<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="tong-etal-2026-trait">
    <titleInfo>
        <title>Trait Collapse in Text-to-Image Generation: Evaluating and Mitigating Personality Inconsistency in Visual Storytelling</title>
    </titleInfo>
    <name type="personal">
        <namePart type="given">Yuanbo</namePart>
        <namePart type="family">Tong</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Haoyuan</namePart>
        <namePart type="family">Li</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Jibiao</namePart>
        <namePart type="family">Jiang</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Zirui</namePart>
        <namePart type="family">Wang</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Yuchen</namePart>
        <namePart type="family">Li</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Jiamou</namePart>
        <namePart type="family">Liu</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>Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing</title>
        </titleInfo>
        <name type="personal">
            <namePart type="given">Hung-Yi</namePart>
            <namePart type="family">Lee</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Sunipa</namePart>
            <namePart type="family">Dev</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Jessy</namePart>
            <namePart type="family">Li</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Desmond</namePart>
            <namePart type="family">Elliott</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">979-8-89176-459-0</identifier>
    </relatedItem>
    <abstract>In visual storytelling, a believable character must not only look consistent across scenes but also behave in ways that reflect their personality. Current text-to-image (T2I) models preserve visual appearance effectively, yet consistently fail to maintain personality-relevant behavior—a failure we term trait collapse. We show that conventional metrics (CLIP Score, DINO similarity) are blind to this failure, and that abstract personality descriptors yield no reliable improvement over neutral prompts regardless of abstraction level. To systematically diagnose trait collapse, we introduce ViStoryBench-Personality, a story-level benchmark combining an expert-validated Visual Cues Lexicon (Fleiss’ ąppa = 0.757) with a six-dimensional metric suite covering personality alignment, cross-scene stability, and behavioral coherence. Evaluating seven T2I systems, we confirm a persistent semantic gap between abstract trait labels and visual evidence, and find that visual identity consistency does not imply personality consistency. To bridge this gap, we propose Trait-Behavior-Visual (TBV), a training-free prompting framework that decomposes personality traits into renderable visual constraints. TBV improves Personality Alignment Score by up to +22.1 %, with larger gains in models with stronger language-conditioning capacity.</abstract>
    <identifier type="citekey">tong-etal-2026-trait</identifier>
    <location>
        <url>https://aclanthology.org/2026.emnlp-main.362/</url>
    </location>
    <part>
        <date>2026-10</date>
        <extent unit="page">
            <start>8281</start>
            <end>8306</end>
        </extent>
    </part>
</mods>
</modsCollection>

%0 Conference Proceedings
%T Trait Collapse in Text-to-Image Generation: Evaluating and Mitigating Personality Inconsistency in Visual Storytelling
%A Tong, Yuanbo
%A Li, Haoyuan
%A Jiang, Jibiao
%A Wang, Zirui
%A Li, Yuchen
%A Liu, Jiamou
%Y Lee, Hung-Yi
%Y Dev, Sunipa
%Y Li, Jessy
%Y Elliott, Desmond
%S Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing
%D 2026
%8 October
%I Association for Computational Linguistics
%C Budapest, Hungary
%@ 979-8-89176-459-0
%F tong-etal-2026-trait
%X In visual storytelling, a believable character must not only look consistent across scenes but also behave in ways that reflect their personality. Current text-to-image (T2I) models preserve visual appearance effectively, yet consistently fail to maintain personality-relevant behavior—a failure we term trait collapse. We show that conventional metrics (CLIP Score, DINO similarity) are blind to this failure, and that abstract personality descriptors yield no reliable improvement over neutral prompts regardless of abstraction level. To systematically diagnose trait collapse, we introduce ViStoryBench-Personality, a story-level benchmark combining an expert-validated Visual Cues Lexicon (Fleiss’ ąppa = 0.757) with a six-dimensional metric suite covering personality alignment, cross-scene stability, and behavioral coherence. Evaluating seven T2I systems, we confirm a persistent semantic gap between abstract trait labels and visual evidence, and find that visual identity consistency does not imply personality consistency. To bridge this gap, we propose Trait-Behavior-Visual (TBV), a training-free prompting framework that decomposes personality traits into renderable visual constraints. TBV improves Personality Alignment Score by up to +22.1 %, with larger gains in models with stronger language-conditioning capacity.
%U https://aclanthology.org/2026.emnlp-main.362/
%P 8281-8306

Markdown (Informal)

[Trait Collapse in Text-to-Image Generation: Evaluating and Mitigating Personality Inconsistency in Visual Storytelling](https://aclanthology.org/2026.emnlp-main.362/) (Tong et al., EMNLP 2026)

• Trait Collapse in Text-to-Image Generation: Evaluating and Mitigating Personality Inconsistency in Visual Storytelling (Tong et al., EMNLP 2026)

ACL

• Yuanbo Tong, Haoyuan Li, Jibiao Jiang, Zirui Wang, Yuchen Li, and Jiamou Liu. 2026. Trait Collapse in Text-to-Image Generation: Evaluating and Mitigating Personality Inconsistency in Visual Storytelling. In Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing, pages 8281–8306, Budapest, Hungary. Association for Computational Linguistics.