{
  "format": "carrier-steward-fresh-test/1",
  "first_test_read_utc": "2026-10-09T10:04:14.928034+00:00",
  "freeze": "protocol/pretest-freeze.json",
  "test_examples": 1200,
  "test_sha256": "9b583dcedfbb21b597e01ce180be18eceec68307c4793910279449122739f4e7",
  "test_matches_generation_manifest": true,
  "evaluation_retry": "First full pass completed predictions but failed JSON serialization of NumPy int64 before reporting metrics. Same frozen weights and numeric evaluation rerun after serialization-only fix. See analysis/evaluation-failure.txt.",
  "methods": {
    "copy": {
      "examples": 1200,
      "micro": {
        "tp": 28329,
        "fp": 3019,
        "fn": 3150,
        "precision": 0.9036940155671813,
        "recall": 0.8999332888592395,
        "f1": 0.9018097314848712
      },
      "added": {
        "tp": 0,
        "fp": 0,
        "fn": 3150,
        "precision": null,
        "recall": 0.0,
        "f1": 0.0
      },
      "removed": {
        "tp": 0,
        "fp": 0,
        "fn": 3019,
        "precision": null,
        "recall": 0.0,
        "f1": 0.0
      },
      "persistent": {
        "correct": 28329,
        "total": 28329,
        "recall": 1.0
      },
      "exact_set_accuracy": 0.2658333333333333
    },
    "prior135": {
      "examples": 1200,
      "micro": {
        "tp": 17451,
        "fp": 14192,
        "fn": 14028,
        "precision": 0.5514963815061783,
        "recall": 0.5543695797198132,
        "f1": 0.5529292481226831
      },
      "added": {
        "tp": 1705,
        "fp": 13890,
        "fn": 1445,
        "precision": 0.10932991343379288,
        "recall": 0.5412698412698412,
        "f1": 0.18191517738063484
      },
      "removed": {
        "tp": 2717,
        "fp": 12583,
        "fn": 302,
        "precision": 0.17758169934640522,
        "recall": 0.8999668764491553,
        "f1": 0.29663191222228286
      },
      "persistent": {
        "correct": 15746,
        "total": 28329,
        "recall": 0.555826185181263
      },
      "exact_set_accuracy": 0.1875
    },
    "copy_edges_predicted_count": {
      "examples": 1200,
      "micro": {
        "tp": 28329,
        "fp": 3019,
        "fn": 3150,
        "precision": 0.9036940155671813,
        "recall": 0.8999332888592395,
        "f1": 0.9018097314848712
      },
      "added": {
        "tp": 0,
        "fp": 0,
        "fn": 3150,
        "precision": null,
        "recall": 0.0,
        "f1": 0.0
      },
      "removed": {
        "tp": 0,
        "fp": 0,
        "fn": 3019,
        "precision": null,
        "recall": 0.0,
        "f1": 0.0
      },
      "persistent": {
        "correct": 28329,
        "total": 28329,
        "recall": 1.0
      },
      "exact_set_accuracy": 0.2658333333333333
    },
    "initial_editor": {
      "examples": 1200,
      "micro": {
        "tp": 6834,
        "fp": 25986,
        "fn": 24645,
        "precision": 0.20822669104204752,
        "recall": 0.21709711236062137,
        "f1": 0.2125694023235198
      },
      "added": {
        "tp": 2595,
        "fp": 25963,
        "fn": 555,
        "precision": 0.09086770782267665,
        "recall": 0.8238095238095238,
        "f1": 0.163681089945755
      },
      "removed": {
        "tp": 2996,
        "fp": 24090,
        "fn": 23,
        "precision": 0.1106106475670088,
        "recall": 0.9923815833057303,
        "f1": 0.1990367048663013
      },
      "persistent": {
        "correct": 4239,
        "total": 28329,
        "recall": 0.14963465000529494
      },
      "exact_set_accuracy": 0.13833333333333334
    },
    "seed136001": {
      "examples": 1200,
      "micro": {
        "tp": 31296,
        "fp": 370,
        "fn": 183,
        "precision": 0.988315543485126,
        "recall": 0.99418660059087,
        "f1": 0.9912423786523081
      },
      "added": {
        "tp": 3090,
        "fp": 370,
        "fn": 60,
        "precision": 0.8930635838150289,
        "recall": 0.9809523809523809,
        "f1": 0.9349470499243571
      },
      "removed": {
        "tp": 3019,
        "fp": 123,
        "fn": 0,
        "precision": 0.960852959898154,
        "recall": 1.0,
        "f1": 0.9800357084888817
      },
      "persistent": {
        "correct": 28206,
        "total": 28329,
        "recall": 0.9956581594832151
      },
      "exact_set_accuracy": 0.7666666666666667
    },
    "seed136002": {
      "examples": 1200,
      "micro": {
        "tp": 31321,
        "fp": 377,
        "fn": 158,
        "precision": 0.9881065051422803,
        "recall": 0.994980780838019,
