1 paper
Zhanghao Hu, Yijun Yang, Junjie Xu +2
Current approaches to question answering rely on pre-trained language models (PLMs) like RoBERTa. This work challenges the existing question-answer encoding convention and explores…