cs.AINov 3, 2026

Reliability Stress Testing of RAG Retrieval under Query-Relevant Corpus Poisoning

Authors: Janet Sung, Kainan Zhou, Gangzhen Qian, Hang Xiao, Lu Yi

Organizations: Google LLC Mountain View, USA · Fortinet, Inc. Sunnyvale, USA

Abstract

External text can affect a retrieval-augmented generation (RAG) system only after the retriever admits it to the model context. We stress this admission step while varying document content, attacker query information, per-target poison budget, and contamination scope. Across 623 Benchmarking Information Retrieval (BEIR) queries from SciFact and NFCorpus, five query-targeted documents per query reduce classical-retriever nDCG@10 by 40.2-61.5% in isolated-target stress. Full-query anchoring explains most Best Matching 25 (BM25) degradation, whereas appended text also shifts latent semantic analysis (LSA) and Hybrid rankings. A pretrained BGE-small validation reaches 100% target exposure for query-only P=5 on both datasets; adding the injection payload changes SciFact nDCG only from 0.269 to 0.271 and lowers NFCorpus exposure slightly to 98.8%, indicating that query anchoring remains the dominant retrieval-side factor for this dense model. Query length materially changes the interpretation of partial-information attacks: on 162 long SciFact queries, two random terms cover 19.9% of eligible terms and still yield 39.5% Hybrid target exposure. Multi-target contamination adds cross-target interference. Overlap-based admission can remove full-query poisons but may sharply reduce clean retrieval utility, and maximal marginal relevance (MMR) leaves exposure unchanged under the tested configuration. The study measures retrieval-layer failures, not downstream instruction following.

Figures & tables

Explore similar work

Aug 17, 2026

Coverage Is Not Containment: A Fundamental Limit of Admission-Time Defenses Against Coordinated Poisoning of Vector Retrieval

Retrieval-augmented generation (RAG) answers a question by retrieving passages from a vector store and trusting them as context, so anyone who can add documents can try to steer the answer. A recent, appealing defense filters poisoning at ingestion, rejecting any document that behaves like a hub. We show it -- and every ingestion-time filter -- is defeated by a coordinated adversary that injects a handful of individually unremarkable documents which together surround one target query and seize its top-k (on BGE-large / BEIR, m=10 documents take 10/10; 9.9/10 on a live HNSW index). The attack is not theoretical. Realized as ordinary fluent text and run end-to-end through a BGE-large + HNSW + Qwen2.5-7B pipeline, it makes the generator emit the attacker's planted claim in 88% of targets, versus 0% without the injection. And no admission-time defense stops it: at ingestion an attack cone is geometrically identical to a legitimate niche upload, so -- measuring this directly -- the strongest trained classifier, given every feature and thousands of examples, separates the two no better than chance, catching 4.2% of attacks at a 1% false-positive rate. We prove this limit for the entire class of ingestion-time statistics (any decision from documents and reference queries alone), and it reproduces -- and worsens -- across two corpora and five encoders. The one signal that separates an attack from legitimate niche ingestion -- a query's demand -- is invisible before retrieval, which is also the escape: a retrieval-time detector that observes demand catches 100% of the attacks at the same 1% false-positive rate. Coverage of the query space by an admission gate is not containment of coordinated poisoning; robust defense must move past the front door, to demand.
Mar 10, 2026

Semantic Chameleon: Corpus-Dependent Poisoning Attacks and Defenses in RAG Systems

Retrieval-Augmented Generation (RAG) systems extend large language models (LLMs) with external knowledge sources but introduce new attack surfaces through the retrieval pipeline. In particular, adversaries can poison retrieval corpora so that malicious documents are preferentially retrieved at inference time, enabling targeted manipulation of model outputs. We study gradient-guided corpus poisoning attacks against modern RAG pipelines and evaluate retrieval-layer defenses that require no modification to the underlying LLM. We implement dual-document poisoning attacks consisting of a sleeper document and a trigger document optimized using Greedy Coordinate Gradient (GCG). In a large-scale evaluation on the Security Stack Exchange corpus (67,941 documents) with 50 attack attempts, gradient-guided poisoning achieves a 38.0 percent co-retrieval rate under pure vector retrieval. We show that a simple architectural modification, hybrid retrieval combining BM25 and vector similarity, substantially mitigates this attack. Across all 50 attacks, hybrid retrieval reduces gradient-guided attack success from 38 percent to 0 percent without modifying the model or retraining the retriever. When attackers jointly optimize payloads for both sparse and dense retrieval signals, hybrid retrieval can be partially circumvented, achieving 20-44 percent success, but still significantly raises attack difficulty relative to vector-only retrieval. Evaluation across five LLM families (GPT-5.3, GPT-4o, Claude Sonnet 4.6, Llama 4, and GPT-4o-mini) shows attack success ranging from 46.7 percent to 93.3 percent. Cross-corpus evaluation on the FEVER Wikipedia dataset (25 attacks) yields 0 percent attack success across all retrieval configurations.
May 7, 2026

Architecture Matters: Comparing RAG Systems under Knowledge Base Poisoning

Retrieval-Augmented Generation (RAG) systems are vulnerable to knowledge base poisoning, yet existing attacks have been evaluated almost exclusively against vanilla retrieve-then-generate pipelines. Architectures designed to handle conflicting retrieved information - multi-agent debate, agentic retrieval, recursive language models - remain untested against adversarially optimized contradictions. We evaluate four RAG architectures (vanilla RAG, agentic RAG, MADAM-RAG, and Recursive Language Models) under controlled single-document (N=1) poisoning on 921 Natural Questions QA pairs, comparing a clean baseline, naive injection, and CorruptRAG-AK - an adversarial attack whose meta-epistemic framing targets credibility assessment. Architecture is a high-impact variable in adversarial robustness: under CorruptRAG-AK, attack success rates range from 81.9% (vanilla) to 24.4% (RLM) - a spread of nearly 58 percentage points across architectures with comparable clean accuracy (~92%). Decomposing this gap, once the poisoned document is retrieved, adversarial framing - not retrieval optimization - drives the majority of CorruptRAG-AK's advantage for three of four architectures, localizing the cross-architecture vulnerability at the content-reasoning stage. Our MADAM-RAG reimplementation shows the highest apparent contradiction detection rate, though our LLM judge over-identifies this behavior (~48.5% precision), so reported rates are upper bounds. Regardless of detection, MADAM-RAG cannot resolve contradictions reliably, producing a 41.4% non-answer rate even on clean inputs - though implementation divergences from the original may contribute. We introduce a seven-category behavioral taxonomy capturing contradiction detection, hedging, and failure modes beyond binary accuracy. Code, data, and analysis notebooks are publicly available.