@inproceedings{muller-etal-2026-semantic,
title = "Semantic Calibration Prevails Where Token Confidence Fails: Benchmarking Long-Form Scientific {QA}",
author = {M{\"u}ller, Philip and
Popovi{\v{c}}, Nicholas and
F{\"a}rber, Michael and
Steinbach, Peter},
editor = {Aziz, Wilker and
Berant, Jonathan and
Eikema, Bryan and
Ilia, Evgenia and
de Marneffe, Marie-Catherine and
Plank, Barbara and
Shelmanov, Artem and
Swayamdipta, Swabha and
Tiedemann, J{\"o}rg and
Vazhentsev, Artem and
V{\'a}zquez, Ra{\'u}l and
Zerva, Chrysoula},
booktitle = "The Proceedings of the Third Workshop on Uncertainty-Aware {NLP}",
month = oct,
year = "2026",
address = "Budapest, Hungary",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2026.uncertainlp-main.11/",
pages = "162--200",
ISBN = "979-8-89176-486-6",
abstract = "Reliable uncertainty quantification (UQ) is essential for safe deployment of large language models (LLMs) in scientific question answering, where long-form outputs exceed practical human verification at scale. We introduce the first large-scale benchmark for UQ calibration in long-form, reasoning-demanding scientific QA, evaluating four UQ methods on 685,000 responses across up to 20 LLMs and seven datasets, supported by an extensible open-source framework whose shared-generation design enables reproducible cross-method comparisons. Instruction tuning is shown to associate with systematic token probability polarization, collapsing confidence distributions and undermining the reliability of token-level uncertainty signals. Reasoning model families diverge: some reproduce this polarization while others actively mitigate it, a pattern that clusters by provider and suggests training pipeline design as a key differentiating factor. Verbalized and token-aggregation sequence-level methods fail systematically. Only semantic consistency, as measured by consistency of the final answer, yields well-calibrated outputs, providing the first large-scale evidence that semantic calibration persists in multi-step, dependency-rich reasoning settings."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="muller-etal-2026-semantic">
<titleInfo>
<title>Semantic Calibration Prevails Where Token Confidence Fails: Benchmarking Long-Form Scientific QA</title>
</titleInfo>
<name type="personal">
<namePart type="given">Philip</namePart>
<namePart type="family">Müller</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Nicholas</namePart>
<namePart type="family">Popovič</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Michael</namePart>
<namePart type="family">Färber</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Peter</namePart>
<namePart type="family">Steinbach</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2026-10</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>The Proceedings of the Third Workshop on Uncertainty-Aware NLP</title>
</titleInfo>
<name type="personal">
<namePart type="given">Wilker</namePart>
<namePart type="family">Aziz</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jonathan</namePart>
<namePart type="family">Berant</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Bryan</namePart>
<namePart type="family">Eikema</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Evgenia</namePart>
<namePart type="family">Ilia</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Marie-Catherine</namePart>
<namePart type="family">de Marneffe</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Barbara</namePart>
<namePart type="family">Plank</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Artem</namePart>
<namePart type="family">Shelmanov</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Swabha</namePart>
<namePart type="family">Swayamdipta</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jörg</namePart>
<namePart type="family">Tiedemann</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Artem</namePart>
<namePart type="family">Vazhentsev</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Raúl</namePart>
<namePart type="family">Vázquez</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Chrysoula</namePart>
<namePart type="family">Zerva</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Budapest, Hungary</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
<identifier type="isbn">979-8-89176-486-6</identifier>
</relatedItem>
<abstract>Reliable uncertainty quantification (UQ) is essential for safe deployment of large language models (LLMs) in scientific question answering, where long-form outputs exceed practical human verification at scale. We introduce the first large-scale benchmark for UQ calibration in long-form, reasoning-demanding scientific QA, evaluating four UQ methods on 685,000 responses across up to 20 LLMs and seven datasets, supported by an extensible open-source framework whose shared-generation design enables reproducible cross-method comparisons. Instruction tuning is shown to associate with systematic token probability polarization, collapsing confidence distributions and undermining the reliability of token-level uncertainty signals. Reasoning model families diverge: some reproduce this polarization while others actively mitigate it, a pattern that clusters by provider and suggests training pipeline design as a key differentiating factor. Verbalized and token-aggregation sequence-level methods fail systematically. Only semantic consistency, as measured by consistency of the final answer, yields well-calibrated outputs, providing the first large-scale evidence that semantic calibration persists in multi-step, dependency-rich reasoning settings.</abstract>
<identifier type="citekey">muller-etal-2026-semantic</identifier>
<location>
<url>https://aclanthology.org/2026.uncertainlp-main.11/</url>
</location>
<part>
<date>2026-10</date>
<extent unit="page">
<start>162</start>
<end>200</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Semantic Calibration Prevails Where Token Confidence Fails: Benchmarking Long-Form Scientific QA
%A Müller, Philip
%A Popovič, Nicholas
%A Färber, Michael
%A Steinbach, Peter
%Y Aziz, Wilker
%Y Berant, Jonathan
%Y Eikema, Bryan
%Y Ilia, Evgenia
%Y de Marneffe, Marie-Catherine
%Y Plank, Barbara
%Y Shelmanov, Artem
%Y Swayamdipta, Swabha
%Y Tiedemann, Jörg
%Y Vazhentsev, Artem
%Y Vázquez, Raúl
%Y Zerva, Chrysoula
%S The Proceedings of the Third Workshop on Uncertainty-Aware NLP
%D 2026
%8 October
%I Association for Computational Linguistics
%C Budapest, Hungary
%@ 979-8-89176-486-6
%F muller-etal-2026-semantic
%X Reliable uncertainty quantification (UQ) is essential for safe deployment of large language models (LLMs) in scientific question answering, where long-form outputs exceed practical human verification at scale. We introduce the first large-scale benchmark for UQ calibration in long-form, reasoning-demanding scientific QA, evaluating four UQ methods on 685,000 responses across up to 20 LLMs and seven datasets, supported by an extensible open-source framework whose shared-generation design enables reproducible cross-method comparisons. Instruction tuning is shown to associate with systematic token probability polarization, collapsing confidence distributions and undermining the reliability of token-level uncertainty signals. Reasoning model families diverge: some reproduce this polarization while others actively mitigate it, a pattern that clusters by provider and suggests training pipeline design as a key differentiating factor. Verbalized and token-aggregation sequence-level methods fail systematically. Only semantic consistency, as measured by consistency of the final answer, yields well-calibrated outputs, providing the first large-scale evidence that semantic calibration persists in multi-step, dependency-rich reasoning settings.
%U https://aclanthology.org/2026.uncertainlp-main.11/
%P 162-200
Markdown (Informal)
[Semantic Calibration Prevails Where Token Confidence Fails: Benchmarking Long-Form Scientific QA](https://aclanthology.org/2026.uncertainlp-main.11/) (Müller et al., UncertaiNLP 2026)
ACL