Metraipolis › archive › dc_3e80447787f5

When Should LLMs Abstain? Chain-of-Self-Questioning for Selective Risk Control — arXiv:2609.17516

Source
arXiv (en)
Original
https://arxiv.org/abs/2609.17516
License (as archived)
CC0-1.0
Archived
2026-10-15T18:39:01.596Z
SHA-256 of the archived text
00d5b5ca09f4d83c5561a893fbb472c05eb8c74a4632ea12e2b82389eb7f4965
Stable URL of this page
https://metraipolis.com/documento/dc_3e80447787f5

How to cite

When Should LLMs Abstain? Chain-of-Self-Questioning for Selective Risk Control — arXiv:2609.17516 — arXiv (en), https://arxiv.org/abs/2609.17516 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_3e80447787f5 (Metraipolis archive dc_3e80447787f5, sha256 00d5b5ca09f4d83c5561a893fbb472c05eb8c74a4632ea12e2b82389eb7f4965)
BibTeX
@misc{dc3e80447787f5,
  title = {When Should LLMs Abstain? Chain-of-Self-Questioning for Selective Risk Control — arXiv:2609.17516},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.17516},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_3e80447787f5; sha256 00d5b5ca09f4d83c5561a893fbb472c05eb8c74a4632ea12e2b82389eb7f4965},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.17516. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Large language models can produce fluent answers when their factual support is weak. This paper introduces Chain-of-Self-Questioning (CoSQ), a prompt-only framework that makes answer commitment conditional on an explicit assessment of the information required to answer a question. We evaluate three CoSQ variants under seventeen conditions on the 817-item TruthfulQA multiple-choice validation set using eleven open-weight and hosted model families. In the final balanced-option protocol, Grounded-CoSQ at {\tau}=0.90 reduces the mean unconditional wrong-commitment rate from 13.1% under chain-of-thought prompting to 8.9%, a 32.1% relative reduction, while increasing answered accuracy from 86.9% to 89.7% and answering 87.6% of questions. Both improvements hold for all eleven models and at every evaluated threshold. Critical-CoSQ and Adaptive-CoSQ provide neighboring operating points with 88.6% and 86.5% coverage, respectively, while remaining more reliable than the baseline. A secondary Natural Questions Short-Answer evaluation provides convergent open-form evidence. These findings show that self-assessment can support explicit, tunable answer-or-abstain decisions when an unsupported commitment is more costly than referral or review.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-16 in cs.AI).
Categorie arXiv: Computation and Language (cs.CL) ; Artificial Intelligence (cs.AI).
Pagina del lavoro: https://arxiv.org/abs/2609.17516 — 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 2026-10-15T18:03:09.768Z dalla pagina https://arxiv.org/catchup/cs.AI/2026-09-16?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).