Joint-GCG: Unified Gradient-Based Poisoning Attacks on Retrieval-Augmented Generation Systems

TL;DR

This paper introduces Joint-GCG, a novel framework for gradient-based poisoning attacks on Retrieval-Augmented Generation systems, unifying retrieval and generation attack strategies to improve effectiveness and transferability.

Contribution

Joint-GCG is the first unified gradient-based attack framework targeting both retrieval and generation components in RAG systems, enhancing attack success and transferability.

Findings

Achieves up to 25% higher attack success rate.

Demonstrates unprecedented transferability to unseen models.

Effectively unifies attack strategies across retrieval and generation stages.

Abstract

Retrieval-Augmented Generation (RAG) systems enhance Large Language Models (LLMs) by retrieving relevant documents from external corpora before generating responses. This approach significantly expands LLM capabilities by leveraging vast, up-to-date external knowledge. However, this reliance on external knowledge makes RAG systems vulnerable to corpus poisoning attacks that manipulate generated outputs via poisoned document injection. Existing poisoning attack strategies typically treat the retrieval and generation stages as disjointed, limiting their effectiveness. We propose Joint-GCG, the first framework to unify gradient-based attacks across both retriever and generator models through three innovations: (1) Cross-Vocabulary Projection for aligning embedding spaces, (2) Gradient Tokenization Alignment for synchronizing token-level gradient signals, and (3) Adaptive Weighted Fusion for dynamically balancing attacking objectives. Evaluations demonstrate that Joint-GCG achieves at most 25% and an average of 5% higher attack success rate than previous methods across multiple retrievers and generators. While optimized under a white-box assumption, the generated poisons show unprecedented transferability to unseen models. Joint-GCG's innovative unification of gradient-based attacks across retrieval and generation stages fundamentally reshapes our understanding of vulnerabilities within RAG systems. Our code is available at https://github.com/NicerWang/Joint-GCG.