Metraipolis › archive › dc_ef6f37f10ca5
Factual Yet One-Sided: Stance Poisoning and Evidence Balancing in Retrieval-Augmented Generation — arXiv:2610.34294
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2610.34294
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-27T08:11:33.599Z
- SHA-256 of the archived text
605477d77ebd796e64e4b55798a9a229f8fa5947adf7c70c697ba7ea68e5306e- Stable URL of this page
https://metraipolis.com/documento/dc_ef6f37f10ca5
How to cite
Factual Yet One-Sided: Stance Poisoning and Evidence Balancing in Retrieval-Augmented Generation — arXiv:2610.34294 — arXiv (en), https://arxiv.org/abs/2610.34294 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_ef6f37f10ca5 (Metraipolis archive dc_ef6f37f10ca5, sha256 605477d77ebd796e64e4b55798a9a229f8fa5947adf7c70c697ba7ea68e5306e)
BibTeX
@misc{dcef6f37f10ca5,
title = {Factual Yet One-Sided: Stance Poisoning and Evidence Balancing in Retrieval-Augmented Generation — arXiv:2610.34294},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2610.34294},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_ef6f37f10ca5; sha256 605477d77ebd796e64e4b55798a9a229f8fa5947adf7c70c697ba7ea68e5306e},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.34294. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
The stance expressed in responses from retrieval-augmented generation systems is significantly influenced by the stance distribution of the retrieved materials. Existing research on knowledge poisoning has primarily targeted factual errors. In contrast, this study focuses on injection attacks in which the facts are correct but the stance on a controversial issue is one-sided. At a mean topic-level poisoning share of 18.9% within the post-injection ProCon documents, all four generation models from different vendors exhibited stance shifts aligned with the poisoning direction; across four concentration levels, the shifts increased monotonically. This paper proposes a Stance-Balancing Middleware positioned between the retriever and the generator. It consists of stance detection, fixed balancing (2 supporting + 2 opposing + 1 neutral), and a debiasing prompt, without introducing additional large language model calls. Across all eight concentration-direction settings, the complete method reduced the magnitude of stance deviation by an estimated 0.106 to 0.350; the six settings at poisoning shares of 11.9%-18.9% remained significant after Holm adjustment (adjusted p <= 0.009). In the cross-model experiment, all eight model-direction combinations moved along the stance axis in the corrective direction, and seven of eight showed a clearly detected reduction in absolute distance from 0. In the concentration sweep, the post-defense intervals included 0 in all eight settings, while the cross-model pro-poisoning results retained small positive residual point estimates of +0.045 to +0.063. For 20 candidate documents, stance detection took approximately 1.5% of the median generation time on a single GPU and approximately 16% on CPU. Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.CL, nuovi lavori del giorno). Categorie arXiv: Computation and Language (cs.CL) ; Cryptography and Security (cs.CR); Information Retrieval (cs.IR). Pagina del lavoro: https://arxiv.org/abs/2610.34294 — 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 2026-10-27T07:56:00.226Z dalla pagina https://arxiv.org/list/cs.CL/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).