Metraipolis › archive › dc_78eb59999e55
RAGSieve: Self-Referenced Local Contrast for Knowledge-Poison Detection in Retrieval-Augmented Generation — arXiv:2608.13010
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2608.13010
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-19T04:24:34.250Z
- SHA-256 of the archived text
2191b87df4b54d25fc4c0f0a35e256165cdf80e8dfe50870f491cb63f8a39d3e- Stable URL of this page
https://metraipolis.com/documento/dc_78eb59999e55
How to cite
RAGSieve: Self-Referenced Local Contrast for Knowledge-Poison Detection in Retrieval-Augmented Generation — arXiv:2608.13010 — arXiv (en), https://arxiv.org/abs/2608.13010 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_78eb59999e55 (Metraipolis archive dc_78eb59999e55, sha256 2191b87df4b54d25fc4c0f0a35e256165cdf80e8dfe50870f491cb63f8a39d3e)
BibTeX
@misc{dc78eb59999e55,
title = {RAGSieve: Self-Referenced Local Contrast for Knowledge-Poison Detection in Retrieval-Augmented Generation — arXiv:2608.13010},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2608.13010},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_78eb59999e55; sha256 2191b87df4b54d25fc4c0f0a35e256165cdf80e8dfe50870f491cb63f8a39d3e},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2608.13010. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Retrieval-augmented generation uses an external corpus as inference-time evidence, allowing an attacker to promote a false answer by injecting a handful of documents. Detection must distinguish this manipulation from ordinary relevance without knowing which queries or documents are targeted. Existing detectors use text irregularity, candidate consensus, or corpus-level graph structure, whose reliability varies with the attack and local context. We present RAGSieve, which constructs a reference matched to each detection scope. At query time, RAGSieve-Query (RSQ) compares generation candidates with the lower-ranked tail of the same retrieval, exposing answer-token concentration and carrier-payload seams. At corpus time, RAGSieve-Graph (RSG) compares each document's strongest semantic relations with its own neighborhood floor to measure coordinated density. Neither requires poison labels, a trusted corpus, or training. Across three QA datasets, three dense retrievers, and six poisoning constructions, RSQ reaches 95.2% AUROC and detects 82.2% of poison at a 5% clean-removal budget, against 81.1% and 52.5% for the strongest query-time baseline; RSG reaches 93.3% and 79.8% against 79.4% and 37.6% for the strongest corpus-time baseline, with a 79.6% versus 1.4% detection rate on camouflaged injections. Joint deployment cuts attack success from 67.4% to 16.1% while retaining unpoisoned-retrieval F1 at 41.0%, compared with 42.1% without filtering. Source code is available at this https URL . Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-08-14 in cs.CL). Categorie arXiv: Computation and Language (cs.CL) ; Cryptography and Security (cs.CR); Information Retrieval (cs.IR). Pagina del lavoro: https://arxiv.org/abs/2608.13010 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito). Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina. Letto il 2026-10-19T04:17:42.391Z dalla pagina https://arxiv.org/catchup/cs.CL/2026-08-14?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).