S 2 G - RAG : Structured Sufficiency and Gap Judging for Iterative Retrieval-Augmented QA
Minghan Li · Junjie Zou · Xinxuan Lv · Chao Zhang · Guodong Zhou
摘要
Retrieval-Augmented Generation (RAG) grounds language models in external evidence, but multi-hop question answering remains difficult because iterative pipelines must control what to retrieve next and when the available evidence is adequate. In practice, systems may answer from incomplete evidenc...