{
 "protocol": "heldout_v1",
 "generated": {
  "timestamp": "2026-11-04T00:02:26+00:00",
  "code_version": "d1aa8dc-dirty",
  "host": {
   "platform": "macOS-27.0-arm64-arm-64bit",
   "python": "3.12.13",
   "machine": "arm64",
   "node": "MacBookPro-de-simon-pierre-2.local"
  }
 },
 "n_runs": 199,
 "n_runs_ok": 199,
 "n_judgments": 33,
 "hypotheses": {
  "H1_D_vs_B": {
   "clarity": {
    "n": 24,
    "mean_diff": 0.042,
    "ci95": [
     -0.406,
     0.281
    ],
    "wilcoxon_p": 0.7920689965469446,
    "cliffs_delta": 0.014,
    "holm_p": 1.0,
    "significant": false
   },
   "argument_logic": {
    "n": 24,
    "mean_diff": 0.333,
    "ci95": [
     -0.219,
     0.656
    ],
    "wilcoxon_p": 0.1120871657779208,
    "cliffs_delta": 0.203,
    "holm_p": 0.4483,
    "significant": false
   },
   "contribution_positioning": {
    "n": 24,
    "mean_diff": 0.083,
    "ci95": [
     -0.406,
     0.438
    ],
    "wilcoxon_p": 0.9665165964249498,
    "cliffs_delta": 0.023,
    "holm_p": 0.9665,
    "significant": false
   },
   "usefulness": {
    "n": 24,
    "mean_diff": 0.167,
    "ci95": [
     -0.5,
     0.688
    ],
    "wilcoxon_p": 0.5747134367649175,
    "cliffs_delta": 0.007,
    "holm_p": 1.0,
    "significant": false
   }
  },
  "H2_D_vs_C": {
   "clarity": {
    "n": 24,
    "mean_diff": 0.0,
    "ci95": [
     -0.281,
     0.25
    ],
    "wilcoxon_p": 1.0,
    "cliffs_delta": 0.005,
    "holm_p": 1.0,
    "significant": false
   },
   "argument_logic": {
    "n": 24,
    "mean_diff": 0.083,
    "ci95": [
     -0.438,
     0.469
    ],
    "wilcoxon_p": 0.7506546409575028,
    "cliffs_delta": 0.0,
    "holm_p": 1.0,
    "significant": false
   },
   "contribution_positioning": {
    "n": 24,
    "mean_diff": -0.208,
    "ci95": [
     -0.719,
     0.156
    ],
    "wilcoxon_p": 0.15271894732542726,
    "cliffs_delta": -0.13,
    "holm_p": 0.6109,
    "significant": false
   },
   "usefulness": {
    "n": 24,
    "mean_diff": 0.042,
    "ci95": [
     -0.5,
     0.438
    ],
    "wilcoxon_p": 0.9426906297302409,
    "cliffs_delta": -0.068,
    "holm_p": 1.0,
    "significant": false
   }
  },
  "H3_E_vs_D": {
   "methodological_rigor": {
    "n": 24,
    "mean_diff": 1.333,
    "ci95": [
     1.0,
     1.845
    ],
    "wilcoxon_p": 9.280186701164916e-05,
    "cliffs_delta": 0.677,
    "holm_p": 0.0003,
    "significant": true
   },
   "fidelity": {
    "n": 24,
    "mean_diff": 0.625,
    "ci95": [
     0.312,
     1.031
    ],
    "wilcoxon_p": 0.0019691496770990976,
    "cliffs_delta": 0.359,
    "holm_p": 0.0039,
    "significant": true
   },
   "usefulness": {
    "n": 24,
    "mean_diff": 0.25,
    "ci95": [
     -0.031,
     0.656
    ],
    "wilcoxon_p": 0.10880943004054569,
    "cliffs_delta": 0.167,
    "holm_p": 0.1088,
    "significant": false
   }
  },
  "E_vs_A": {
   "clarity": {
    "n": 24,
    "mean_diff": -0.042,
    "ci95": [
     -0.406,
     0.281
    ],
    "wilcoxon_p": 0.7815112949987133,
    "cliffs_delta": -0.059,
    "holm_p": 0.7815,
    "significant": false
   },
   "argument_logic": {
    "n": 24,
    "mean_diff": 0.917,
    "ci95": [
     0.406,
     1.438
    ],
    "wilcoxon_p": 0.0008241896037559798,
    "cliffs_delta": 0.516,
    "holm_p": 0.0041,
    "significant": true
   },
   "contribution_positioning": {
    "n": 24,
    "mean_diff": 0.5,
    "ci95": [
     0.125,
     0.907
    ],
    "wilcoxon_p": 0.018735985158564714,
    "cliffs_delta": 0.198,
    "holm_p": 0.0562,
    "significant": false
   },
   "methodological_rigor": {
    "n": 24,
    "mean_diff": 2.042,
    "ci95": [
     1.438,
     2.625
    ],
    "wilcoxon_p": 4.6023226073469e-05,
    "cliffs_delta": 0.847,
    "holm_p": 0.0003,
    "significant": true
   },
   "fidelity": {
    "n": 24,
    "mean_diff": 0.417,
    "ci95": [
     0.156,
     0.781
    ],
    "wilcoxon_p": 0.012419330651552268,
    "cliffs_delta": 0.281,
    "holm_p": 0.0497,
    "significant": true
   },
   "usefulness": {
    "n": 24,
    "mean_diff": 0.458,
    "ci95": [
     -0.094,
     1.031
    ],
    "wilcoxon_p": 0.08329941488171003,
    "cliffs_delta": 0.163,
    "holm_p": 0.1666,
    "significant": false
