作者位次 #4/5ACL 2026Findings· Findings of ACL
Align Documents to Questions: Question-Oriented Document Rewriting for Retrieval-Augmented Generation
Jiaang Li · Zhendong Mao · Quan Wang · Yuning Wan · Yongdong Zhang
摘要
Retrieval-Augmented Generation (RAG) enhances the factuality of Large Language Models (LLMs) by incorporating retrieved documents and/or generated context. However, LLMs often exhibit a stylistic bias when presented with mixed contexts, favoring fluent but hallucinated generated content over fact...