@inproceedings{soudani-etal-2025-uncertainty,
title = "Why Uncertainty Estimation Methods Fall Short in {RAG}: An Axiomatic Analysis",
author = "Soudani, Heydar and
Kanoulas, Evangelos and
Hasibi, Faegheh",
editor = "Che, Wanxiang and
Nabende, Joyce and
Shutova, Ekaterina and
Pilehvar, Mohammad Taher",
booktitle = "Findings of the Association for Computational Linguistics: ACL 2025",
month = jul,
year = "2025",
address = "Vienna, Austria",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2025.findings-acl.852/",
doi = "10.18653/v1/2025.findings-acl.852",
pages = "16596--16616",
ISBN = "979-8-89176-256-5",
abstract = "Large Language Models (LLMs) are valued for their strong performance across various tasks, but they also produce inaccurate or misleading outputs. Uncertainty Estimation (UE) quantifies the model{'}s confidence and helps users assess response reliability. However, existing UE methods have not been thoroughly examined in scenarios like Retrieval-Augmented Generation (RAG), where the input prompt includes non-parametric knowledge. This paper shows that current UE methods cannot reliably estimate the correctness of LLM responses in the RAG setting. We propose an axiomatic framework to identify deficiencies in existing UE methods. Our framework introduces five constraints that an effective UE method should meet after incorporating retrieved documents into the LLM{'}s prompt. Experimental results reveal that no existing UE method fully satisfies all the axioms, explaining their suboptimal performance in RAG. We further introduce a simple yet effective calibration function based on our framework, which not only satisfies more axioms than baseline methods but also improves the correlation between uncertainty estimates and correctness."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="soudani-etal-2025-uncertainty">
<titleInfo>
<title>Why Uncertainty Estimation Methods Fall Short in RAG: An Axiomatic Analysis</title>
</titleInfo>
<name type="personal">
<namePart type="given">Heydar</namePart>
<namePart type="family">Soudani</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Evangelos</namePart>
<namePart type="family">Kanoulas</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Faegheh</namePart>
<namePart type="family">Hasibi</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2025-07</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Findings of the Association for Computational Linguistics: ACL 2025</title>
</titleInfo>
<name type="personal">
<namePart type="given">Wanxiang</namePart>
<namePart type="family">Che</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Joyce</namePart>
<namePart type="family">Nabende</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Ekaterina</namePart>
<namePart type="family">Shutova</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Mohammad</namePart>
<namePart type="given">Taher</namePart>
<namePart type="family">Pilehvar</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Vienna, Austria</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
<identifier type="isbn">979-8-89176-256-5</identifier>
</relatedItem>
<abstract>Large Language Models (LLMs) are valued for their strong performance across various tasks, but they also produce inaccurate or misleading outputs. Uncertainty Estimation (UE) quantifies the model’s confidence and helps users assess response reliability. However, existing UE methods have not been thoroughly examined in scenarios like Retrieval-Augmented Generation (RAG), where the input prompt includes non-parametric knowledge. This paper shows that current UE methods cannot reliably estimate the correctness of LLM responses in the RAG setting. We propose an axiomatic framework to identify deficiencies in existing UE methods. Our framework introduces five constraints that an effective UE method should meet after incorporating retrieved documents into the LLM’s prompt. Experimental results reveal that no existing UE method fully satisfies all the axioms, explaining their suboptimal performance in RAG. We further introduce a simple yet effective calibration function based on our framework, which not only satisfies more axioms than baseline methods but also improves the correlation between uncertainty estimates and correctness.</abstract>
<identifier type="citekey">soudani-etal-2025-uncertainty</identifier>
<identifier type="doi">10.18653/v1/2025.findings-acl.852</identifier>
<location>
<url>https://aclanthology.org/2025.findings-acl.852/</url>
</location>
<part>
<date>2025-07</date>
<extent unit="page">
<start>16596</start>
<end>16616</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Why Uncertainty Estimation Methods Fall Short in RAG: An Axiomatic Analysis
%A Soudani, Heydar
%A Kanoulas, Evangelos
%A Hasibi, Faegheh
%Y Che, Wanxiang
%Y Nabende, Joyce
%Y Shutova, Ekaterina
%Y Pilehvar, Mohammad Taher
%S Findings of the Association for Computational Linguistics: ACL 2025
%D 2025
%8 July
%I Association for Computational Linguistics
%C Vienna, Austria
%@ 979-8-89176-256-5
%F soudani-etal-2025-uncertainty
%X Large Language Models (LLMs) are valued for their strong performance across various tasks, but they also produce inaccurate or misleading outputs. Uncertainty Estimation (UE) quantifies the model’s confidence and helps users assess response reliability. However, existing UE methods have not been thoroughly examined in scenarios like Retrieval-Augmented Generation (RAG), where the input prompt includes non-parametric knowledge. This paper shows that current UE methods cannot reliably estimate the correctness of LLM responses in the RAG setting. We propose an axiomatic framework to identify deficiencies in existing UE methods. Our framework introduces five constraints that an effective UE method should meet after incorporating retrieved documents into the LLM’s prompt. Experimental results reveal that no existing UE method fully satisfies all the axioms, explaining their suboptimal performance in RAG. We further introduce a simple yet effective calibration function based on our framework, which not only satisfies more axioms than baseline methods but also improves the correlation between uncertainty estimates and correctness.
%R 10.18653/v1/2025.findings-acl.852
%U https://aclanthology.org/2025.findings-acl.852/
%U https://doi.org/10.18653/v1/2025.findings-acl.852
%P 16596-16616
Markdown (Informal)
[Why Uncertainty Estimation Methods Fall Short in RAG: An Axiomatic Analysis](https://aclanthology.org/2025.findings-acl.852/) (Soudani et al., Findings 2025)
ACL