Critical Questions Dataset

该数据集包含了通过在论证方案模板的背景下应用大型语言模型而获得的分类为有帮助、无帮助和无效的生成关键问题。具体来说，这个数据集利用了GPT-4o-mini和LLaMa模型，并结合了多种提示技巧来优化有用关键问题的生成。数据集的规模包括60个有帮助的问题、25个无帮助的问题以及17个无效的问题，其任务旨在从论证性文本中生成关键问题。

This dataset comprises generated critical questions categorized into three classes: helpful, unhelpful, and invalid. These questions were obtained by applying large language models (LLMs) under the framework of argumentation scheme templates. Specifically, this dataset leverages GPT-4o-mini and LLaMA models, combined with diverse prompting techniques to optimize the generation of helpful critical questions. The dataset includes 60 helpful questions, 25 unhelpful questions, and 17 invalid questions, with its core task targeting the generation of critical questions from argumentative texts.