@inproceedings{lewis-lim-etal-2026-think,
title = "Think When Unsure: Leveraging Model Confidence to Decide When to Use Chain-of-Thought",
author = "Lewis-Lim, Samuel and
Tan, Xingwei and
Zhao, Zhixue and
Aletras, Nikolaos",
editor = "Zhang, Yue and
Xu, Ruifeng and
Purwarianti, Ayu and
Rong, Tong and
Suzuki, Jun",
booktitle = "Findings of the {A}ssociation for {C}omputational {L}inguistics: {IJCNLP}-{AACL} 2026",
month = nov,
year = "2026",
address = "Hengqin, China",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2026.findings-aacl.21/",
pages = "369--396",
ISBN = "979-8-89176-444-6",
abstract = "Chain-of-thought (CoT) prompting improves performance on complex reasoning tasks, but often triggers unnecessary computation on simpler queries. A natural approach is to use the model{'}s own confidence as a training-free signal to gate this decision, answering directly when confident and invoking CoT only when uncertain. Across four training-free confidence signals, three model families, and seven reasoning benchmarks, we find that the effectiveness of confidence-gated CoT varies, and it is unclear how to predict whether it will work for a new model. We address this with two key contributions. First, we identify that gating fails in the following circumstances: signal saturation, where confidence scores collapse to a narrow range and lose discriminative power; and selection bias, when confidence-gated routing sends the longest queries to CoT. Second, we propose a novel Pareto-optimal threshold selection method. This has the dual ability to preserve accuracy while reducing costs where the confidence signal is strong, and maintains accuracy where it is weak. Where the confidence signal is strong, gating reduces CoT usage by 30{\%} without sacrificing accuracy. These findings provide practical guidance for evaluating and deploying models that selectively use CoT."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="lewis-lim-etal-2026-think">
<titleInfo>
<title>Think When Unsure: Leveraging Model Confidence to Decide When to Use Chain-of-Thought</title>
</titleInfo>
<name type="personal">
<namePart type="given">Samuel</namePart>
<namePart type="family">Lewis-Lim</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Xingwei</namePart>
<namePart type="family">Tan</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Zhixue</namePart>
<namePart type="family">Zhao</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Nikolaos</namePart>
<namePart type="family">Aletras</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2026-11</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Findings of the Association for Computational Linguistics: IJCNLP-AACL 2026</title>
</titleInfo>
<name type="personal">
<namePart type="given">Yue</namePart>
<namePart type="family">Zhang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Ruifeng</namePart>
<namePart type="family">Xu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Ayu</namePart>
<namePart type="family">Purwarianti</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Tong</namePart>
<namePart type="family">Rong</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jun</namePart>
<namePart type="family">Suzuki</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Hengqin, China</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
<identifier type="isbn">979-8-89176-444-6</identifier>
</relatedItem>
<abstract>Chain-of-thought (CoT) prompting improves performance on complex reasoning tasks, but often triggers unnecessary computation on simpler queries. A natural approach is to use the model’s own confidence as a training-free signal to gate this decision, answering directly when confident and invoking CoT only when uncertain. Across four training-free confidence signals, three model families, and seven reasoning benchmarks, we find that the effectiveness of confidence-gated CoT varies, and it is unclear how to predict whether it will work for a new model. We address this with two key contributions. First, we identify that gating fails in the following circumstances: signal saturation, where confidence scores collapse to a narrow range and lose discriminative power; and selection bias, when confidence-gated routing sends the longest queries to CoT. Second, we propose a novel Pareto-optimal threshold selection method. This has the dual ability to preserve accuracy while reducing costs where the confidence signal is strong, and maintains accuracy where it is weak. Where the confidence signal is strong, gating reduces CoT usage by 30% without sacrificing accuracy. These findings provide practical guidance for evaluating and deploying models that selectively use CoT.</abstract>
<identifier type="citekey">lewis-lim-etal-2026-think</identifier>
<location>
<url>https://aclanthology.org/2026.findings-aacl.21/</url>
</location>
<part>
<date>2026-11</date>
<extent unit="page">
<start>369</start>
<end>396</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Think When Unsure: Leveraging Model Confidence to Decide When to Use Chain-of-Thought
%A Lewis-Lim, Samuel
%A Tan, Xingwei
%A Zhao, Zhixue
%A Aletras, Nikolaos
%Y Zhang, Yue
%Y Xu, Ruifeng
%Y Purwarianti, Ayu
%Y Rong, Tong
%Y Suzuki, Jun
%S Findings of the Association for Computational Linguistics: IJCNLP-AACL 2026
%D 2026
%8 November
%I Association for Computational Linguistics
%C Hengqin, China
%@ 979-8-89176-444-6
%F lewis-lim-etal-2026-think
%X Chain-of-thought (CoT) prompting improves performance on complex reasoning tasks, but often triggers unnecessary computation on simpler queries. A natural approach is to use the model’s own confidence as a training-free signal to gate this decision, answering directly when confident and invoking CoT only when uncertain. Across four training-free confidence signals, three model families, and seven reasoning benchmarks, we find that the effectiveness of confidence-gated CoT varies, and it is unclear how to predict whether it will work for a new model. We address this with two key contributions. First, we identify that gating fails in the following circumstances: signal saturation, where confidence scores collapse to a narrow range and lose discriminative power; and selection bias, when confidence-gated routing sends the longest queries to CoT. Second, we propose a novel Pareto-optimal threshold selection method. This has the dual ability to preserve accuracy while reducing costs where the confidence signal is strong, and maintains accuracy where it is weak. Where the confidence signal is strong, gating reduces CoT usage by 30% without sacrificing accuracy. These findings provide practical guidance for evaluating and deploying models that selectively use CoT.
%U https://aclanthology.org/2026.findings-aacl.21/
%P 369-396
Markdown (Informal)
[Think When Unsure: Leveraging Model Confidence to Decide When to Use Chain-of-Thought](https://aclanthology.org/2026.findings-aacl.21/) (Lewis-Lim et al., Findings 2026)
ACL