MCQ Dataset
收藏arXiv2025-09-30 收录
下载链接:
https://github.com/DRSY/DGen
下载链接
链接失效反馈官方服务:
资源简介:
该数据集是由多种开源多项选择题数据集汇编而成,包括SciQ、MCQL、AI2科学问题等,同时还加入了从网站上搜集的额外常识和词汇多项选择题。该数据集重点关注提取式填空式的干扰项生成。此外,该数据集被随机划分为训练集、验证集和测试集,比例分别为8:1:1。数据集规模为2,880条目,其任务是针对多项选择题进行干扰项生成。
This dataset is compiled from multiple open-source multiple-choice question datasets, including SciQ, MCQL, AI2 Science Questions, among others. Additionally, it incorporates extra common-sense and vocabulary multiple-choice questions collected from websites. This dataset focuses on extractive fill-in-the-blank distractor generation. Furthermore, it is randomly split into training, validation and test sets with a ratio of 8:1:1 respectively. The dataset contains a total of 2,880 entries, and its task is to generate distractors for multiple-choice questions.
提供机构:
Various open-source sources