   }
  },
  "E_vs_B": {
   "clarity": {
    "n": 24,
    "mean_diff": -0.042,
    "ci95": [
     -0.406,
     0.25
    ],
    "wilcoxon_p": 0.8127423725657857,
    "cliffs_delta": -0.05,
    "holm_p": 0.8127,
    "significant": false
   },
   "argument_logic": {
    "n": 24,
    "mean_diff": 0.792,
    "ci95": [
     0.406,
     1.062
    ],
    "wilcoxon_p": 0.0002901870694228414,
    "cliffs_delta": 0.483,
    "holm_p": 0.0015,
    "significant": true
   },
   "contribution_positioning": {
    "n": 24,
    "mean_diff": 0.25,
    "ci95": [
     -0.125,
     0.688
    ],
    "wilcoxon_p": 0.2720394478957354,
    "cliffs_delta": 0.089,
    "holm_p": 0.5441,
    "significant": false
   },
   "methodological_rigor": {
    "n": 24,
    "mean_diff": 1.667,
    "ci95": [
     1.125,
     2.188
    ],
    "wilcoxon_p": 8.456524858492154e-05,
    "cliffs_delta": 0.741,
    "holm_p": 0.0005,
    "significant": true
   },
   "fidelity": {
    "n": 24,
    "mean_diff": 1.042,
    "ci95": [
     0.781,
     1.438
    ],
    "wilcoxon_p": 0.0003061844202490992,
    "cliffs_delta": 0.573,
    "holm_p": 0.0012,
    "significant": true
   },
   "usefulness": {
    "n": 24,
    "mean_diff": 0.417,
    "ci95": [
     -0.125,
     0.969
    ],
    "wilcoxon_p": 0.11972288596282951,
    "cliffs_delta": 0.149,
    "holm_p": 0.3592,
    "significant": false
   }
  }
 },
 "ablations": {
  "E_no_retrieval": {
   "clarity": {
    "n": 16,
    "mean_diff": -0.125,
    "ci95": [
     -0.375,
     0.125
    ],
    "wilcoxon_p": 0.31731050786291415,
    "cliffs_delta": -0.117
   },
   "argument_logic": {
    "n": 16,
    "mean_diff": 0.188,
    "ci95": [
     -0.062,
     0.438
    ],
    "wilcoxon_p": 0.17971249487899976,
    "cliffs_delta": 0.141
   },
   "contribution_positioning": {
    "n": 16,
    "mean_diff": -0.062,
    "ci95": [
     -0.25,
     0.125
    ],
    "wilcoxon_p": 0.5637028616507731,
    "cliffs_delta": -0.047
   },
   "methodological_rigor": {
    "n": 16,
    "mean_diff": 0.125,
    "ci95": [
     -0.125,
     0.375
    ],
    "wilcoxon_p": 0.4142161782425253,
    "cliffs_delta": 0.109
   },
   "fidelity": {
    "n": 16,
    "mean_diff": 0.062,
    "ci95": [
     -0.312,
     0.438
    ],
    "wilcoxon_p": 0.7388826803635273,
    "cliffs_delta": 0.051
   },
   "usefulness": {
    "n": 16,
    "mean_diff": 0.0,
    "ci95": [
     -0.25,
     0.25
    ],
    "wilcoxon_p": 1.0,
    "cliffs_delta": 0.02
   }
  },
  "E_no_profile": {
   "clarity": {
    "n": 16,
    "mean_diff": -0.188,
    "ci95": [
     -0.562,
     0.125
    ],
    "wilcoxon_p": 0.31731050786291415,
    "cliffs_delta": -0.121
   },
   "argument_logic": {
    "n": 16,
    "mean_diff": 0.062,
    "ci95": [
     -0.25,
     0.375
    ],
    "wilcoxon_p": 0.7054569861112734,
    "cliffs_delta": 0.008
   },
   "contribution_positioning": {
    "n": 16,
    "mean_diff": -0.125,
    "ci95": [
     -0.438,
     0.188
    ],
    "wilcoxon_p": 0.4142161782425253,
    "cliffs_delta": -0.07
   },
   "methodological_rigor": {
    "n": 16,
    "mean_diff": 0.062,
    "ci95": [
     -0.312,
     0.5
    ],
    "wilcoxon_p": 0.763024600552995,
    "cliffs_delta": 0.02
   },
   "fidelity": {
    "n": 16,
    "mean_diff": 0.0,
    "ci95": [
     -0.438,
     0.438
    ],
    "wilcoxon_p": 1.0,
    "cliffs_delta": 0.008
   },
   "usefulness": {
    "n": 16,
    "mean_diff": -0.062,
    "ci95": [
     -0.312,
     0.188
    ],
    "wilcoxon_p": 0.6547208460185769,
    "cliffs_delta": -0.031
   }
  },
  "E_no_referee": {
   "clarity": {
    "n": 16,
    "mean_diff": -0.125,
    "ci95": [
     -0.438,
     0.188
    ],
    "wilcoxon_p": 0.4142161782425253,
    "cliffs_delta": -0.062
   },
   "argument_logic": {
    "n": 16,
    "mean_diff": 0.688,
    "ci95": [
     0.312,
     1.125
    ],
    "wilcoxon_p": 0.00838632096179826,
    "cliffs_delta": 0.43
   },
   "contribution_positioning": {
    "n": 16,
    "mean_diff": 0.25,
    "ci95": [
     -0.062,
     0.562
    ],
    "wilcoxon_p": 0.15729920705028516,
    "cliffs_delta": 0.09
   },
   "methodological_rigor": {
    "n": 16,
    "mean_diff": 1.625,
    "ci95": [
     1.0,
     2.188
    ],
