arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2606.11265cs.CRcs.AI
收藏

当投毒在检索后失败:重新审视分块与重排序管道下的语料库投毒

When Poison Fails After Retrieval: Revisiting Corpus Poisoning under Chunking and Reranking Pipelines

School of Computer Science, Shandong University(山东大学计算机学院);School of Information, Shandong University(山东大学信息学院);School of Software Engineering, Shandong University(山东大学软件学院)

Xi Nie, Hongwei Li, Shenghao Wu, Mingxuan Li, Jiachen Li, Wenbo Jiang

AI总结:

针对RAG系统,提出CRCP框架,通过联合优化检索相关性、重排序一致性和分块边界鲁棒性,解决现有投毒方法在真实多阶段检索管道中因分块和重排序导致效果下降的问题。

摘要:

检索增强生成(RAG)系统容易受到语料库投毒攻击,这些攻击通过恶意知识注入操纵下游模型输出。现有研究主要在简化的检索设置下评估投毒,忽视了涉及文档分块、密集检索、重排序和基于生成的生成等实际RAG管道。在本文中,我们重新审视了在现实多阶段检索管道下的语料库投毒,并表明许多现有攻击在重排序后效果显著下降,尽管在检索阶段实现了高相关性。我们识别出检索粒度不匹配是这种失败的关键原因:文档级别的对抗信号在分块过程中经常被碎片化,而重排序器偏好局部连贯且包含答案的段落,而非全局优化的语义相似性。基于这一观察,我们提出了分块感知和重排序一致的投毒(CRCP),这是一个联合优化检索相关性、重排序一致性和分块边界鲁棒性的投毒框架。CRCP在优化过程中显式建模分块变换,以生成在变化的分块配置下仍然有效的局部自包含对抗段落。在多个检索器和重排序器的标准RAG基准上的实验表明,现有投毒方法对分块大小和重排序策略高度敏感,而CRCP在现实检索管道中实现了显著更高的攻击成功率和更强的鲁棒性。我们的发现凸显了当前RAG安全评估中的一个重要现实差距,并表明现代RAG系统中的投毒应被视为一个多阶段检索一致性问题,而不仅仅是检索问题。

英文摘要:

Retrieval-Augmented Generation (RAG) systems are vulnerable to corpus poisoning attacks that manipulate downstream model outputs through malicious knowledge injection. Existing studies mainly evaluate poisoning under simplified retrieval settings, overlooking practical RAG pipelines involving document chunking, dense retrieval, reranking, and grounded generation. In this paper, we revisit corpus poisoning under realistic multi-stage retrieval pipelines and show that many existing attacks substantially degrade after reranking despite achieving high retrieval-stage relevance. We identify retrieval granularity mismatch as a key reason for this failure: document-level adversarial signals are often fragmented during chunking, while rerankers favor locally coherent and answer-bearing passages rather than globally optimized semantic similarity. Based on this observation, we propose Chunk-aware and Rerank-Consistent Poisoning (CRCP), a poisoning framework that jointly optimizes retrieval relevance, reranker consistency, and chunk-boundary robustness. CRCP explicitly models chunking transformations during optimization to generate locally self-contained adversarial passages that remain effective under varying chunking configurations. Experiments on standard RAG benchmarks with multiple retrievers and rerankers show that existing poisoning methods are highly sensitive to chunk size and reranking strategies, whereas CRCP achieves substantially higher attack success rates and stronger robustness across realistic retrieval pipelines. Our findings highlight an important realism gap in current RAG security evaluation and suggest that poisoning in modern RAG systems should be studied as a multi-stage retrieval consistency problem rather than a retrieval-only problem.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