作者位次 #4/10ACL 2026Findings· Findings of ACL
PRA - RAG : Provably Robust Aggregation in Retrieval-Augmented Generation against Retrieval Corruption
Xue Tan · Yi Zheng · Chang Huo · Yunruo Zhang · Yu Liu · Hao Luan · Zhuyang Yu · Jun Dai · Xiaoyan Sun · Ping Chen
摘要
Retrieval-Augmented Generation (RAG) enhances Large Language Models (LLMs) by incorporating external knowledge, effectively mitigating their inherent knowledge limitations. However, RAG remains vulnerable to poisoning attacks that manipulate retrieved texts to mislead model outputs. Existing defe...