Metraipolis › archive › dc_0e2744bd0acc

SCoNE: Selective Context-aware Neuron Editing for Robust Retrieval-Augmented Generation — arXiv:2609.00689

Source
arXiv (en)
Original
https://arxiv.org/abs/2609.00689
License (as archived)
CC0-1.0
Archived
2026-10-17T06:41:19.365Z
SHA-256 of the archived text
7e71660c9512b7b43d628251c06739bd522fa52f7a226f52981d9d2d05a87725
Stable URL of this page
https://metraipolis.com/documento/dc_0e2744bd0acc

How to cite

SCoNE: Selective Context-aware Neuron Editing for Robust Retrieval-Augmented Generation — arXiv:2609.00689 — arXiv (en), https://arxiv.org/abs/2609.00689 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_0e2744bd0acc (Metraipolis archive dc_0e2744bd0acc, sha256 7e71660c9512b7b43d628251c06739bd522fa52f7a226f52981d9d2d05a87725)
BibTeX
@misc{dc0e2744bd0acc,
  title = {SCoNE: Selective Context-aware Neuron Editing for Robust Retrieval-Augmented Generation — arXiv:2609.00689},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.00689},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_0e2744bd0acc; sha256 7e71660c9512b7b43d628251c06739bd522fa52f7a226f52981d9d2d05a87725},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.00689. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Retrieval-Augmented Generation (RAG) is highly sensitive to retrieval noise: when retrieved documents mix informative and irrelevant context, LLMs are easily distracted, leading to hallucinations. To overcome this, we propose SCoNE (Selective Context-aware Neuron Editing), a training-free model editing approach that improves retrieval noise robustness by selectively strengthening context-aware FFN neurons that are identified by both high attribution and high cross-input variability. SCoNE requires only a small number of mining samples, no fine-tuning, and no inference-time overhead. Across various knowledge-intensive question-answering benchmarks and two LLM backbones, SCoNE consistently outperforms competitive baseline methods. Our code is available at this https URL .

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-02 in cs.CL).
Categorie arXiv: Computation and Language (cs.CL).
Pagina del lavoro: https://arxiv.org/abs/2609.00689 — 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 2026-10-17T06:33:38.821Z dalla pagina https://arxiv.org/catchup/cs.CL/2026-09-02?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).