{ "epoch": 2.999098751126561, "eval_kl/chosen": 0.0069907549768686295, "eval_kl/rejected": 0.005060141906142235, "eval_logits/chosen": -1.9296875, "eval_logits/rejected": -1.984375, "eval_logps/chosen": -330.0, "eval_logps/rejected": -316.0, "eval_loss": 1.5701874494552612, "eval_nll_loss": 1.0859375, "eval_ref_probs/chosen": 0.0038271306548267603, "eval_ref_probs/rejected": 0.0034569373819977045, "eval_rejected_term/max": 0.005060141906142235, "eval_rejected_term/min": 0.005060141906142235, "eval_rewards/accuracies": 0.7639999985694885, "eval_rewards/chosen": -0.83984375, "eval_rewards/margins": 1.03125, "eval_rewards/rejected": -1.8671875, "eval_runtime": 25.2389, "eval_samples_per_second": 39.621, "eval_steps_per_second": 9.905 }