{
  "date": "2026-10-05",
  "candidate": "aw1-answerability-entity-ab-20261005-r1",
  "scope": "Synthetic bilingual QA answerability scorer; exposed development evidence-group transfer; frozen reader/ranking. Candidate only.",
  "source_receipt_sha256": "dff08f355829f4c5496a3875461e74b3d307457cf07967173fdf9497ad795008",
  "QUALITY": "WITHHELD",
  "promotion": false,
  "execution": "PASS_SCOPED_COLD",
  "dev_gate": "FAIL_OR_NOT_RUN",
  "final": "NOT_RUN_DEV_FAIL",
  "canary": "NOT_RUN",
  "answerability_updates": 1200,
  "LM_updates": 0,
  "reader_updates": 0,
  "start_updates": 0,
  "end_updates": 0,
  "metrics": {
    "C": {
      "en": {
        "answerable": 40,
        "raw_exact": 27,
        "gated_exact": 13,
        "coverage": 20,
        "noanswer": 20,
        "noanswer_correct": 10,
        "correct_rejected": 14,
        "wrong_accepted": 7,
        "noanswer_falseanswer": 10,
        "perfect_answerability_exact_ceiling_TEST_ONLY": 27
      },
      "vi": {
        "answerable": 40,
        "raw_exact": 36,
        "gated_exact": 36,
        "coverage": 40,
        "noanswer": 20,
        "noanswer_correct": 20,
        "correct_rejected": 0,
        "wrong_accepted": 4,
        "noanswer_falseanswer": 0,
        "perfect_answerability_exact_ceiling_TEST_ONLY": 36
      }
    },
    "R": {
      "en": {
        "answerable": 40,
        "raw_exact": 27,
        "gated_exact": 13,
        "coverage": 20,
        "noanswer": 20,
        "noanswer_correct": 10,
        "correct_rejected": 14,
        "wrong_accepted": 7,
        "noanswer_falseanswer": 10,
        "perfect_answerability_exact_ceiling_TEST_ONLY": 27
      },
      "vi": {
        "answerable": 40,
        "raw_exact": 36,
        "gated_exact": 36,
        "coverage": 40,
        "noanswer": 20,
        "noanswer_correct": 20,
        "correct_rejected": 0,
        "wrong_accepted": 4,
        "noanswer_falseanswer": 0,
        "perfect_answerability_exact_ceiling_TEST_ONLY": 36
      }
    }
  },
  "pair_metrics": {
    "C": {
      "question_both": 13,
      "position_both": 19,
      "question_raw_ceiling_TEST_ONLY": 19,
      "position_raw_ceiling_TEST_ONLY": 19
    },
    "R": {
      "question_both": 13,
      "position_both": 19,
      "question_raw_ceiling_TEST_ONLY": 19,
      "position_raw_ceiling_TEST_ONLY": 19
    }
  },
  "train_fitting": {
    "C": {
      "en": {
        "numerator": 8000,
        "denominator": 8000,
        "scope": "train fitting; not held-out or raw-answer accuracy"
      },
      "vi": {
        "numerator": 8000,
        "denominator": 8000,
        "scope": "train fitting; not held-out or raw-answer accuracy"
      }
    },
    "R": {
      "en": {
        "numerator": 9600,
        "denominator": 9600,
        "scope": "train fitting; not held-out or raw-answer accuracy"
      },
      "vi": {
        "numerator": 9600,
        "denominator": 9600,
        "scope": "train fitting; not held-out or raw-answer accuracy"
      }
    }
  },
  "production": "No promotion; production not evaluated by this candidate receipt",
  "first_failed_transition": "Already-perfect TRAIN answerability fitting → EN evidence-group answerability transfer; reassignment did not improve frozen operating-point decisions",
  "historical_residual_receipt_sha256": "89d0b2c97bdc8efa79d0b984fc401e4b4cd62b1941ef1b55f616842de86d93ca",
  "comparison_limit": "Residual and entity candidates use different exposed DEV packets; VI 35/40 versus 36/40 is not a paired improvement claim."
}