@inproceedings{seo-etal-2026-advice,
title = "{ADVICE}: Answer-Dependent Verbalized Confidence Estimation",
author = "Seo, KiJung and
Lim, Sehun and
Kim, Taeuk",
editor = "Liakata, Maria and
Moreira, Viviane P. and
Zhang, Jiajun and
Jurgens, David",
booktitle = "Proceedings of the 64th Annual Meeting of the {A}ssociation for {C}omputational {L}inguistics (Volume 1: Long Papers)",
month = jul,
year = "2026",
address = "San Diego, California, United States",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2026.acl-long.1098/",
doi = "10.18653/v1/2026.acl-long.1098",
pages = "23942--23962",
ISBN = "979-8-89176-390-6",
abstract = "Recent progress in large language models (LLMs) has enabled them to communicate their confidence in natural language, improving transparency and reliability.However, this expressiveness is often accompanied by systematic overconfidence, whose underlying causes remain poorly understood. In this work, we analyze the dynamics of verbalized confidence estimation and identify answer-independence-the failure to condition confidence on the model{'}s own answer-as a primary driver of this behavior.To address this, we introduce ADVICE (Answer-Dependent VerbalIzed Confidence Estimation), a fine-tuning framework that promotes answer-grounded confidence estimation.Extensive experiments show that ADVICE substantially improves confidence calibration, while exhibiting strong generalization to unseen settings without degrading task performance.We further demonstrate that these gains stem from enhanced answer dependence, shedding light on the origins of overconfidence and enabling trustworthy confidence verbalization."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="seo-etal-2026-advice">
<titleInfo>
<title>ADVICE: Answer-Dependent Verbalized Confidence Estimation</title>
</titleInfo>
<name type="personal">
<namePart type="given">KiJung</namePart>
<namePart type="family">Seo</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Sehun</namePart>
<namePart type="family">Lim</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Taeuk</namePart>
<namePart type="family">Kim</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2026-07</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)</title>
</titleInfo>
<name type="personal">
<namePart type="given">Maria</namePart>
<namePart type="family">Liakata</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Viviane</namePart>
<namePart type="given">P</namePart>
<namePart type="family">Moreira</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jiajun</namePart>
<namePart type="family">Zhang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">David</namePart>
<namePart type="family">Jurgens</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">San Diego, California, United States</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
<identifier type="isbn">979-8-89176-390-6</identifier>
</relatedItem>
<abstract>Recent progress in large language models (LLMs) has enabled them to communicate their confidence in natural language, improving transparency and reliability.However, this expressiveness is often accompanied by systematic overconfidence, whose underlying causes remain poorly understood. In this work, we analyze the dynamics of verbalized confidence estimation and identify answer-independence-the failure to condition confidence on the model’s own answer-as a primary driver of this behavior.To address this, we introduce ADVICE (Answer-Dependent VerbalIzed Confidence Estimation), a fine-tuning framework that promotes answer-grounded confidence estimation.Extensive experiments show that ADVICE substantially improves confidence calibration, while exhibiting strong generalization to unseen settings without degrading task performance.We further demonstrate that these gains stem from enhanced answer dependence, shedding light on the origins of overconfidence and enabling trustworthy confidence verbalization.</abstract>
<identifier type="citekey">seo-etal-2026-advice</identifier>
<identifier type="doi">10.18653/v1/2026.acl-long.1098</identifier>
<location>
<url>https://aclanthology.org/2026.acl-long.1098/</url>
</location>
<part>
<date>2026-07</date>
<extent unit="page">
<start>23942</start>
<end>23962</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T ADVICE: Answer-Dependent Verbalized Confidence Estimation
%A Seo, KiJung
%A Lim, Sehun
%A Kim, Taeuk
%Y Liakata, Maria
%Y Moreira, Viviane P.
%Y Zhang, Jiajun
%Y Jurgens, David
%S Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)
%D 2026
%8 July
%I Association for Computational Linguistics
%C San Diego, California, United States
%@ 979-8-89176-390-6
%F seo-etal-2026-advice
%X Recent progress in large language models (LLMs) has enabled them to communicate their confidence in natural language, improving transparency and reliability.However, this expressiveness is often accompanied by systematic overconfidence, whose underlying causes remain poorly understood. In this work, we analyze the dynamics of verbalized confidence estimation and identify answer-independence-the failure to condition confidence on the model’s own answer-as a primary driver of this behavior.To address this, we introduce ADVICE (Answer-Dependent VerbalIzed Confidence Estimation), a fine-tuning framework that promotes answer-grounded confidence estimation.Extensive experiments show that ADVICE substantially improves confidence calibration, while exhibiting strong generalization to unseen settings without degrading task performance.We further demonstrate that these gains stem from enhanced answer dependence, shedding light on the origins of overconfidence and enabling trustworthy confidence verbalization.
%R 10.18653/v1/2026.acl-long.1098
%U https://aclanthology.org/2026.acl-long.1098/
%U https://doi.org/10.18653/v1/2026.acl-long.1098
%P 23942-23962
Markdown (Informal)
[ADVICE: Answer-Dependent Verbalized Confidence Estimation](https://aclanthology.org/2026.acl-long.1098/) (Seo et al., ACL 2026)
ACL
- KiJung Seo, Sehun Lim, and Taeuk Kim. 2026. ADVICE: Answer-Dependent Verbalized Confidence Estimation. In Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 23942–23962, San Diego, California, United States. Association for Computational Linguistics.