all_results.json

{
    "epoch": 1.0,
    "eval_logits/chosen": 0.9053679704666138,
    "eval_logits/rejected": 1.7481720447540283,
    "eval_logps/chosen": -228.00465393066406,
    "eval_logps/rejected": -198.00367736816406,
    "eval_loss": 4449.90234375,
    "eval_rewards/accuracies": 0.616104006767273,
    "eval_rewards/chosen": -0.8765509724617004,
    "eval_rewards/margins": 0.08215557038784027,
    "eval_rewards/rejected": -0.9587064981460571,
    "eval_rewards/safe_rewards": -0.865267813205719,
    "eval_rewards/unsafe_rewards": -0.860846996307373,
    "eval_runtime": 2354.4025,
    "eval_samples": 35044,
    "eval_samples_per_second": 14.884,
    "eval_steps_per_second": 0.466,
    "train_loss": 5859.617769083399,
    "train_runtime": 32772.3871,
    "train_samples": 120613,
    "train_samples_per_second": 3.68,
    "train_steps_per_second": 0.057
}