Datasets in CoT prompting as pedagogical scaffolding in AI-mediated translation tasks: effects on discourse-level performance across genres and directions

Original Chinese prompts, six source texts, six reference translations, and 24 translation outputs generated under minimal and CoT prompting conditions in this study.本研究涵盖原始中文提示词、6个源文本、6个参考译文，以及在最小提示与思维链（Chain of Thought，CoT）提示条件下生成的24条翻译输出。