{
    "epoch": 0.9997382884061764,
    "eval_logits/chosen": -2.736534833908081,
    "eval_logits/rejected": -2.716714382171631,
    "eval_logps/chosen": -271.84613037109375,
    "eval_logps/rejected": -276.6851501464844,
    "eval_loss": 0.5081929564476013,
    "eval_rewards/accuracies": 0.7222222089767456,
    "eval_rewards/chosen": 0.002502653282135725,
    "eval_rewards/margins": 0.9071723222732544,
    "eval_rewards/rejected": -0.9046697020530701,
    "eval_runtime": 352.642,
    "eval_samples": 2000,
    "eval_samples_per_second": 5.671,
    "eval_steps_per_second": 0.179
}