{
  "schema": "silktrain-pilot-sample/v1",
  "synthetic": true,
  "locale": "zh-hant",
  "workflow": "multilingual-evaluation",
  "service": "training-evaluation",
  "title": "多語言模型評測",
  "source_scenario": "華語訂單中的「十四件」被系統記為「四十件」。",
  "expected_delivery": "針對使用地區的提問、修正建議、錯誤嚴重程度與術語依據。",
  "acceptance_check": "由第二位熟悉當地語言的人員複核保留樣本的語意、文字及用詞。",
  "scoping_checklist": "準備一小組回答、參考來源、評分規準，以及需由第二人複核的範例。",
  "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"
    }
  ]
}
