all_results.json

{
    "epoch": 0.00012874983906270118,
    "eval_logits/chosen": -0.00010515563189983368,
    "eval_logits/rejected": -8.74076213221997e-05,
    "eval_logps/chosen": -2241.19482421875,
    "eval_logps/rejected": -2296.974853515625,
    "eval_loss": 0.6931134462356567,
    "eval_rewards/accuracies": 0.625,
    "eval_rewards/chosen": -7.68661848269403e-05,
    "eval_rewards/margins": 7.162100519053638e-05,
    "eval_rewards/rejected": -0.00014848708815407008,
    "eval_runtime": 3.5865,
    "eval_samples_per_second": 2.231,
    "eval_steps_per_second": 0.279
}