TellMeWhy

TellMeWhy是一个由石溪大学创建的大型数据集，包含超过30,000个关于短篇故事中角色行为的‘为什么’问题及其自由形式的答案。该数据集特别关注那些答案不在故事文本中的问题，旨在测试模型是否能利用外部常识知识来回答问题。数据集的创建过程涉及使用简单的模板转换从故事中生成问题，并通过众包方式收集答案。TellMeWhy的应用领域主要集中在测试和提升自然语言处理系统在理解和推理叙事文本方面的能力，特别是在需要常识推理的情境中。

TellMeWhy is a large-scale dataset created by Stony Brook University, comprising over 30,000 'why' questions regarding the behaviors of characters in short stories along with their free-form answers. This dataset specifically focuses on questions whose answers are not present within the original story text, and aims to evaluate whether models can leverage external commonsense knowledge to answer such questions. The construction of this dataset involves generating questions from stories using simple template-based conversion, and collecting answers via crowdsourcing. Its primary application scenarios focus on testing and enhancing the capabilities of natural language processing (NLP) systems to understand and reason over narrative texts, particularly in scenarios requiring commonsense reasoning.