        "f1": 0.9915317283188502
      },
      "added": {
        "tp": 3104,
        "fp": 377,
        "fn": 46,
        "precision": 0.8916977879919563,
        "recall": 0.9853968253968254,
        "f1": 0.9362087166339919
      },
      "removed": {
        "tp": 3019,
        "fp": 112,
        "fn": 0,
        "precision": 0.9642286809326094,
        "recall": 1.0,
        "f1": 0.9817886178861789
      },
      "persistent": {
        "correct": 28217,
        "total": 28329,
        "recall": 0.996046454163578
      },
      "exact_set_accuracy": 0.7733333333333333
    },
    "seed136003": {
      "examples": 1200,
      "micro": {
        "tp": 31331,
        "fp": 396,
        "fn": 148,
        "precision": 0.987518517351152,
        "recall": 0.9952984529368786,
        "f1": 0.991393222162453
      },
      "added": {
        "tp": 3105,
        "fp": 396,
        "fn": 45,
        "precision": 0.8868894601542416,
        "recall": 0.9857142857142858,
        "f1": 0.9336941813261164
      },
      "removed": {
        "tp": 3019,
        "fp": 103,
        "fn": 0,
        "precision": 0.9670083279948751,
        "recall": 1.0,
        "f1": 0.9832274873799055
      },
      "persistent": {
        "correct": 28226,
        "total": 28329,
        "recall": 0.9963641498111476
      },
      "exact_set_accuracy": 0.7741666666666667
    },
    "raw_seed136001": {
      "examples": 1200,
      "micro": {
        "tp": 31300,
        "fp": 394,
        "fn": 179,
        "precision": 0.9875686249763362,
        "recall": 0.9943136694304139,
        "f1": 0.9909296693207541
      },
      "added": {
        "tp": 3094,
        "fp": 394,
        "fn": 56,
        "precision": 0.8870412844036697,
        "recall": 0.9822222222222222,
        "f1": 0.932208496535101
      },
      "removed": {
        "tp": 3019,
        "fp": 123,
        "fn": 0,
        "precision": 0.960852959898154,
        "recall": 1.0,
        "f1": 0.9800357084888817
      },
      "persistent": {
        "correct": 28206,
        "total": 28329,
        "recall": 0.9956581594832151
      },
      "exact_set_accuracy": 0.7558333333333334
    },
    "raw_seed136002": {
      "examples": 1200,
      "micro": {
        "tp": 31325,
        "fp": 415,
        "fn": 154,
        "precision": 0.9869250157529931,
        "recall": 0.9951078496775628,
        "f1": 0.9909995412771477
      },
      "added": {
        "tp": 3108,
        "fp": 415,
        "fn": 42,
        "precision": 0.882202668180528,
        "recall": 0.9866666666666667,
        "f1": 0.9315150606923422
      },
      "removed": {
        "tp": 3019,
        "fp": 112,
        "fn": 0,
        "precision": 0.9642286809326094,
        "recall": 1.0,
        "f1": 0.9817886178861789
      },
      "persistent": {
        "correct": 28217,
        "total": 28329,
        "recall": 0.996046454163578
      },
      "exact_set_accuracy": 0.7541666666666667
    },
    "raw_seed136003": {
      "examples": 1200,
      "micro": {
        "tp": 31335,
        "fp": 436,
        "fn": 144,
        "precision": 0.986276793302068,
        "recall": 0.9954255217764224,
        "f1": 0.9908300395256917
      },
      "added": {
        "tp": 3109,
        "fp": 436,
        "fn": 41,
        "precision": 0.8770098730606488,
        "recall": 0.986984126984127,
        "f1": 0.9287528005974608
      },
      "removed": {
        "tp": 3019,
        "fp": 103,
        "fn": 0,
        "precision": 0.9670083279948751,
        "recall": 1.0,
        "f1": 0.9832274873799055
      },
      "persistent": {
        "correct": 28226,
        "total": 28329,
        "recall": 0.9963641498111476
      },
      "exact_set_accuracy": 0.7533333333333333
    },
    "seed136001_without_additions": {
      "examples": 1200,
      "micro": {
        "tp": 31126,
        "fp": 123,
        "fn": 353,
        "precision": 0.9960638740439695,
        "recall": 0.9887861749102577,
        "f1": 0.9924116821833949
      },
      "added": {
        "tp": 2920,
        "fp": 123,
        "fn": 230,
        "precision": 0.9595793624712455,
        "recall": 0.926984126984127,
        "f1": 0.9430001614726304
      },
      "removed": {
        "tp": 3019,
        "fp": 123,
        "fn": 0,
        "precision": 0.960852959898154,
        "recall": 1.0,
        "f1": 0.9800357084888817
      },
      "persistent": {
        "correct": 28206,
        "total": 28329,
        "recall": 0.9956581594832151
      },
      "exact_set_accuracy": 0.7341666666666666
    }
