{ "repo_name": "ReasoningRelease-AuditRepo", "policy_name": "balanced_release_v1", "selected_checkpoint": "step_700", "eligible_checkpoints": [ "step_600", "step_700" ], "eligible_rankings": [ { "rank": 1, "checkpoint": "step_700", "weighted_overall": "0.739", "min_core_reasoning": "0.592", "release_score": "0.723", "release_score_margin_vs_selected": "+0.000", "tie_breaker_weighted_overall": "0.739", "tie_breaker_step": 700 }, { "rank": 2, "checkpoint": "step_600", "weighted_overall": "0.731", "min_core_reasoning": "0.583", "release_score": "0.714", "release_score_margin_vs_selected": "-0.009", "tie_breaker_weighted_overall": "0.731", "tie_breaker_step": 600 } ], "minimum_margins": [ { "checkpoint": "step_100", "safety_evaluation": "-0.140", "instruction_following": "-0.139", "question_answering": "-0.148", "limiting_minimum": "question_answering", "limiting_margin": "-0.148" }, { "checkpoint": "step_200", "safety_evaluation": "-0.109", "instruction_following": "-0.105", "question_answering": "-0.113", "limiting_minimum": "question_answering", "limiting_margin": "-0.113" }, { "checkpoint": "step_300", "safety_evaluation": "-0.077", "instruction_following": "-0.072", "question_answering": "-0.075", "limiting_minimum": "safety_evaluation", "limiting_margin": "-0.077" }, { "checkpoint": "step_400", "safety_evaluation": "-0.044", "instruction_following": "-0.041", "question_answering": "-0.038", "limiting_minimum": "safety_evaluation", "limiting_margin": "-0.044" }, { "checkpoint": "step_500", "safety_evaluation": "-0.012", "instruction_following": "-0.012", "question_answering": "-0.008", "limiting_minimum": "safety_evaluation", "limiting_margin": "-0.012" }, { "checkpoint": "step_600", "safety_evaluation": "+0.014", "instruction_following": "+0.005", "question_answering": "+0.011", "limiting_minimum": "instruction_following", "limiting_margin": "+0.005" }, { "checkpoint": "step_700", "safety_evaluation": "+0.026", "instruction_following": "+0.011", "question_answering": "+0.021", "limiting_minimum": "instruction_following", "limiting_margin": "+0.011" }, { "checkpoint": "step_800", "safety_evaluation": "-0.056", "instruction_following": "+0.015", "question_answering": "+0.028", "limiting_minimum": "safety_evaluation", "limiting_margin": "-0.056" }, { "checkpoint": "step_900", "safety_evaluation": "-0.038", "instruction_following": "-0.022", "question_answering": "+0.034", "limiting_minimum": "safety_evaluation", "limiting_margin": "-0.038" }, { "checkpoint": "step_1000", "safety_evaluation": "-0.039", "instruction_following": "-0.017", "question_answering": "+0.039", "limiting_minimum": "safety_evaluation", "limiting_margin": "-0.039" } ], "audit_file_hashes": [ { "path": "audit/benchmark_trace.jsonl", "sha256": "42a17098a02213b55f3f49d31b3e8d12c6d45b8ec3ef895fb6d1f16378efd9fb" }, { "path": "audit/checkpoint_scorecard.csv", "sha256": "922b77cb8ecdd3d740c8086b062d3406a42f3385f53c41ea5121fb6cd3ab8c88" } ] }