{
    "epoch": 0.9959925193694897,
    "eval_logits/chosen": -2.370966672897339,
    "eval_logits/rejected": -2.2963974475860596,
    "eval_logps/chosen": -223.0048065185547,
    "eval_logps/ref_chosen": -237.70138549804688,
    "eval_logps/ref_rejected": -244.81951904296875,
    "eval_logps/rejected": -232.80235290527344,
    "eval_loss": 0.4936164915561676,
    "eval_rewards/accuracies": 0.6572580933570862,
    "eval_rewards/chosen": 0.14696598052978516,
    "eval_rewards/margins": 0.026794558390975,
    "eval_rewards/rejected": 0.12017140537500381,
    "eval_runtime": 141.5135,
    "eval_samples": 1961,
    "eval_samples_per_second": 13.857,
    "eval_steps_per_second": 0.438
}