eval_results.json

{
    "epoch": 0.9994767137624281,
    "eval_logits/chosen": 0.45220446586608887,
    "eval_logits/rejected": 0.9319149851799011,
    "eval_logps/chosen": -137.00851440429688,
    "eval_logps/rejected": -190.1944580078125,
    "eval_loss": 0.5959618091583252,
    "eval_rewards/accuracies": 0.369047611951828,
    "eval_rewards/chosen": -0.6251727342605591,
    "eval_rewards/margins": 0.4654882848262787,
    "eval_rewards/rejected": -1.0906610488891602,
    "eval_runtime": 114.506,
    "eval_samples": 2000,
    "eval_samples_per_second": 17.466,
    "eval_steps_per_second": 0.55
}