ClarQ

该数据集是一个多样化的、大规模的澄清问题集合，基于从Stack Exchange提取的帖子-评论对。它包含了大约200万个示例，分布在173个领域内。此外，该数据集还包括每个帖子的实际答案，这使得研究人员能够在独立评估和问答情境中，对澄清问题的质量进行评价。规模上，该数据集大约包含了2,079,300个澄清问题元组，其任务定位于澄清问题的生成。This dataset is a diverse, large-scale collection of clarification questions, constructed from post-comment pairs extracted from Stack Exchange. It contains approximately 2 million examples across 173 domains. Additionally, the dataset includes the ground-truth answers corresponding to each post, allowing researchers to assess the quality of clarification questions in both standalone evaluation and question-answering scenarios. In terms of scale, the dataset comprises roughly 2,079,300 clarification question tuples, with its task targeting clarification question generation.