    "wilcoxon_p": 0.0011302166601564717,
    "cliffs_delta": 0.781
   },
   "fidelity": {
    "n": 16,
    "mean_diff": 0.5,
    "ci95": [
     0.125,
     0.814
    ],
    "wilcoxon_p": 0.02092133533779403,
    "cliffs_delta": 0.289
   },
   "usefulness": {
    "n": 16,
    "mean_diff": 0.438,
    "ci95": [
     0.0,
     0.938
    ],
    "wilcoxon_p": 0.08326451666355039,
    "cliffs_delta": 0.246
   }
  },
  "E_no_factcheck": {
   "clarity": {
    "n": 16,
    "mean_diff": 0.0,
    "ci95": [
     -0.438,
     0.438
    ],
    "wilcoxon_p": 1.0,
    "cliffs_delta": 0.027
   },
   "argument_logic": {
    "n": 16,
    "mean_diff": 0.125,
    "ci95": [
     -0.25,
     0.438
    ],
    "wilcoxon_p": 0.47950012218695337,
    "cliffs_delta": 0.094
   },
   "contribution_positioning": {
    "n": 16,
    "mean_diff": 0.125,
    "ci95": [
     -0.125,
     0.375
    ],
    "wilcoxon_p": 0.31731050786291415,
    "cliffs_delta": 0.066
   },
   "methodological_rigor": {
    "n": 16,
    "mean_diff": 0.25,
    "ci95": [
     0.0,
     0.5
    ],
    "wilcoxon_p": 0.10247043485974945,
    "cliffs_delta": 0.113
   },
   "fidelity": {
    "n": 16,
    "mean_diff": 0.0,
    "ci95": [
     -0.25,
     0.25
    ],
    "wilcoxon_p": 1.0,
    "cliffs_delta": -0.008
   },
   "usefulness": {
    "n": 16,
    "mean_diff": 0.125,
    "ci95": [
     -0.188,
     0.438
    ],
    "wilcoxon_p": 0.47950012218695337,
    "cliffs_delta": 0.102
   }
  }
 },
 "judge_means": {
  "A_baseline": {
   "clarity": 5.67,
   "argument_logic": 5.04,
   "contribution_positioning": 4.12,
   "methodological_rigor": 4.17,
   "fidelity": 5.83,
   "usefulness": 4.83,
   "n_invented_facts": 0.79,
   "n": 24
  },
  "B_strong_prompt": {
   "clarity": 5.67,
   "argument_logic": 5.17,
   "contribution_positioning": 4.38,
   "methodological_rigor": 4.54,
   "fidelity": 5.21,
   "usefulness": 4.88,
   "n_invented_facts": 1.54,
   "n": 24
  },
  "C_conventional_rag": {
   "clarity": 5.71,
   "argument_logic": 5.42,
   "contribution_positioning": 4.67,
   "methodological_rigor": 4.58,
   "fidelity": 5.29,
   "usefulness": 5.0,
   "n_invented_facts": 1.33,
   "n": 24
  },
  "D_profiles_retrieval": {
   "clarity": 5.71,
   "argument_logic": 5.5,
   "contribution_positioning": 4.46,
   "methodological_rigor": 4.88,
   "fidelity": 5.62,
   "usefulness": 5.04,
   "n_invented_facts": 1.08,
   "n": 24
  },
  "E_full_engine": {
   "clarity": 5.62,
   "argument_logic": 5.96,
   "contribution_positioning": 4.62,
   "methodological_rigor": 6.21,
   "fidelity": 6.25,
   "usefulness": 5.29,
   "n_invented_facts": 0.42,
   "n": 24
  },
  "E_no_factcheck": {
   "clarity": 5.56,
   "argument_logic": 5.81,
   "contribution_positioning": 4.44,
   "methodological_rigor": 5.94,
   "fidelity": 6.12,
   "usefulness": 5.19,
   "n_invented_facts": 0.62,
   "n": 16
  },
  "E_no_profile": {
   "clarity": 5.75,
   "argument_logic": 5.88,
   "contribution_positioning": 4.69,
   "methodological_rigor": 6.12,
   "fidelity": 6.12,
   "usefulness": 5.38,
   "n_invented_facts": 0.69,
   "n": 16
  },
  "E_no_referee": {
   "clarity": 5.69,
   "argument_logic": 5.25,
   "contribution_positioning": 4.31,
   "methodological_rigor": 4.56,
   "fidelity": 5.62,
   "usefulness": 4.88,
   "n_invented_facts": 1.19,
   "n": 16
  },
  "E_no_retrieval": {
   "clarity": 5.69,
   "argument_logic": 5.75,
   "contribution_positioning": 4.62,
   "methodological_rigor": 6.06,
   "fidelity": 6.06,
   "usefulness": 5.31,
   "n_invented_facts": 0.69,
   "n": 16
  }
 },
 "task_means": {
  "('draft_introduction', 'A_baseline')": {
   "clarity": 6.25,
   "argument_logic": 5.75,
   "contribution_positioning": 5.0,
   "methodological_rigor": 4.25,
   "fidelity": 5.5,
   "usefulness": 5.75
  },
  "('draft_introduction', 'B_strong_prompt')": {
   "clarity": 6.0,
   "argument_logic": 5.5,
   "contribution_positioning": 5.0,
   "methodological_rigor": 4.25,
   "fidelity": 5.25,
   "usefulness": 5.5
  },
  "('draft_introduction', 'C_conventional_rag')": {
   "clarity": 6.0,
   "argument_logic": 5.75,
   "contribution_positioning": 5.5,
   "methodological_rigor": 4.75,
   "fidelity": 5.0,
   "usefulness": 5.5
  },
