{
    "epoch": 3.0,
    "eval_logits/chosen": -2.6887364387512207,
    "eval_logits/rejected": -2.6727285385131836,
    "eval_logps/chosen": -75.25064849853516,
    "eval_logps/rejected": -85.66250610351562,
    "eval_loss": 0.9378770589828491,
    "eval_rewards/accuracies": 0.3472222089767456,
    "eval_rewards/chosen": -0.34173667430877686,
    "eval_rewards/margins": 1.698632836341858,
    "eval_rewards/rejected": -2.0403695106506348,
    "eval_runtime": 119.487,
    "eval_samples": 2000,
    "eval_samples_per_second": 16.738,
    "eval_steps_per_second": 0.527,
    "total_flos": 0.0,
    "train_loss": 0.49774221859604084,
    "train_runtime": 9884.8529,
    "train_samples": 18340,
    "train_samples_per_second": 5.566,
    "train_steps_per_second": 0.087
}