{
  "schema": "completed_extension_round2_public_summary_v1",
  "observed_at_beijing": "2026-11-02 12:20:03",
  "rounds": [
    {
      "arm": "WDT_OPT",
      "round": 2,
      "status": "COMPLETE",
      "questions": 40,
      "answers": 320,
      "seed_questions": 20,
      "seed_answers": 160,
      "seed_successes": 125,
      "review_questions": 10,
      "new_questions": 10,
      "effective_questions": 17,
      "final_policy_id": "69cec642ec57251107eebc71b4920feb9db9e5c0e8a660fe9c30f070019e92e9",
      "completed_at_utc": "2026-11-02T00:59:16.661732+00:00",
      "completed_at_beijing": "2026-11-02 08:59:16",
      "started_at_utc": "2026-11-02T00:05:42.761413+00:00",
      "complete_round_seconds": 3213.900319,
      "stage_seconds": {
        "plan-round": 2116.67438,
        "train-round": 1097.208328
      },
      "pure_parameter_training_seconds": null,
      "all_correct_questions": 16,
      "all_wrong_questions": 7,
      "effective_new_questions": 4,
      "new_difficulty_by_kind": {
        "deep_down": {
          "all_failed": 0,
          "all_succeeded": 0,
          "effective_questions": 1,
          "minimum_actions_delta_range": [
            -17,
            -17
          ],
          "questions": 1,
          "required_chain_depth_delta_range": [
            -1,
            -1
          ]
        },
        "deep_up": {
          "all_failed": 3,
          "all_succeeded": 0,
          "effective_questions": 2,
          "minimum_actions_delta_range": [
            1,
            8
          ],
          "questions": 5,
          "required_chain_depth_delta_range": [
            1,
            2
          ]
        },
        "wide": {
          "all_failed": 1,
          "all_succeeded": 2,
          "effective_questions": 1,
          "minimum_actions_delta_range": [
            0,
            1
          ],
          "questions": 4,
          "required_chain_depth_delta_range": [
            0,
            1
          ]
        }
      },
      "signal_coverage": {
        "failed_questions": 16,
        "first_answer_failed_questions": 10,
        "independent_student_answers": 0,
        "later_failures_considered": 6,
        "mixed_source_questions": 13,
        "regressed_question_ids": [
          "generated:local_textcraft_357",
          "generated:local_textcraft_190",
          "generated:cluster_wdt_r1_s4",
          "generated:cluster_wdt_r1_s18"
        ],
        "source_questions": 30,
        "upward_sources_skipped_without_teacher_proposals": 1
      },
      "reserved_new_question_positions": 10,
      "main_learning_updates": 1,
      "main_learning_complete": true,
      "model_saved": true,
      "repair": null,
      "answers_collected_before_second_learning": true,
      "completed_main_learning_rounds_before_collection": 1,
      "answer_meaning": "本轮答卷在第二次学习之前收集，反映第一轮学习后的表现；第二轮模型已经保存。",
      "timing_scope": "完整一轮，包含出题、作答、学习、纠错（如有）、保存和阶段切换。",
      "source_records": {
        "CAPTURE.json": {
          "bytes": 41850,
          "sha256": "a26b94fde3da906b1ea34543dda655657c43d2664c1479394c0f334e692336f5"
        },
        "complete.json": {
          "bytes": 8406,
          "sha256": "6890c386929017b6f72448edd03bd29afe02aa2f451945d58450a79aaa82dc0e"
        },
        "collection.json": {
          "bytes": 11663,
          "sha256": "21b5604ffc73ff9cb319affda280e0baf6ef9b65710b7f98abe4db37f17671a2"
        },
        "plan.json": {
          "bytes": 86396,
          "sha256": "e34be224d2fad9f691ce6ae821d1e5952031e7072705e45620d2f7fd2fb20002"
        },
        "main.json": {
          "bytes": 7702,
          "sha256": "3f6a9c3729731485b372e375f74a7d13464366c0e11400d6e4fcf909a2005147"
        }
      }
    },
    {
      "arm": "WDT_REPAIR",
      "round": 2,
      "status": "COMPLETE",
      "questions": 40,
      "answers": 320,
      "seed_questions": 20,
      "seed_answers": 160,
      "seed_successes": 125,
      "review_questions": 10,
      "new_questions": 10,
      "effective_questions": 15,
      "final_policy_id": "64bba71af88c1427cc676e23f26a83f3f6f70ead23698dae58c93bd283d64fe7",
      "completed_at_utc": "2026-11-02T02:13:21.055887+00:00",
      "completed_at_beijing": "2026-11-02 10:13:21",
      "started_at_utc": "2026-11-02T00:59:16.683471+00:00",
      "complete_round_seconds": 4444.372416,
      "stage_seconds": {
        "plan-round": 1982.335677,
        "train-round": 2462.02242
      },
      "pure_parameter_training_seconds": null,
      "all_correct_questions": 17,
      "all_wrong_questions": 8,
      "effective_new_questions": 3,
      "new_difficulty_by_kind": {