  "('draft_introduction', 'D_profiles_retrieval')": {
   "clarity": 6.0,
   "argument_logic": 6.0,
   "contribution_positioning": 5.0,
   "methodological_rigor": 5.25,
   "fidelity": 4.75,
   "usefulness": 5.5
  },
  "('draft_introduction', 'E_full_engine')": {
   "clarity": 5.75,
   "argument_logic": 6.5,
   "contribution_positioning": 5.0,
   "methodological_rigor": 6.5,
   "fidelity": 6.0,
   "usefulness": 6.0
  },
  "('draft_introduction', 'E_no_factcheck')": {
   "clarity": 5.5,
   "argument_logic": 5.5,
   "contribution_positioning": 5.0,
   "methodological_rigor": 6.0,
   "fidelity": 5.5,
   "usefulness": 5.5
  },
  "('draft_introduction', 'E_no_profile')": {
   "clarity": 5.5,
   "argument_logic": 6.0,
   "contribution_positioning": 4.5,
   "methodological_rigor": 6.0,
   "fidelity": 6.5,
   "usefulness": 5.5
  },
  "('draft_introduction', 'E_no_referee')": {
   "clarity": 6.0,
   "argument_logic": 6.5,
   "contribution_positioning": 5.0,
   "methodological_rigor": 5.0,
   "fidelity": 4.5,
   "usefulness": 5.0
  },
  "('draft_introduction', 'E_no_retrieval')": {
   "clarity": 6.0,
   "argument_logic": 6.0,
   "contribution_positioning": 5.0,
   "methodological_rigor": 7.0,
   "fidelity": 5.5,
   "usefulness": 6.0
  },
  "('explain_specification', 'A_baseline')": {
   "clarity": 6.0,
   "argument_logic": 5.0,
   "contribution_positioning": 3.5,
   "methodological_rigor": 4.5,
   "fidelity": 5.5,
   "usefulness": 4.5
  },
  "('explain_specification', 'B_strong_prompt')": {
   "clarity": 6.0,
   "argument_logic": 5.0,
   "contribution_positioning": 3.5,
   "methodological_rigor": 4.5,
   "fidelity": 5.5,
   "usefulness": 4.5
  },
  "('explain_specification', 'C_conventional_rag')": {
   "clarity": 5.5,
   "argument_logic": 5.0,
   "contribution_positioning": 4.5,
   "methodological_rigor": 4.0,
   "fidelity": 5.0,
   "usefulness": 4.5
  },
  "('explain_specification', 'D_profiles_retrieval')": {
   "clarity": 6.0,
   "argument_logic": 5.5,
   "contribution_positioning": 3.5,
   "methodological_rigor": 4.5,
   "fidelity": 5.5,
   "usefulness": 4.5
  },
  "('explain_specification', 'E_full_engine')": {
   "clarity": 5.5,
   "argument_logic": 5.5,
   "contribution_positioning": 3.5,
   "methodological_rigor": 6.0,
   "fidelity": 6.5,
   "usefulness": 5.0
  },
  "('explain_specification', 'E_no_factcheck')": {
   "clarity": 6.5,
   "argument_logic": 6.0,
   "contribution_positioning": 3.5,
   "methodological_rigor": 5.0,
   "fidelity": 6.0,
   "usefulness": 5.0
  },
  "('explain_specification', 'E_no_profile')": {
   "clarity": 6.0,
   "argument_logic": 6.0,
   "contribution_positioning": 3.5,
   "methodological_rigor": 6.0,
   "fidelity": 5.0,
   "usefulness": 5.5
  },
  "('explain_specification', 'E_no_referee')": {
   "clarity": 6.0,
   "argument_logic": 5.0,
   "contribution_positioning": 3.5,
   "methodological_rigor": 4.5,
   "fidelity": 6.0,
   "usefulness": 5.0
  },
  "('explain_specification', 'E_no_retrieval')": {
   "clarity": 6.0,
   "argument_logic": 5.5,
   "contribution_positioning": 3.5,
   "methodological_rigor": 5.5,
   "fidelity": 5.5,
   "usefulness": 5.0
  },
  "('flag_causal_claims', 'A_baseline')": {
   "clarity": 5.5,
   "argument_logic": 5.0,
   "contribution_positioning": 3.5,
   "methodological_rigor": 5.0,
   "fidelity": 6.0,
   "usefulness": 4.5
  },
  "('flag_causal_claims', 'B_strong_prompt')": {
   "clarity": 6.0,
   "argument_logic": 6.0,
   "contribution_positioning": 4.0,
   "methodological_rigor": 6.0,
   "fidelity": 5.5,
   "usefulness": 5.5
  },
  "('flag_causal_claims', 'C_conventional_rag')": {
   "clarity": 6.0,
   "argument_logic": 6.0,
   "contribution_positioning": 4.5,
   "methodological_rigor": 6.0,
   "fidelity": 6.0,
   "usefulness": 5.0
  },
  "('flag_causal_claims', 'D_profiles_retrieval')": {
   "clarity": 5.5,
   "argument_logic": 5.0,
   "contribution_positioning": 3.5,
   "methodological_rigor": 5.0,
   "fidelity": 6.0,
   "usefulness": 4.0
  },
  "('flag_causal_claims', 'E_full_engine')": {
   "clarity": 6.0,
   "argument_logic": 6.5,
   "contribution_positioning": 4.0,
   "methodological_rigor": 6.5,
   "fidelity": 6.5,
   "usefulness": 5.0
