@inproceedings{jeon-suh-2026-team,
title = "Team hccl-jy at {SHROOM}-Visions: Confidence Signals from an Unmodified {VLM} Verifier",
author = "Jeon, Junyoung and
Suh, Bongwon",
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.60/",
pages = "842--848",
ISBN = "979-8-89176-486-6",
abstract = "Vision-language models (VLMs) can describe details an image does not support. Detecting such hallucinations requires locating the unsupported text rather than flagging whole responses. The SHROOM-Visions shared task asks systems to localize these spans, classify their error type, and assign character-level probabilities. This paper describes the hccl-jy English-track submission. We ask whether confidence from a pretrained VLM verifier can support this task without fine-tuning, and whether candidate-level gains survive extraction and aggregation. The system runs an unmodified Qwen2.5-VL verifier over spaCy candidates with exact character offsets. An aggregation model fitted on held-out calibration data converts candidate judgments into output spans. We compare verbalized confidence, verdict-token probability, and five-draw consistency against annotator agreement, calibration error, and the official metrics. Verdict-token probability separates high-agreement candidates from low-agreement ones more reliably at 32B than at 7B. Averaging repeated judgments improves this ranking further but does not improve the official score. Error analysis shows that the initial score comes mostly from correct empty predictions and that error typing is weak for invented content. Revising only the aggregation over cached judgments raises Cor on development and test data. The findings support evaluating confidence discrimination, calibration, and end-to-end performance as separate targets."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="jeon-suh-2026-team">
<titleInfo>
<title>Team hccl-jy at SHROOM-Visions: Confidence Signals from an Unmodified VLM Verifier</title>
</titleInfo>
<name type="personal">
<namePart type="given">Junyoung</namePart>
<namePart type="family">Jeon</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Bongwon</namePart>
<namePart type="family">Suh</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>Vision-language models (VLMs) can describe details an image does not support. Detecting such hallucinations requires locating the unsupported text rather than flagging whole responses. The SHROOM-Visions shared task asks systems to localize these spans, classify their error type, and assign character-level probabilities. This paper describes the hccl-jy English-track submission. We ask whether confidence from a pretrained VLM verifier can support this task without fine-tuning, and whether candidate-level gains survive extraction and aggregation. The system runs an unmodified Qwen2.5-VL verifier over spaCy candidates with exact character offsets. An aggregation model fitted on held-out calibration data converts candidate judgments into output spans. We compare verbalized confidence, verdict-token probability, and five-draw consistency against annotator agreement, calibration error, and the official metrics. Verdict-token probability separates high-agreement candidates from low-agreement ones more reliably at 32B than at 7B. Averaging repeated judgments improves this ranking further but does not improve the official score. Error analysis shows that the initial score comes mostly from correct empty predictions and that error typing is weak for invented content. Revising only the aggregation over cached judgments raises Cor on development and test data. The findings support evaluating confidence discrimination, calibration, and end-to-end performance as separate targets.</abstract>
<identifier type="citekey">jeon-suh-2026-team</identifier>
<location>
<url>https://aclanthology.org/2026.uncertainlp-main.60/</url>
</location>
<part>
<date>2026-10</date>
<extent unit="page">
<start>842</start>
<end>848</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Team hccl-jy at SHROOM-Visions: Confidence Signals from an Unmodified VLM Verifier
%A Jeon, Junyoung
%A Suh, Bongwon
%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 jeon-suh-2026-team
%X Vision-language models (VLMs) can describe details an image does not support. Detecting such hallucinations requires locating the unsupported text rather than flagging whole responses. The SHROOM-Visions shared task asks systems to localize these spans, classify their error type, and assign character-level probabilities. This paper describes the hccl-jy English-track submission. We ask whether confidence from a pretrained VLM verifier can support this task without fine-tuning, and whether candidate-level gains survive extraction and aggregation. The system runs an unmodified Qwen2.5-VL verifier over spaCy candidates with exact character offsets. An aggregation model fitted on held-out calibration data converts candidate judgments into output spans. We compare verbalized confidence, verdict-token probability, and five-draw consistency against annotator agreement, calibration error, and the official metrics. Verdict-token probability separates high-agreement candidates from low-agreement ones more reliably at 32B than at 7B. Averaging repeated judgments improves this ranking further but does not improve the official score. Error analysis shows that the initial score comes mostly from correct empty predictions and that error typing is weak for invented content. Revising only the aggregation over cached judgments raises Cor on development and test data. The findings support evaluating confidence discrimination, calibration, and end-to-end performance as separate targets.
%U https://aclanthology.org/2026.uncertainlp-main.60/
%P 842-848
Markdown (Informal)
[Team hccl-jy at SHROOM-Visions: Confidence Signals from an Unmodified VLM Verifier](https://aclanthology.org/2026.uncertainlp-main.60/) (Jeon & Suh, UncertaiNLP 2026)
ACL