        "deep_down": {
          "all_failed": 0,
          "all_succeeded": 1,
          "effective_questions": 1,
          "minimum_actions_delta_range": [
            -17,
            -17
          ],
          "questions": 2,
          "required_chain_depth_delta_range": [
            -1,
            -1
          ]
        },
        "deep_up": {
          "all_failed": 3,
          "all_succeeded": 0,
          "effective_questions": 1,
          "minimum_actions_delta_range": [
            1,
            4
          ],
          "questions": 4,
          "required_chain_depth_delta_range": [
            1,
            1
          ]
        },
        "wide": {
          "all_failed": 1,
          "all_succeeded": 2,
          "effective_questions": 1,
          "minimum_actions_delta_range": [
            0,
            1
          ],
          "questions": 4,
          "required_chain_depth_delta_range": [
            0,
            0
          ]
        }
      },
      "signal_coverage": {
        "failed_questions": 16,
        "first_answer_failed_questions": 10,
        "independent_student_answers": 0,
        "later_failures_considered": 6,
        "mixed_source_questions": 12,
        "regressed_question_ids": [
          "generated:local_textcraft_280",
          "generated:local_textcraft_448",
          "generated:local_textcraft_357",
          "generated:cluster_wdt_r1_s4",
          "generated:cluster_wdt_r1_s5",
          "generated:cluster_wdt_r1_s18"
        ],
        "source_questions": 30,
        "upward_sources_skipped_without_teacher_proposals": 1
      },
      "reserved_new_question_positions": 10,
      "main_learning_updates": 1,
      "main_learning_complete": true,
      "model_saved": true,
      "repair": {
        "accepted_by_mode": {
          "prevention": 28,
          "recovery": 13
        },
        "accepted_proposals": 41,
        "optimizer_steps": 1,
        "proposal_count": 57,
        "rejection_reasons": {
          "NO_SUCCESSFUL_RELEVANT_CORRECTION": 1,
          "SUFFIX_DID_NOT_COMPLETE_ORIGINAL_TASK": 15
        },
        "repaired_target_ids": [
          "generated:cluster_wdt_r1_s0",
          "generated:cluster_wdt_r1_s18",
          "generated:cluster_wdt_r1_s4",
          "generated:cluster_wdt_repair_r2_s12",
          "generated:cluster_wdt_repair_r2_s2",
          "generated:cluster_wdt_repair_r2_s3",
          "generated:cluster_wdt_repair_r2_s5",
          "generated:local_textcraft_215",
          "generated:local_textcraft_248",
          "generated:local_textcraft_280",
          "generated:local_textcraft_357",
          "generated:local_textcraft_448",
          "generated:local_textcraft_462",
          "generated:local_textcraft_480",
          "generated:local_textcraft_489",
          "generated:local_textcraft_502"
        ],
        "selected_actions": 65
      },
      "answers_collected_before_second_learning": true,
      "completed_main_learning_rounds_before_collection": 1,
      "answer_meaning": "本轮答卷在第二次学习之前收集，反映第一轮学习后的表现；第二轮模型已经保存。",
      "timing_scope": "完整一轮，包含出题、作答、学习、纠错（如有）、保存和阶段切换。",
      "source_records": {
        "CAPTURE.json": {
          "bytes": 43850,
          "sha256": "d08c7b28c2c8e7f959c8a3b39234c9bcb64dda4a4f2eec6b17ed327d5f989199"
        },
        "complete.json": {
          "bytes": 9348,
          "sha256": "2bad3cd116d11f963246df352be563ae4a5c70501957b5fc37def83e0d193964"
        },
        "collection.json": {
          "bytes": 11662,
          "sha256": "3dae285d9255d10c86bb599c9e832f4d0f0d7d5733506d056f8631789fdb17c5"
        },
        "plan.json": {
          "bytes": 90906,
          "sha256": "ef8012fc618a629be84d4c8ac1a1dbb6990abdc57ad29091d548e1cfad38bcbe"
        },
        "main.json": {
          "bytes": 7690,
          "sha256": "a48a7b6a0280e17a2ccb51f341939339875e2bbab3049bb5482235e302b8a9fb"
        }
      }
    },
    {
      "arm": "WDT_SCALE50",
      "round": 2,
      "status": "COMPLETE",
      "questions": 75,
      "answers": 600,
      "seed_questions": 50,
      "seed_answers": 400,
      "seed_successes": 279,
      "review_questions": 25,
      "new_questions": 0,
      "effective_questions": 40,
      "final_policy_id": "20c9a42e5b230319dadf41a42c85c32b7c148dead5b0a32fe9a499c0b098070f",
      "completed_at_utc": "2026-11-02T03:44:15.742210+00:00",
      "completed_at_beijing": "2026-11-02 11:44:15",
      "started_at_utc": "2026-11-02T02:13:21.078609+00:00",
      "complete_round_seconds": 5454.663601,
      "stage_seconds": {
        "plan-round": 4271.307887,
        "train-round": 1183.324984
