PAQ

该数据集名为PAQ，包含6500万个自动生成的问题-答案对，旨在提升开放领域问答模型的性能。这些数据集来源于维基百科，并根据答案提取器和问题生成器的不同版本进行了分类。规模上，该数据集包含了6500万个问答对，其任务专注于问答领域。

This dataset, named PAQ, contains 65 million automatically generated question-answer pairs, aiming to enhance the performance of open-domain question answering models. Derived from Wikipedia, it is categorized based on different versions of answer extractors and question generators. In terms of scale, this dataset has 65 million question-answer pairs, focusing on the question answering domain.