rewards/accuracies": 0.893750011920929,
      "rewards/chosen": -2.181811809539795,
      "rewards/margins": 2.5547263622283936,
      "rewards/rejected": -4.736538410186768,
      "step": 70
    },
    {
      "epoch": 1.5165876777251186,
      "grad_norm": 84.8043898828848,
      "learning_rate": 7.775827023107834e-08,
      "logits/chosen": 115.09355163574219,
      "logits/rejected": 131.64649963378906,
      "logps/chosen": -409.45123291015625,
      "logps/rejected": -520.7732543945312,
      "loss": 0.1818,
      "rewards/accuracies": 0.9125000238418579,
      "rewards/chosen": -2.7900097370147705,
      "rewards/margins": 2.5867667198181152,
      "rewards/rejected": -5.376776218414307,
      "step": 80
    },
    {
      "epoch": 1.7061611374407581,
      "grad_norm": 73.29783309901748,
      "learning_rate": 2.7440387297912122e-08,
      "logits/chosen": 115.2972412109375,
      "logits/rejected": 128.11497497558594,
      "logps/chosen": -435.94677734375,
      "logps/rejected": -542.7080688476562,
      "loss": 0.162,
      "rewards/accuracies": 0.9312499761581421,
      "rewards/chosen": -2.475785493850708,
      "rewards/margins": 2.7755281925201416,
      "rewards/rejected": -5.251313209533691,
      "step": 90
    },
    {
      "epoch": 1.8957345971563981,
      "grad_norm": 72.22436049386164,
      "learning_rate": 2.27878296044029e-09,
      "logits/chosen": 121.9096908569336,
      "logits/rejected": 121.7502212524414,
      "logps/chosen": -424.9623107910156,
      "logps/rejected": -510.92376708984375,
      "loss": 0.1531,
      "rewards/accuracies": 0.949999988079071,
      "rewards/chosen": -2.149359703063965,
      "rewards/margins": 2.558631420135498,
      "rewards/rejected": -4.707991600036621,
      "step": 100
    },
    {
      "epoch": 1.971563981042654,
      "step": 104,
      "total_flos": 0.0,
      "train_loss": 0.41261253735193837,
      "train_runtime": 2174.4689,
      "train_samples_per_second": 6.208,
      "train_steps_per_second": 0.048
    }
  ],
  "logging_steps": 10,
  "max_steps": 104,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 2,
  "save_steps": 500,
  "total_flos": 0.0,
  "train_batch_size": 2,
  "trial_name": null,
  "trial_params": null
}