      },
      "pure_parameter_training_seconds": null,
      "all_correct_questions": 28,
      "all_wrong_questions": 7,
      "effective_new_questions": 0,
      "new_difficulty_by_kind": {},
      "signal_coverage": {
        "failed_questions": 47,
        "first_answer_failed_questions": 27,
        "independent_student_answers": 0,
        "later_failures_considered": 20,
        "mixed_source_questions": 40,
        "regressed_question_ids": [
          "generated:local_textcraft_248",
          "generated:local_textcraft_280",
          "generated:local_textcraft_448",
          "generated:local_textcraft_357",
          "generated:local_textcraft_190",
          "generated:random_r3_textcraft_219",
          "generated:matched_random_r3_textcraft_282",
          "generated:matched_random_r2_textcraft_304",
          "generated:matched_random_r3_textcraft_350",
          "generated:matched_random_r2_textcraft_401",
          "generated:cluster_wdt_scale50_r1_s12",
          "generated:cluster_wdt_scale50_r1_s11",
          "generated:cluster_wdt_scale50_r1_s10",
          "generated:cluster_wdt_scale50_r1_s20",
          "generated:cluster_wdt_scale50_r1_s35",
          "generated:cluster_wdt_scale50_r1_s1"
        ],
        "source_questions": 75,
        "upward_sources_skipped_without_teacher_proposals": 0
      },
      "reserved_new_question_positions": 25,
      "main_learning_updates": 1,
      "main_learning_complete": true,
      "model_saved": true,
      "repair": null,
      "answers_collected_before_second_learning": true,
      "completed_main_learning_rounds_before_collection": 1,
      "answer_meaning": "本轮答卷在第二次学习之前收集，反映第一轮学习后的表现；第二轮模型已经保存。",
      "timing_scope": "完整一轮，包含出题、作答、学习、纠错（如有）、保存和阶段切换。",
      "source_records": {
        "CAPTURE.json": {
          "bytes": 81497,
          "sha256": "91e067ae2740e14e5df137b4df9fbc001f52428b422b65b7df3d1c5be57bd069"
        },
        "complete.json": {
          "bytes": 13748,
          "sha256": "3a3719b6addf1e63cdbfbcbeea0f32e15ee74e41c34011bb198dc13ca9e63ad5"
        },
        "collection.json": {
          "bytes": 21637,
          "sha256": "9d7fc5d35cc7a9980483f1231b04b63ada06f4a01d5b47c916fbac52e6ee613a"
        },
        "plan.json": {
          "bytes": 126053,
          "sha256": "6b3cc77887a2b7503bad20e03b4a6242b8a30475a7e17a641b270d26d2613d34"
        },
        "main.json": {
          "bytes": 13037,
          "sha256": "d94176b7a4f9d90b43afb37b46d8e6394f2ae7192519fd6961ebe70b4e8dcb1f"
        }
      },
      "new_question_issue": {
        "valid_individual_diagnoses": 55,
        "partition_assigned_diagnoses": 54,
        "omitted_diagnosis_id": "D44",
        "generation_candidate_requests": 0,
        "reason": "老师分组漏掉一条诊断，整份分组被拒，55 份有效单题诊断未进入出题。",
        "training_continued_on_existing_questions": true,
        "successful_expansion_this_round": false,
        "evidence": {
          "bytes": 1367166,
          "sha256": "b710a6d99f7f90af62c0fe62af62e5529354b0c5456e415d7e79660b9ef59f39"
        }
      }
    }
  ],
  "same_30_old_questions": {
    "questions": 30,
    "answers_each": 240,
    "successes": {
      "OPT": 182,
      "REPAIR": 178
    },
    "task_ids": [
      "generated:cluster_wdt_r1_s0",
      "generated:cluster_wdt_r1_s1",
      "generated:cluster_wdt_r1_s12",
      "generated:cluster_wdt_r1_s16",
      "generated:cluster_wdt_r1_s17",
      "generated:cluster_wdt_r1_s18",
      "generated:cluster_wdt_r1_s3",
      "generated:cluster_wdt_r1_s4",
      "generated:cluster_wdt_r1_s5",
      "generated:cluster_wdt_r1_s7",
      "generated:local_textcraft_100",
      "generated:local_textcraft_188",
      "generated:local_textcraft_190",
      "generated:local_textcraft_215",
      "generated:local_textcraft_248",
      "generated:local_textcraft_273",
      "generated:local_textcraft_280",
      "generated:local_textcraft_328",
      "generated:local_textcraft_355",
      "generated:local_textcraft_357",
      "generated:local_textcraft_42",
      "generated:local_textcraft_448",
      "generated:local_textcraft_45",
      "generated:local_textcraft_462",
      "generated:local_textcraft_477",
      "generated:local_textcraft_480",
      "generated:local_textcraft_489",
      "generated:local_textcraft_502",
      "generated:local_textcraft_517",
      "generated:local_textcraft_72"
    ]
  },
  "same_20_seed_questions": {
    "questions": 20,
    "answers_each": 160,
    "successes": {
      "OPT": 125,
      "REPAIR": 125
    }
  },
  "repair_effect": "共同旧题上尚未看到纠错组的优势；红旗首轮老师修复 3/3 后，第二轮学生仍为 0/8。",
  "final_test_results_available": false
}
