{
    "epoch": 1.0,
    "eval_logits/chosen": 98.66954803466797,
    "eval_logits/rejected": 98.6399917602539,
    "eval_logps/chosen": -32.599971771240234,
    "eval_logps/rejected": -36.293174743652344,
    "eval_loss": 25.394121170043945,
    "eval_rewards/accuracies": 0.550664484500885,
    "eval_rewards/chosen": -0.015679799020290375,
    "eval_rewards/margins": 0.016984686255455017,
    "eval_rewards/rejected": -0.03266448527574539,
    "eval_runtime": 103.7202,
    "eval_samples": 343,
    "eval_samples_per_second": 3.307,
    "eval_steps_per_second": 0.415
}