Official dataset link for ''Reasoning in Dialog: Improving Response Generation by Context Reading Comprehension''
DRCD corpus consists of 113,116 training, 3,000 validation, and 3,000 test cases.
We appreciate your citation if you find our dataset and code beneficial.
@inproceedings{chen2021reasoning,
title={Reasoning in Dialog: Improving Response Generation by Context Reading Comprehension},
author={Xiuying Chen, Zhi Cui, Jiayi Zhang, Chen Wei, Jianwei Cui, Bin Wang, Dongyan Zhao, and Rui Yan},
booktitle={AAAI},
year={2021}
}