{
    "epoch": 1.0,
    "eval_logits/chosen": 1.3032478094100952,
    "eval_logits/rejected": 2.0913753509521484,
    "eval_logps/chosen": -363.4786376953125,
    "eval_logps/rejected": -457.9681396484375,
    "eval_loss": 0.5057953596115112,
    "eval_rewards/accuracies": 0.76171875,
    "eval_rewards/chosen": -1.0084874629974365,
    "eval_rewards/margins": 0.9445733428001404,
    "eval_rewards/rejected": -1.9530608654022217,
    "eval_runtime": 101.8853,
    "eval_samples": 2000,
    "eval_samples_per_second": 19.63,
    "eval_steps_per_second": 0.314,
    "total_flos": 0.0,
    "train_loss": 0.5417051395113,
    "train_runtime": 12463.9447,
    "train_samples": 61134,
    "train_samples_per_second": 4.905,
    "train_steps_per_second": 0.038
}