{
  "schema": "silktrain-pilot-sample/v1",
  "synthetic": true,
  "locale": "en",
  "workflow": "multilingual-evaluation",
  "service": "training-evaluation",
  "title": "Multilingual model evaluation",
  "source_scenario": "A Mandarin order for fourteen items is recorded as forty.",
  "expected_delivery": "Locale-specific prompts, corrections, severity ratings and references for each disputed term.",
  "acceptance_check": "Have a second speaker review meaning, writing system and regional terminology on a held-out sample.",
  "scoping_checklist": "Prepare a small answer set, the source material, a scoring rubric and examples that should go to a second reviewer.",
  "examples": [
    {
      "id": "L-014",
      "locale": "zh-Hans",
      "utterance": "订十四件",
      "expected_quantity": 14,
      "observed_quantity": 40,
      "label": "number_error"
    },
    {
      "id": "L-015",
      "locale": "de-DE",
      "utterance": "vierzehn Stück",
      "expected_quantity": 14,
      "observed_quantity": 14,
      "label": "pass"
    },
    {
      "id": "L-016",
      "locale": "sr-Latn",
      "utterance": "četrdeset komada",
      "expected_quantity": 40,
      "observed_quantity": 14,
      "label": "number_error"
    }
  ]
}