Problem
Research questions and friction points this paper is trying to address.
multi-passage corpus poisoning
similarity-aware defenses
conflict-aware defenses
Divide and Doubt: Diverse Distributed Poisoning for Retrieval-Augmented Generation
🤖 AI Summary
本文提出DnD方法，通过分散多样风格的段落支持目标答案并加入质疑参考答案的段落，以提高多段落语料中毒攻击效果，尤其在对抗集群和冲突感知防御方面。
📝 Abstract
Multi-passage corpus poisoning often repeats one target claim across similar documents, creating correlated lexical and semantic patterns that similarity- and conflict-aware defenses can suppress jointly. We introduce DnD (Divide and Doubt), a targeted attack based on two principles: distributing support for the target answer across stylistically diverse passages, and including a passage that casts doubt on evidence for the reference answer. The first disperses poison-passage representations in embedding space, while the second strengthens target adoption when multiple poisoned passages are retrieved. We evaluate DnD on two open-domain QA datasets across three LLMs and nine RAG configurations, under both black-box and white-box access to the retriever. Across these settings, DnD matches or outperforms prior attacks in most configurations, with its largest gains against clustering- and conflict-aware defenses.
Innovation
Methods, ideas, or system contributions that make the work stand out.
DnD
stylistically diverse passages
conflict-aware defenses
embedding space dispersion
target answer adoption
🔎 Similar Papers
No similar papers found.