  },
  "('flag_causal_claims', 'E_no_factcheck')": {
   "clarity": 5.5,
   "argument_logic": 6.0,
   "contribution_positioning": 4.0,
   "methodological_rigor": 6.0,
   "fidelity": 6.0,
   "usefulness": 5.0
  },
  "('flag_causal_claims', 'E_no_profile')": {
   "clarity": 6.0,
   "argument_logic": 6.0,
   "contribution_positioning": 5.0,
   "methodological_rigor": 6.5,
   "fidelity": 6.5,
   "usefulness": 5.0
  },
  "('flag_causal_claims', 'E_no_referee')": {
   "clarity": 5.5,
   "argument_logic": 5.5,
   "contribution_positioning": 4.0,
   "methodological_rigor": 6.0,
   "fidelity": 6.0,
   "usefulness": 4.5
  },
  "('flag_causal_claims', 'E_no_retrieval')": {
   "clarity": 6.0,
   "argument_logic": 6.0,
   "contribution_positioning": 4.0,
   "methodological_rigor": 6.0,
   "fidelity": 6.5,
   "usefulness": 5.0
  },
  "('improve_abstract', 'A_baseline')": {
   "clarity": 5.0,
   "argument_logic": 5.25,
   "contribution_positioning": 4.5,
   "methodological_rigor": 4.0,
   "fidelity": 6.25,
   "usefulness": 4.75
  },
  "('improve_abstract', 'B_strong_prompt')": {
   "clarity": 5.0,
   "argument_logic": 5.0,
   "contribution_positioning": 5.0,
   "methodological_rigor": 4.25,
   "fidelity": 5.75,
   "usefulness": 4.75
  },
  "('improve_abstract', 'C_conventional_rag')": {
   "clarity": 5.5,
   "argument_logic": 5.5,
   "contribution_positioning": 5.25,
   "methodological_rigor": 4.25,
   "fidelity": 6.0,
   "usefulness": 5.25
  },
  "('improve_abstract', 'D_profiles_retrieval')": {
   "clarity": 5.25,
   "argument_logic": 5.25,
   "contribution_positioning": 4.75,
   "methodological_rigor": 4.25,
   "fidelity": 6.25,
   "usefulness": 5.0
  },
  "('improve_abstract', 'E_full_engine')": {
   "clarity": 5.0,
   "argument_logic": 5.25,
   "contribution_positioning": 4.75,
   "methodological_rigor": 5.25,
   "fidelity": 6.75,
   "usefulness": 4.75
  },
  "('improve_abstract', 'E_no_factcheck')": {
   "clarity": 5.0,
   "argument_logic": 6.0,
   "contribution_positioning": 5.0,
   "methodological_rigor": 5.0,
   "fidelity": 6.5,
   "usefulness": 5.0
  },
  "('improve_abstract', 'E_no_profile')": {
   "clarity": 5.5,
   "argument_logic": 5.5,
   "contribution_positioning": 5.0,
   "methodological_rigor": 5.0,
   "fidelity": 6.0,
   "usefulness": 5.0
  },
  "('improve_abstract', 'E_no_referee')": {
   "clarity": 5.0,
   "argument_logic": 5.0,
   "contribution_positioning": 4.5,
   "methodological_rigor": 4.0,
   "fidelity": 6.5,
   "usefulness": 5.0
  },
  "('improve_abstract', 'E_no_retrieval')": {
   "clarity": 4.5,
   "argument_logic": 5.5,
   "contribution_positioning": 5.0,
   "methodological_rigor": 5.5,
   "fidelity": 6.5,
   "usefulness": 5.0
  },
  "('improve_table', 'A_baseline')": {
   "clarity": 5.5,
   "argument_logic": 5.0,
   "contribution_positioning": 4.0,
   "methodological_rigor": 4.0,
   "fidelity": 6.0,
   "usefulness": 5.5
  },
  "('improve_table', 'B_strong_prompt')": {
   "clarity": 5.5,
   "argument_logic": 5.0,
   "contribution_positioning": 4.0,
   "methodological_rigor": 4.0,
   "fidelity": 5.5,
   "usefulness": 5.0
  },
  "('improve_table', 'C_conventional_rag')": {
   "clarity": 5.5,
   "argument_logic": 4.5,
   "contribution_positioning": 3.5,
   "methodological_rigor": 4.0,
   "fidelity": 5.0,
   "usefulness": 5.0
  },
  "('improve_table', 'D_profiles_retrieval')": {
   "clarity": 5.0,
   "argument_logic": 4.5,
   "contribution_positioning": 3.5,
   "methodological_rigor": 4.5,
   "fidelity": 5.5,
   "usefulness": 5.0
  },
  "('improve_table', 'E_full_engine')": {
   "clarity": 5.5,
   "argument_logic": 5.5,
   "contribution_positioning": 4.5,
   "methodological_rigor": 6.0,
   "fidelity": 6.5,
   "usefulness": 5.5
  },
  "('improve_table', 'E_no_factcheck')": {
   "clarity": 4.5,
   "argument_logic": 5.5,
   "contribution_positioning": 3.5,
   "methodological_rigor": 6.0,
   "fidelity": 6.5,
   "usefulness": 4.5
  },
  "('improve_table', 'E_no_profile')": {
   "clarity": 5.0,
   "argument_logic": 5.5,
   "contribution_positioning": 4.0,
   "methodological_rigor": 6.5,
   "fidelity": 6.5,
   "usefulness": 5.0
  },
  "('improve_table', 'E_no_referee')": {
