RAGSentinel: Certifiable Geometric Consensus for Robust Retrieval-Augmented Generation

August 25, 2026 ยท Grace Period ยท ๐Ÿ› EMNLP 2026

โณ Grace Period
This paper is less than 90 days old. We give authors time to release their code before passing judgment.
Authors Yueyang Quan, Anjun Gao, Yufei Xia, Minghong Fang, Zhuqing Liu arXiv ID 2608.23965 Category cs.CR: Cryptography & Security Cross-listed cs.AI, cs.IR, cs.LG Citations 0 Venue EMNLP 2026
Abstract
Retrieval-augmented generation (RAG) improves the factuality of large language models by grounding responses in external documents, but it also exposes a critical security vulnerability: adversarial documents injected into the knowledge database can enter the context window and steer the model toward targeted incorrect answers. Existing post-retrieval defenses rely on instruction following, parametric knowledge, or text-level consistency, all of which can be imitated or optimized against by adaptive attackers. We propose RAGSentinel, a training-free, label-free defense for black-box RAG systems. RAGSentinel uses a surrogate encoder to measure query-conditioned hidden-state shifts induced by retrieved documents, removes shared topic directions, and filters poisoned documents as geometric outliers from a robust majority consensus. We prove that, under an honest-majority assumption and a representation-level separation condition, RAGSentinel exactly recovers a poison-free majority-sized context. Experiments across three question-answering datasets, three LLM families, and multiple poisoning attacks show that RAGSentinel consistently achieves low attack success rates while preserving competitive accuracy and remaining effective against adaptive attacks with full pipeline knowledge.
Community shame:
Not yet rated
Community Contributions

Found the code? Know the venue? Think something is wrong? Let us know!

๐Ÿ“œ Similar Papers

In the same crypt โ€” Cryptography & Security