@inproceedings{liu-etal-2025-metafaith,
title = "{M}eta{F}aith: Faithful Natural Language Uncertainty Expression in {LLM}s",
author = "Liu, Gabrielle Kaili-May and
Yona, Gal and
Caciularu, Avi and
Szpektor, Idan and
Rudner, Tim G. J. and
Cohan, Arman",
editor = "Christodoulopoulos, Christos and
Chakraborty, Tanmoy and
Rose, Carolyn and
Peng, Violet",
booktitle = "Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing",
month = nov,
year = "2025",
address = "Suzhou, China",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2025.emnlp-main.1505/",
pages = "29600--29644",
ISBN = "979-8-89176-332-6",
abstract = "A critical component in the trustworthiness of LLMs is reliable uncertainty communication, yet LLMs often use assertive language when conveying false claims, leading to over-reliance and eroded trust. We present the first systematic study of {\_}faithful confidence calibration{\_} of LLMs, benchmarking models' ability to use linguistic expressions of uncertainty that {\_}faithfully reflect{\_} their intrinsic uncertainty, across a comprehensive array of models, datasets, and prompting strategies. Our results demonstrate that LLMs largely fail at this task, and that existing interventions are insufficient: standard prompt approaches provide only marginal gains, and existing, factuality-based calibration techniques can even harm faithful calibration. To address this critical gap, we introduce MetaFaith, a novel prompt-based calibration approach inspired by human metacognition. We show that MetaFaith robustly improves faithful calibration across diverse models and task domains, enabling up to 61{\%} improvement in faithfulness and achieving an 83{\%} win rate over original generations as judged by humans."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="liu-etal-2025-metafaith">
<titleInfo>
<title>MetaFaith: Faithful Natural Language Uncertainty Expression in LLMs</title>
</titleInfo>
<name type="personal">
<namePart type="given">Gabrielle</namePart>
<namePart type="given">Kaili-May</namePart>
<namePart type="family">Liu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Gal</namePart>
<namePart type="family">Yona</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Avi</namePart>
<namePart type="family">Caciularu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Idan</namePart>
<namePart type="family">Szpektor</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Tim</namePart>
<namePart type="given">G</namePart>
<namePart type="given">J</namePart>
<namePart type="family">Rudner</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Arman</namePart>
<namePart type="family">Cohan</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2025-11</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing</title>
</titleInfo>
<name type="personal">
<namePart type="given">Christos</namePart>
<namePart type="family">Christodoulopoulos</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Tanmoy</namePart>
<namePart type="family">Chakraborty</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Carolyn</namePart>
<namePart type="family">Rose</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Violet</namePart>
<namePart type="family">Peng</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Suzhou, China</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
<identifier type="isbn">979-8-89176-332-6</identifier>
</relatedItem>
<abstract>A critical component in the trustworthiness of LLMs is reliable uncertainty communication, yet LLMs often use assertive language when conveying false claims, leading to over-reliance and eroded trust. We present the first systematic study of _faithful confidence calibration_ of LLMs, benchmarking models’ ability to use linguistic expressions of uncertainty that _faithfully reflect_ their intrinsic uncertainty, across a comprehensive array of models, datasets, and prompting strategies. Our results demonstrate that LLMs largely fail at this task, and that existing interventions are insufficient: standard prompt approaches provide only marginal gains, and existing, factuality-based calibration techniques can even harm faithful calibration. To address this critical gap, we introduce MetaFaith, a novel prompt-based calibration approach inspired by human metacognition. We show that MetaFaith robustly improves faithful calibration across diverse models and task domains, enabling up to 61% improvement in faithfulness and achieving an 83% win rate over original generations as judged by humans.</abstract>
<identifier type="citekey">liu-etal-2025-metafaith</identifier>
<location>
<url>https://aclanthology.org/2025.emnlp-main.1505/</url>
</location>
<part>
<date>2025-11</date>
<extent unit="page">
<start>29600</start>
<end>29644</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T MetaFaith: Faithful Natural Language Uncertainty Expression in LLMs
%A Liu, Gabrielle Kaili-May
%A Yona, Gal
%A Caciularu, Avi
%A Szpektor, Idan
%A Rudner, Tim G. J.
%A Cohan, Arman
%Y Christodoulopoulos, Christos
%Y Chakraborty, Tanmoy
%Y Rose, Carolyn
%Y Peng, Violet
%S Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing
%D 2025
%8 November
%I Association for Computational Linguistics
%C Suzhou, China
%@ 979-8-89176-332-6
%F liu-etal-2025-metafaith
%X A critical component in the trustworthiness of LLMs is reliable uncertainty communication, yet LLMs often use assertive language when conveying false claims, leading to over-reliance and eroded trust. We present the first systematic study of _faithful confidence calibration_ of LLMs, benchmarking models’ ability to use linguistic expressions of uncertainty that _faithfully reflect_ their intrinsic uncertainty, across a comprehensive array of models, datasets, and prompting strategies. Our results demonstrate that LLMs largely fail at this task, and that existing interventions are insufficient: standard prompt approaches provide only marginal gains, and existing, factuality-based calibration techniques can even harm faithful calibration. To address this critical gap, we introduce MetaFaith, a novel prompt-based calibration approach inspired by human metacognition. We show that MetaFaith robustly improves faithful calibration across diverse models and task domains, enabling up to 61% improvement in faithfulness and achieving an 83% win rate over original generations as judged by humans.
%U https://aclanthology.org/2025.emnlp-main.1505/
%P 29600-29644
Markdown (Informal)
[MetaFaith: Faithful Natural Language Uncertainty Expression in LLMs](https://aclanthology.org/2025.emnlp-main.1505/) (Liu et al., EMNLP 2025)
ACL
- Gabrielle Kaili-May Liu, Gal Yona, Avi Caciularu, Idan Szpektor, Tim G. J. Rudner, and Arman Cohan. 2025. MetaFaith: Faithful Natural Language Uncertainty Expression in LLMs. In Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing, pages 29600–29644, Suzhou, China. Association for Computational Linguistics.