   "clarity": 6.0,
   "argument_logic": 5.0,
   "contribution_positioning": 4.0,
   "methodological_rigor": 4.5,
   "fidelity": 5.5,
   "usefulness": 6.0
  },
  "('improve_table', 'E_no_retrieval')": {
   "clarity": 5.5,
   "argument_logic": 5.5,
   "contribution_positioning": 4.5,
   "methodological_rigor": 6.0,
   "fidelity": 6.0,
   "usefulness": 5.0
  },
  "('interpret_results', 'A_baseline')": {
   "clarity": 6.0,
   "argument_logic": 5.0,
   "contribution_positioning": 4.0,
   "methodological_rigor": 4.5,
   "fidelity": 6.0,
   "usefulness": 5.0
  },
  "('interpret_results', 'B_strong_prompt')": {
   "clarity": 6.0,
   "argument_logic": 5.25,
   "contribution_positioning": 4.5,
   "methodological_rigor": 5.0,
   "fidelity": 5.0,
   "usefulness": 5.0
  },
  "('interpret_results', 'C_conventional_rag')": {
   "clarity": 6.0,
   "argument_logic": 6.0,
   "contribution_positioning": 4.5,
   "methodological_rigor": 5.25,
   "fidelity": 5.0,
   "usefulness": 5.25
  },
  "('interpret_results', 'D_profiles_retrieval')": {
   "clarity": 6.0,
   "argument_logic": 5.5,
   "contribution_positioning": 4.25,
   "methodological_rigor": 5.0,
   "fidelity": 5.5,
   "usefulness": 5.0
  },
  "('interpret_results', 'E_full_engine')": {
   "clarity": 5.75,
   "argument_logic": 6.25,
   "contribution_positioning": 4.5,
   "methodological_rigor": 6.75,
   "fidelity": 5.75,
   "usefulness": 5.0
  },
  "('interpret_results', 'E_no_factcheck')": {
   "clarity": 6.0,
   "argument_logic": 6.0,
   "contribution_positioning": 4.5,
   "methodological_rigor": 7.0,
   "fidelity": 6.5,
   "usefulness": 5.5
  },
  "('interpret_results', 'E_no_profile')": {
   "clarity": 6.0,
   "argument_logic": 6.0,
   "contribution_positioning": 4.5,
   "methodological_rigor": 6.0,
   "fidelity": 6.5,
   "usefulness": 5.5
  },
  "('interpret_results', 'E_no_referee')": {
   "clarity": 6.0,
   "argument_logic": 5.0,
   "contribution_positioning": 4.5,
   "methodological_rigor": 5.0,
   "fidelity": 6.0,
   "usefulness": 5.0
  },
  "('interpret_results', 'E_no_retrieval')": {
   "clarity": 6.0,
   "argument_logic": 6.0,
   "contribution_positioning": 4.5,
   "methodological_rigor": 6.5,
   "fidelity": 6.5,
   "usefulness": 5.5
  },
  "('referee_report', 'A_baseline')": {
   "clarity": 5.25,
   "argument_logic": 4.25,
   "contribution_positioning": 3.25,
   "methodological_rigor": 3.75,
   "fidelity": 6.0,
   "usefulness": 3.75
  },
  "('referee_report', 'B_strong_prompt')": {
   "clarity": 5.25,
   "argument_logic": 4.75,
   "contribution_positioning": 3.25,
   "methodological_rigor": 4.5,
   "fidelity": 5.25,
   "usefulness": 4.0
  },
  "('referee_report', 'C_conventional_rag')": {
   "clarity": 5.25,
   "argument_logic": 4.5,
   "contribution_positioning": 3.5,
   "methodological_rigor": 4.0,
   "fidelity": 5.25,
   "usefulness": 3.75
  },
  "('referee_report', 'D_profiles_retrieval')": {
   "clarity": 5.75,
   "argument_logic": 6.0,
   "contribution_positioning": 4.75,
   "methodological_rigor": 5.75,
   "fidelity": 6.25,
   "usefulness": 5.25
  },
  "('referee_report', 'E_full_engine')": {
   "clarity": 6.0,
   "argument_logic": 6.25,
   "contribution_positioning": 4.75,
   "methodological_rigor": 6.25,
   "fidelity": 6.5,
   "usefulness": 5.5
  },
  "('referee_report', 'E_no_factcheck')": {
   "clarity": 5.5,
   "argument_logic": 5.5,
   "contribution_positioning": 4.0,
   "methodological_rigor": 6.0,
   "fidelity": 6.5,
   "usefulness": 5.0
  },
  "('referee_report', 'E_no_profile')": {
   "clarity": 6.0,
   "argument_logic": 6.0,
   "contribution_positioning": 5.0,
   "methodological_rigor": 6.5,
   "fidelity": 6.0,
   "usefulness": 5.5
  },
  "('referee_report', 'E_no_referee')": {
   "clarity": 5.5,
   "argument_logic": 4.5,
   "contribution_positioning": 3.5,
   "methodological_rigor": 4.0,
   "fidelity": 6.0,
   "usefulness": 3.5
  },
  "('referee_report', 'E_no_retrieval')": {
   "clarity": 6.0,
   "argument_logic": 6.0,
   "contribution_positioning": 5.0,
   "methodological_rigor": 6.0,
   "fidelity": 6.5,
   "usefulness": 5.5
  },
  "('rewrite_literature', 'A_baseline')": {
   "clarity": 6.0,
   "argument_logic": 5.0,
