Research

Paper

TESTING March 24, 2026

ProGRank: Probe-Gradient Reranking to Defend Dense-Retriever RAG from Corpus Poisoning

Authors

Xiangyu Yin, Yi Qi, Chih-hong Cheng

Abstract

Retrieval-Augmented Generation (RAG) improves the reliability of large language model applications by grounding generation in retrieved evidence, but it also introduces a new attack surface: corpus poisoning. In this setting, an adversary injects or edits passages so that they are ranked into the Top-$K$ results for target queries and then affect downstream generation. Existing defences against corpus poisoning often rely on content filtering, auxiliary models, or generator-side reasoning, which can make deployment more difficult. We propose ProGRank, a post hoc, training-free retriever-side defence for dense-retriever RAG. ProGRank stress-tests each query--passage pair under mild randomized perturbations and extracts probe gradients from a small fixed parameter subset of the retriever. From these signals, it derives two instability signals, representational consistency and dispersion risk, and combines them with a score gate in a reranking step. ProGRank preserves the original passage content, requires no retraining, and also supports a surrogate-based variant when the deployed retriever is unavailable. Extensive experiments across three datasets, three dense retriever backbones, representative corpus poisoning attacks, and both retrieval-stage and end-to-end settings show that ProGRank provides stronger defence performance and a favorable robustness--utility trade-off. It also remains competitive under adaptive evasive attacks.

Metadata

arXiv ID: 2603.22934
Provider: ARXIV
Primary Category: cs.AI
Published: 2026-03-24
Fetched: 2026-03-25 06:02

Related papers

Raw Data (Debug)
{
  "raw_xml": "<entry>\n    <id>http://arxiv.org/abs/2603.22934v1</id>\n    <title>ProGRank: Probe-Gradient Reranking to Defend Dense-Retriever RAG from Corpus Poisoning</title>\n    <updated>2026-03-24T08:29:15Z</updated>\n    <link href='https://arxiv.org/abs/2603.22934v1' rel='alternate' type='text/html'/>\n    <link href='https://arxiv.org/pdf/2603.22934v1' rel='related' title='pdf' type='application/pdf'/>\n    <summary>Retrieval-Augmented Generation (RAG) improves the reliability of large language model applications by grounding generation in retrieved evidence, but it also introduces a new attack surface: corpus poisoning. In this setting, an adversary injects or edits passages so that they are ranked into the Top-$K$ results for target queries and then affect downstream generation. Existing defences against corpus poisoning often rely on content filtering, auxiliary models, or generator-side reasoning, which can make deployment more difficult. We propose ProGRank, a post hoc, training-free retriever-side defence for dense-retriever RAG. ProGRank stress-tests each query--passage pair under mild randomized perturbations and extracts probe gradients from a small fixed parameter subset of the retriever. From these signals, it derives two instability signals, representational consistency and dispersion risk, and combines them with a score gate in a reranking step. ProGRank preserves the original passage content, requires no retraining, and also supports a surrogate-based variant when the deployed retriever is unavailable. Extensive experiments across three datasets, three dense retriever backbones, representative corpus poisoning attacks, and both retrieval-stage and end-to-end settings show that ProGRank provides stronger defence performance and a favorable robustness--utility trade-off. It also remains competitive under adaptive evasive attacks.</summary>\n    <category scheme='http://arxiv.org/schemas/atom' term='cs.AI'/>\n    <published>2026-03-24T08:29:15Z</published>\n    <arxiv:primary_category term='cs.AI'/>\n    <author>\n      <name>Xiangyu Yin</name>\n    </author>\n    <author>\n      <name>Yi Qi</name>\n    </author>\n    <author>\n      <name>Chih-hong Cheng</name>\n    </author>\n  </entry>"
}