  },
  "comparisons": {
    "seed136001_minus_copy": {
      "difference": 0.08943264716743693,
      "paired_scene_bootstrap_95": [
        0.08604033727214601,
        0.09294270293754407
      ],
      "resamples": 2000,
      "seed": 136010,
      "scope": "descriptive procedural-scene resampling; not general semantic population inference"
    },
    "seed136001_minus_prior135": {
      "difference": 0.43831313052962506,
      "paired_scene_bootstrap_95": [
        0.43039492874343566,
        0.4463554609355329
      ],
      "resamples": 2000,
      "seed": 136010,
      "scope": "descriptive procedural-scene resampling; not general semantic population inference"
    },
    "seed136001_minus_copy_edges_predicted_count": {
      "difference": 0.08943264716743693,
      "paired_scene_bootstrap_95": [
        0.08604033727214601,
        0.09294270293754407
      ],
      "resamples": 2000,
      "seed": 136010,
      "scope": "descriptive procedural-scene resampling; not general semantic population inference"
    },
    "seed136002_minus_copy": {
      "difference": 0.08972199683397897,
      "paired_scene_bootstrap_95": [
        0.08628812098244716,
        0.09330025383524992
      ],
      "resamples": 2000,
      "seed": 136010,
      "scope": "descriptive procedural-scene resampling; not general semantic population inference"
    },
    "seed136002_minus_prior135": {
      "difference": 0.4386024801961671,
      "paired_scene_bootstrap_95": [
        0.4307201455125669,
        0.44654233269638866
      ],
      "resamples": 2000,
      "seed": 136010,
      "scope": "descriptive procedural-scene resampling; not general semantic population inference"
    },
    "seed136002_minus_copy_edges_predicted_count": {
      "difference": 0.08972199683397897,
      "paired_scene_bootstrap_95": [
        0.08628812098244716,
        0.09330025383524992
      ],
      "resamples": 2000,
      "seed": 136010,
      "scope": "descriptive procedural-scene resampling; not general semantic population inference"
    },
    "seed136003_minus_copy": {
      "difference": 0.08958349067758176,
      "paired_scene_bootstrap_95": [
        0.08621989601958865,
        0.09308530663461743
      ],
      "resamples": 2000,
      "seed": 136010,
      "scope": "descriptive procedural-scene resampling; not general semantic population inference"
    },
    "seed136003_minus_prior135": {
      "difference": 0.4384639740397699,
      "paired_scene_bootstrap_95": [
        0.43052393183569154,
        0.44643473189326743
      ],
      "resamples": 2000,
      "seed": 136010,
      "scope": "descriptive procedural-scene resampling; not general semantic population inference"
    },
    "seed136003_minus_copy_edges_predicted_count": {
      "difference": 0.08958349067758176,
      "paired_scene_bootstrap_95": [
        0.08621989601958865,
        0.09308530663461743
      ],
      "resamples": 2000,
      "seed": 136010,
      "scope": "descriptive procedural-scene resampling; not general semantic population inference"
    }
  },
  "editor_adapter_events": {
    "136001": {
      "duplicate_collapse": 87,
      "endpoint_filter": 26,
      "capacity_filter": 1
    },
    "136002": {
      "duplicate_collapse": 88,
      "endpoint_filter": 41,
      "capacity_filter": 1
    },
    "136003": {
      "duplicate_collapse": 88,
      "endpoint_filter": 42,
      "capacity_filter": 1
    }
  },
  "backbone_node_count_accuracy": 0.9291666666666667,
  "elapsed_seconds": 13.04102611541748,
  "birth_links": {
    "raw_counts": [
      174,
      271,
      56
    ],
    "adapted_counts": [
      170,
      247,
      60
    ],
    "raw_f1": 0.5155555555555555,
    "adapted_f1": 0.5255023183925811
  },
  "prediction_archive": {
    "path": "analysis/fresh-test-predictions.jsonl",
    "sha256": "5dee8913d40f9839af27da105ad4b5e0e78a3615ec26c6627a2629d9b53ef586"
  },
  "limits": [
    "Authored numeric teacher reused; this is fresh examples, not novel meanings.",
    "Geometric distance and signed membrane distance are engineered features.",
    "No direct REMOVE or retype2..5 target examples; only r0/r1 transitions and birth-return links trained.",
    "Existinglinkidentity and restricted addition support are designed priors.",
    "Other channels unchanged from frozen135; no general improvement claim.",
    "Raw editor may propose absent endpoints. Filtering uses predicted count and is explicitly not learned validity.",
    "No conscience, qualia, semantic comprehension or Carrier-specific advantage established."
  ]
}