   "contribution_positioning": 5.0,
   "methodological_rigor": 3.5,
   "fidelity": 5.0,
   "usefulness": 5.0
  },
  "('rewrite_literature', 'B_strong_prompt')": {
   "clarity": 6.0,
   "argument_logic": 5.0,
   "contribution_positioning": 5.5,
   "methodological_rigor": 4.0,
   "fidelity": 3.5,
   "usefulness": 5.0
  },
  "('rewrite_literature', 'C_conventional_rag')": {
   "clarity": 6.0,
   "argument_logic": 6.0,
   "contribution_positioning": 6.0,
   "methodological_rigor": 4.5,
   "fidelity": 5.0,
   "usefulness": 6.0
  },
  "('rewrite_literature', 'D_profiles_retrieval')": {
   "clarity": 6.0,
   "argument_logic": 5.5,
   "contribution_positioning": 5.5,
   "methodological_rigor": 4.0,
   "fidelity": 5.0,
   "usefulness": 5.5
  },
  "('rewrite_literature', 'E_full_engine')": {
   "clarity": 5.5,
   "argument_logic": 5.5,
   "contribution_positioning": 5.5,
   "methodological_rigor": 6.5,
   "fidelity": 5.5,
   "usefulness": 5.5
  },
  "('rewrite_literature', 'E_no_factcheck')": {
   "clarity": 6.0,
   "argument_logic": 6.0,
   "contribution_positioning": 6.0,
   "methodological_rigor": 6.5,
   "fidelity": 5.5,
   "usefulness": 6.0
  },
  "('rewrite_literature', 'E_no_profile')": {
   "clarity": 6.0,
   "argument_logic": 6.0,
   "contribution_positioning": 6.0,
   "methodological_rigor": 6.5,
   "fidelity": 6.0,
   "usefulness": 6.0
  },
  "('rewrite_literature', 'E_no_referee')": {
   "clarity": 5.5,
   "argument_logic": 5.5,
   "contribution_positioning": 5.5,
   "methodological_rigor": 3.5,
   "fidelity": 4.5,
   "usefulness": 5.0
  },
  "('rewrite_literature', 'E_no_retrieval')": {
   "clarity": 5.5,
   "argument_logic": 5.5,
   "contribution_positioning": 5.5,
   "methodological_rigor": 6.0,
   "fidelity": 5.5,
   "usefulness": 5.5
  }
 },
 "out_of_domain_means": {
  "A_baseline": {
   "clarity": 5.67,
   "argument_logic": 5.33,
   "contribution_positioning": 3.67,
   "methodological_rigor": 4.0,
   "fidelity": 5.33,
   "usefulness": 4.0
  },
  "B_strong_prompt": {
   "clarity": 5.67,
   "argument_logic": 5.33,
   "contribution_positioning": 4.0,
   "methodological_rigor": 3.67,
   "fidelity": 5.0,
   "usefulness": 4.0
  },
  "C_conventional_rag": {
   "clarity": 6.33,
   "argument_logic": 5.67,
   "contribution_positioning": 5.33,
   "methodological_rigor": 5.0,
   "fidelity": 4.67,
   "usefulness": 5.67
  },
  "D_profiles_retrieval": {
   "clarity": 5.33,
   "argument_logic": 5.33,
   "contribution_positioning": 4.0,
   "methodological_rigor": 4.33,
   "fidelity": 5.33,
   "usefulness": 4.0
  },
  "E_full_engine": {
   "clarity": 5.33,
   "argument_logic": 5.33,
   "contribution_positioning": 3.67,
   "methodological_rigor": 5.0,
   "fidelity": 4.67,
   "usefulness": 4.0
  }
 },
 "judge_consistency": {
  "n_pairs": 50,
  "clarity": {
   "mean_abs_diff": 0.18,
   "spearman": 0.712
  },
  "argument_logic": {
   "mean_abs_diff": 0.38,
   "spearman": 0.599
  },
  "contribution_positioning": {
   "mean_abs_diff": 0.3,
   "spearman": 0.909
  },
  "methodological_rigor": {
   "mean_abs_diff": 0.28,
   "spearman": 0.893
  },
  "fidelity": {
   "mean_abs_diff": 0.44,
   "spearman": 0.739
  },
  "usefulness": {
   "mean_abs_diff": 0.24,
   "spearman": 0.818
  }
 },
 "secondary_means": {
  "A_baseline": {
   "preserved_rate": 0.199,
   "new_numbers": 1.583,
   "n_unverifiable_citations": 0.0,
   "n_unsupported_claims": 1.167,
   "max_overlap_share": 0.041,
   "todo_placeholders": 12.292,
   "n_words": 769.333,
   "cost_usd": 0.076,
   "latency_s": 33.675,
   "n": 24
  },
  "B_strong_prompt": {
   "preserved_rate": 0.276,
   "new_numbers": 1.208,
   "n_unverifiable_citations": 0.0,
   "n_unsupported_claims": 1.792,
   "max_overlap_share": 0.043,
   "todo_placeholders": 15.458,
   "n_words": 828.417,
   "cost_usd": 0.081,
   "latency_s": 36.117,
   "n": 24
  },
  "C_conventional_rag": {
   "preserved_rate": 0.244,
   "new_numbers": 1.458,
   "n_unverifiable_citations": 0.0,
   "n_unsupported_claims": 1.5,
   "max_overlap_share": 0.033,
   "todo_placeholders": 12.625,
   "n_words": 844.625,
   "cost_usd": 0.087,
   "latency_s": 37.613,
   "n": 24
  },
  "D_profiles_retrieval": {
   "preserved_rate": 0.238,
