{
    "epoch": 0.9989071038251366,
    "eval_logits/chosen": -1.2960765361785889,
    "eval_logits/rejected": -1.2513173818588257,
    "eval_logps/chosen": -2.094325304031372,
    "eval_logps/rejected": -2.8013558387756348,
    "eval_loss": 1.674479603767395,
    "eval_rewards/accuracies": 0.8433734774589539,
    "eval_rewards/chosen": -20.943252563476562,
    "eval_rewards/margins": 7.0703020095825195,
    "eval_rewards/rejected": -28.01355743408203,
    "eval_runtime": 33.5359,
    "eval_samples": 1318,
    "eval_samples_per_second": 39.301,
    "eval_steps_per_second": 2.475,
    "total_flos": 0.0,
    "train_loss": 2.442784007842483,
    "train_runtime": 5988.2154,
    "train_samples": 58558,
    "train_samples_per_second": 9.779,
    "train_steps_per_second": 0.076
}