FairytaleQA

该数据集名为FairytaleQA，专为教育目的构建，基于故事书内容。每本书被划分为若干章节，标注者据此创作了叙事性质的问题和答案对。问题根据七种问题类型进行标注，以捕捉叙事元素和关系，并分为显性或隐性。该数据集涵盖了232本书，共8,548个问答对，其任务是生成问题（Question Generation, Qg）。

This dataset, named FairytaleQA, was constructed for educational purposes and is grounded in storybook content. Each book is divided into multiple chapters, based on which annotators create narrative-style question-answer pairs. The questions are annotated with seven question types to capture narrative elements and their relationships, and are categorized into explicit or implicit types. This dataset covers 232 books with a total of 8,548 question-answer pairs, and its target task is Question Generation (Qg).