   "new_numbers": 1.833,
   "n_unverifiable_citations": 0.0,
   "n_unsupported_claims": 1.458,
   "max_overlap_share": 0.032,
   "todo_placeholders": 12.458,
   "n_words": 847.5,
   "cost_usd": 0.095,
   "latency_s": 38.204,
   "n": 24
  },
  "E_full_engine": {
   "preserved_rate": 0.302,
   "new_numbers": 0.708,
   "n_unverifiable_citations": 0.0,
   "n_unsupported_claims": 0.875,
   "max_overlap_share": 0.048,
   "todo_placeholders": 12.25,
   "n_words": 981.25,
   "cost_usd": 0.202,
   "latency_s": 80.062,
   "n": 24
  },
  "E_no_factcheck": {
   "preserved_rate": 0.4,
   "new_numbers": 1.938,
   "n_unverifiable_citations": 0.0,
   "n_unsupported_claims": 1.0,
   "max_overlap_share": 0.025,
   "todo_placeholders": 12.875,
   "n_words": 909.062,
   "cost_usd": 0.197,
   "latency_s": 77.606,
   "n": 16
  },
  "E_no_profile": {
   "preserved_rate": 0.364,
   "new_numbers": 0.625,
   "n_unverifiable_citations": 0.0,
   "n_unsupported_claims": 1.125,
   "max_overlap_share": 0.03,
   "todo_placeholders": 13.062,
   "n_words": 931.5,
   "cost_usd": 0.194,
   "latency_s": 77.194,
   "n": 16
  },
  "E_no_referee": {
   "preserved_rate": 0.221,
   "new_numbers": 1.438,
   "n_unverifiable_citations": 0.0,
   "n_unsupported_claims": 0.938,
   "max_overlap_share": 0.041,
   "todo_placeholders": 12.125,
   "n_words": 790.812,
   "cost_usd": 0.097,
   "latency_s": 39.5,
   "n": 16
  },
  "E_no_retrieval": {
   "preserved_rate": 0.305,
   "new_numbers": 0.938,
   "n_unverifiable_citations": 0.0,
   "n_unsupported_claims": 0.75,
   "max_overlap_share": 0.047,
   "todo_placeholders": 14.312,
   "n_words": 897.938,
   "cost_usd": 0.195,
   "latency_s": 78.019,
   "n": 16
  }
 },
 "H4": {
  "E_vs_A": {
   "n_unverifiable_citations": {
    "n": 24,
    "mean_diff": 0.0,
    "ci95": [
     0.0,
     0.0
    ],
    "wilcoxon_p": 1.0,
    "cliffs_delta": 0.0
   },
   "n_unsupported_claims": {
    "n": 24,
    "mean_diff": -0.292,
    "ci95": [
     -0.875,
     0.469
    ],
    "wilcoxon_p": 0.22484337530538745,
    "cliffs_delta": -0.109
   },
   "new_numbers": {
    "n": 24,
    "mean_diff": -0.875,
    "ci95": [
     -1.938,
     -0.031
    ],
    "wilcoxon_p": 0.040684908780744745,
    "cliffs_delta": -0.113
   }
  },
  "E_vs_B": {
   "n_unverifiable_citations": {
    "n": 24,
    "mean_diff": 0.0,
    "ci95": [
     0.0,
     0.0
    ],
    "wilcoxon_p": 1.0,
    "cliffs_delta": 0.0
   },
   "n_unsupported_claims": {
    "n": 24,
    "mean_diff": -0.917,
    "ci95": [
     -1.688,
     0.125
    ],
    "wilcoxon_p": 0.05044606501119172,
    "cliffs_delta": -0.312
   },
   "new_numbers": {
    "n": 24,
    "mean_diff": -0.5,
    "ci95": [
     -1.125,
     0.0
    ],
    "wilcoxon_p": 0.08446903225325182,
    "cliffs_delta": -0.083
   }
  },
  "E_vs_C": {
   "n_unverifiable_citations": {
    "n": 24,
    "mean_diff": 0.0,
    "ci95": [
     0.0,
     0.0
    ],
    "wilcoxon_p": 1.0,
    "cliffs_delta": 0.0
   },
   "n_unsupported_claims": {
    "n": 24,
    "mean_diff": -0.625,
    "ci95": [
     -1.281,
     0.219
    ],
    "wilcoxon_p": 0.07156571124393864,
    "cliffs_delta": -0.24
   },
   "new_numbers": {
    "n": 24,
    "mean_diff": -0.75,
    "ci95": [
     -1.344,
     -0.25
    ],
    "wilcoxon_p": 0.026014479622894222,
    "cliffs_delta": -0.125
   }
  }
 },
 "H5_overlap": {
  "preregistered_metric": {
   "max_overlap_share_overall": 0.36,
   "share_outputs_below_0.02": 0.5652173913043478
  },
  "excluding_source_document": {
   "max_other_share": 0.0063,
   "share_outputs_below_0.02": 1.0,
   "mean_other_share": 0.0005831521739130436
  },
  "self_overlap_by_task": {
   "draft_introduction": 0.018,
   "explain_specification": 0.012,
   "flag_causal_claims": 0.038,
   "improve_abstract": 0.117,
   "improve_table": 0.035,
   "interpret_results": 0.012,
   "referee_report": 0.006,
   "rewrite_literature": 0.063
  },
  "self_overlap_by_condition": {
   "A_baseline": 0.04,
   "B_strong_prompt": 0.043,
   "C_conventional_rag": 0.033,
   "D_profiles_retrieval": 0.032,
   "E_full_engine": 0.048,
   "E_no_factcheck": 0.025,
   "E_no_profile": 0.03,
   "E_no_referee": 0.041,
   "E_no_retrieval": 0.047
  }
 }
}