eval_results.json

{
    "epoch": 0.995910949568378,
    "eval_logits/chosen": -15.370370864868164,
    "eval_logits/rejected": -15.3493070602417,
    "eval_logps/chosen": -1.7481694221496582,
    "eval_logps/rejected": -1.9914860725402832,
    "eval_loss": 4.618270397186279,
    "eval_rewards/accuracies": 0.6290983557701111,
    "eval_rewards/chosen": -17.481693267822266,
    "eval_rewards/margins": 2.4331674575805664,
    "eval_rewards/rejected": -19.91486167907715,
    "eval_runtime": 86.386,
    "eval_samples": 1941,
    "eval_samples_per_second": 22.469,
    "eval_steps_per_second": 1.412
}