eval_results.json

{
    "epoch": 1.0,
    "eval_logits/chosen": -1.9401931762695312,
    "eval_logits/rejected": -1.9123154878616333,
    "eval_logps/chosen": -77.5232162475586,
    "eval_logps/rejected": -95.19373321533203,
    "eval_loss": 0.6917868852615356,
    "eval_pred_label": 4738.58740234375,
    "eval_rewards/accuracies": 0.3591269850730896,
    "eval_rewards/chosen": -0.0862266793847084,
    "eval_rewards/margins": 0.11172995716333389,
    "eval_rewards/rejected": -0.19795666635036469,
    "eval_runtime": 247.3331,
    "eval_samples": 2000,
    "eval_samples_per_second": 8.086,
    "eval_steps_per_second": 0.255,
    "eval_use_label": 15333.4130859375
}