{
  "status": "PRELIMINARY",
  "release_date": "2026-10-22",
  "dataset": "clinc150",
  "source_rows": 23700,
  "excluded_out_of_scope": 23250,
  "rows_in_scope_before_cleaning": 450,
  "rows_with_redaction_before_deduplication": 67,
  "retained_rows": 447,
  "retained_rows_with_redaction": 67,
  "excluded_normalized_duplicates": 3,
  "excluded_cross_source_duplicates": 3,
  "redacted_span_counts_before_deduplication": {
    "AMOUNT": 2,
    "ORGANIZATION": 11,
    "DATE": 42,
    "LOCATION": 11,
    "TIME": 1,
    "PERSON": 1
  },
  "splits": {
    "test": 89,
    "train": 298,
    "validation": 60
  },
  "intent_count": 3,
  "checks": {
    "count_reconciliation": true,
    "no_normalized_duplicate_text": true,
    "no_cross_split_normalized_duplicates": true,
    "identifier_pattern_scan_passed": true
  },
  "privacy_method": "Regular-expression identifier/contact/address screening and spaCy en_core_web_sm 3.8.0 named-entity redaction. Automated screening is not a guarantee of anonymity.",
  "deduplication": "NFKC, casefold, retain alphanumeric characters only. Identical normalized texts are deduplicated across all three released customer-service datasets, with evaluation preferred over training and BANKING77 then CLINC150 then Bitext as source tie-breaker. Conflicting labels for identical normalized text within a source are excluded.",
  "source_files": [
    {
      "file": "clinc-full.json",
      "url": "https://raw.githubusercontent.com/clinc/oos-eval/master/data/data_full.json",
      "bytes": 2495390,
      "sha256": "36923c3705a59e08fe9c3883d8bc2dd966ef93e22cb78ac41171782a698d56e0"
    },
    {
      "file": "clinc-domains.json",
      "url": "https://raw.githubusercontent.com/clinc/oos-eval/master/data/domains.json",
      "bytes": 3818,
      "sha256": "b947b579d3b8e74b06f93b01083d8efaff2888b43a3e362533bd88a6e1211b3a"
    },
    {
      "file": "clinc-license.txt",
      "url": "https://raw.githubusercontent.com/clinc/oos-eval/master/LICENSE",
      "bytes": 19467,
      "sha256": "e6bc9e9c474700b708f568bac9e5a8a9bcb2b1dad53442f5ba449fcb848b8e76"
    },
    {
      "file": "clinc-readme.md",
      "url": "https://raw.githubusercontent.com/clinc/oos-eval/master/README.md",
      "bytes": 2668,
      "sha256": "8a8df26c4de3d25b6c4cff385ca23602c30e4e56828b015206cbc202f6db363a"
    }
  ]
}