eval_results.json

{
    "epoch": 1.0,
    "eval_logits/chosen": -2.4939169883728027,
    "eval_logits/rejected": -2.495774507522583,
    "eval_logps/chosen": -269.28546142578125,
    "eval_logps/rejected": -253.23594665527344,
    "eval_loss": 0.6354129910469055,
    "eval_pred_label": 13796.32421875,
    "eval_rewards/accuracies": 0.6259999871253967,
    "eval_rewards/chosen": 0.027118388563394547,
    "eval_rewards/margins": 0.056793875992298126,
    "eval_rewards/rejected": -0.029675481840968132,
    "eval_runtime": 1013.6049,
    "eval_samples": 2000,
    "eval_samples_per_second": 1.973,
    "eval_steps_per_second": 0.247,
    "eval_use_label": 18265.67578125
}