{ "method": "dpo", "loss_type": "dpo", "seed": 42, "beta": 0.05, "backbone": "Qwen/Qwen3-8B", "status": "completed", "returncode": 0, "completed_at": "2026-07-08T13:03:41Z", "dataset": "/home/sjkim/mnpo_shared/revision_qwen3_8b/full_iter1/precomputed/avg_oracle", "output_dir": "/ext_hdd/sjkim/mnpo/revision_qwen3_8b/full_iter1/train/dpo_avg_beta0p05_s42_odin2", "config": "/ext_hdd/sjkim/mnpo/revision_qwen3_8b/full_iter1/train/dpo_avg_beta0p05_s42_odin2/config.yaml", "train_metrics_exists": true, "trainer_state_exists": true }