{
  "n": 300,
  "split": "track_b_applied / skills_unseen_catalogs, complete 300 rows",
  "learned_models_run": false,
  "accuracy": {
    "first_option": 0.08666666666666667,
    "always_none": 0.22,
    "bm25_or_none": 0.33,
    "bm25_name_description": 0.37
  },
  "unmatched_query_count": 32,
  "none_gold_count": 66,
  "methods": "Same tokenization/BM25; each choice description is a document, no training, no translation, fallback none only if all scores zero.",
  "counts": {
    "first_option": {
      "correct": 26,
      "unnecessary_selection_on_none": 59,
      "missed_required_selection": 28,
      "wrong_non_none_selection": 187
    },
    "always_none": {
      "correct": 66,
      "unnecessary_selection_on_none": 0,
      "missed_required_selection": 234,
      "wrong_non_none_selection": 0
    },
    "bm25_or_none": {
      "correct": 99,
      "unnecessary_selection_on_none": 34,
      "missed_required_selection": 49,
      "wrong_non_none_selection": 118
    },
    "bm25_name_description": {
      "correct": 111,
      "unnecessary_selection_on_none": 36,
      "missed_required_selection": 46,
      "wrong_non_none_selection": 107
    }
  },
  "exploratory_ablation": "name+description variant was added after initial outcomes, no parameter or threshold tuning; no significance/equivalence claim"
}