Metraipolis › archive › dc_7bc202ceca6a

Language Models Show Conformity Bias in RAG: When the Majority Endorses a Poisoned Claim — arXiv:2610.41012

Source: arXiv (en)
Original: https://arxiv.org/abs/2610.41012
License (as archived): CC0-1.0
Archived: 2026-11-02T04:45:38.656Z
SHA-256 of the archived text: cbab9824561a75b166a460ad16aabd697186508e40ac828c99e8e2d322088b76
Stable URL of this page: https://metraipolis.com/documento/dc_7bc202ceca6a

How to cite

Language Models Show Conformity Bias in RAG: When the Majority Endorses a Poisoned Claim — arXiv:2610.41012 — arXiv (en), https://arxiv.org/abs/2610.41012 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_7bc202ceca6a (Metraipolis archive dc_7bc202ceca6a, sha256 cbab9824561a75b166a460ad16aabd697186508e40ac828c99e8e2d322088b76)

BibTeX

@misc{dc7bc202ceca6a,
  title = {Language Models Show Conformity Bias in RAG: When the Majority Endorses a Poisoned Claim — arXiv:2610.41012},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2610.41012},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_7bc202ceca6a; sha256 cbab9824561a75b166a460ad16aabd697186508e40ac828c99e8e2d322088b76},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.41012. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Retrieval-augmented generation (RAG) can expose language models to misleading evidence in retrieved documents. We investigate whether increasing the number of mutually agreeing poisoned documents affects a model's ability to detect misinformation. We evaluate 200 randomly selected questions from SciFact (random seed 38), comparing contexts with one poisoned and six healthy documents against contexts with six mutually agreeing poisoned documents and one healthy document. For Gemma-3-12b, poison-detection rates decrease from 25% (50/200) in the minority condition to 10% (20/200) in the majority condition. Among the 38 questions with different outcomes across conditions, 34 change from detected to undetected, while only four change in the opposite direction (exact McNemar test, p = 6.04e-7). This pattern is consistent with a conformity-like response to a majority of mutually agreeing poisoned documents. Our results provide initial evidence of conformity bias in RAG-based misinformation detection: Gemma-3-12b is substantially less likely to detect poisoned claims when they dominate the retrieved evidence.

Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.CL, nuovi lavori del giorno).
Categorie arXiv: Computation and Language (cs.CL).
Pagina del lavoro: https://arxiv.org/abs/2610.41012 — 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 2026-11-02T04:42:26.886Z dalla pagina https://arxiv.org/list/cs.CL/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).