论文研究
arxiv.org ·
大语言模型何时应当弃权？面向选择性风险控制的自我提问链
论文提出 Chain-of-Self-Questioning（CoSQ），一种仅依赖提示的框架，使模型在明确评估回答问题所需信息后才决定是否作答。在 TruthfulQA 817 题多选题的平衡选项协议下，Grounded-CoSQ 在 τ=0.90 时将平均无条件错误承诺率从 CoT 的 13.1% 降至 8.9%（相对下降 32.1%），回答准确率从 86.9% 升至 89.7%，覆盖 87.6% 的问题；该改进在 11 个开源与托管模型及所有评估阈值上均成立。Critical-CoSQ 与 Adaptive-CoSQ 提供覆盖率 88.6% 和 86.5% 的相邻工作点，可靠性仍优于基线。Natural Questions 短答案评估提供了一致的开放式证据。
AI推荐理由 arXiv 新论文，提出仅靠提示的 Chain-of-Self-Questioning（CoSQ）框架，让 LLM 在信息不足时选择弃权。在 TruthfulQA 817 题多选题上，Grounded-CoSQ 在 τ=0.90 时把平均无条件错误承诺率从 CoT 的 13.1% 降到 8.9%（相对降幅 32.1%），同时回答准确率从 86.9% 升到 89.7%，覆盖 87.6% 的问题，并在 11 个模型和所有阈值上成立；另有 Natural Questions 短答案的开放式验证。局限是仅摘要、单一作者、未提供代码或复现细节，且弃权机制对下游应用的实际收益仍需验证。
72
正文
Computer Science > Computation and Language
arXiv:2609.17516 (cs)
[Submitted on 15 Sep 2026]
Title:When Should LLMs Abstain? Chain-of-Self-Questioning for Selective Risk Control
Authors:Ali Şenol
View a PDF of the paper titled When Should LLMs Abstain? Chain-of-Self-Questioning for Selective Risk Control, by Ali \c{S}enol
View PDF
HTML (experimental)
Abstract: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.
Subjects:
Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
Cite as:
arXiv:2609.17516 [cs.CL]
(or
arXiv:2609.17516v1 [cs.CL] for this version)
https://doi.org/10.48550/arXiv.2609.17516
Focus to learn more
arXiv-issued DOI via DataCite (pending registration)
Submission history
From: Ali Şenol [view email]
[v1]
Tue, 15 Sep 2026 17:52:24 UTC (834 KB)
Full-text links:
Access Paper:
View a PDF of the paper titled When Should LLMs Abstain? Chain-of-Self-Questioning for Selective Risk Control, by Ali \c{S}enol
View PDF
HTML (experimental)
TeX Source
view license
Current browse context:
cs.CL
new
|
recent
| 2026-09
Change to browse by:
cs
cs.AI
References & Citations
NASA ADS
Google Scholar
Semantic Scholar
export BibTeX citation
Loading...
BibTeX formatted citation
×
loading...
Data provided by:
Bookmark
Bibliographic Tools
Bibliographic and Citation Tools
Bibliographic Explorer Toggle
Bibliographic Explorer (What is the Explorer?)
Connected Papers Toggle
Connected Papers (What is Connected Papers?)
Litmaps Toggle
Litmaps (What is Litmaps?)
scite.ai Toggle
scite Smart Citations (What are Smart Citations?)
Code, Data, Media
Code, Data and Media Associated with this Article
alphaXiv Toggle
alphaXiv (What is alphaXiv?)
Links to Code Toggle
CatalyzeX Code Finder for Papers (What is CatalyzeX?)
DagsHub Toggle
DagsHub (What is DagsHub?)
GotitPub Toggle
Gotit.pub (What is GotitPub?)
Huggingface Toggle
Hugging Face (What is Huggingface?)
ScienceCast Toggle
ScienceCast (What is ScienceCast?)
Demos
Demos
Replicate Toggle
Replicate (What is Replicate?)
Spaces Toggle
Hugging Face Spaces (What is Spaces?)
Spaces Toggle
TXYZ.AI (What is TXYZ.AI?)
Related Papers
Recommenders and Search Tools
Link to Influence Flower
Influence Flower (What are Influence Flowers?)
Core recommender toggle
CORE Recommender (What is CORE?)
Author
Venue
Institution
Topic
About arXivLabs
arXivLabs: experimental projects with community collaborators
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Which authors of this paper are endorsers? |
Disable MathJax (What is MathJax?)