{
  "schema": "silktrain-pilot-sample/v1",
  "synthetic": true,
  "locale": "zh-hans",
  "workflow": "multilingual-evaluation",
  "service": "training-evaluation",
  "title": "多语言模型评测",
  "source_scenario": "普通话订单中的“十四件”被系统记成了“四十件”。",
  "expected_delivery": "面向目标市场的测试问题、纠正建议、错误严重度和术语依据。",
  "acceptance_check": "由第二位熟悉当地语言的审核员检查预留样本的含义、文字和用语。",
  "scoping_checklist": "准备一小组回答、参考来源、评分准则，以及需要第二人复核的示例。",
  "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"
    }
  ]
}
