<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="wu-etal-2026-counter">
    <titleInfo>
        <title>Counter-Stereotypic Evidence Is Not Enough: Unequal Uptake in LLMs</title>
    </titleInfo>
    <name type="personal">
        <namePart type="given">Jinxuan</namePart>
        <namePart type="family">Wu</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Bin</namePart>
        <namePart type="family">Li</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Xiangyang</namePart>
        <namePart type="family">Xue</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>Recent work shows that context can both amplify and mitigate social bias in LLMs. Yet a key question remains underexplored: when context provides counter-stereotypic evidence, is that evidence taken up equitably across demographic targets? Using agency–communality bias as a testbed, we introduce a sequential evaluation framework that holds total counter-evidence fixed while varying its evidence structure, adapting the bookkeeping–subtyping contrast from social cognition. Across three instruction-tuned LLMs evaluated in English, counter-evidence is taken up in explicit group-level judgement and transfers to explicit judgement about held-out individuals. The prior-category structure of this uptake, however, does not carry over to generation: communal-prior targets lose their correction advantage, absorb evidence more slowly, and the weakest uptake repeatedly appears for Asian and Hispanic female targets within our controlled target set. The tested prompt instruction and DPO fine-tuning variants yield partial improvements at prior-category level but do not reliably repair these target-level gaps. These findings show that fairness audits for context-rich LLM systems should go beyond checking whether corrective evidence is present and test whether models use that evidence equitably in downstream generation.</abstract>
    <identifier type="citekey">wu-etal-2026-counter</identifier>
    <location>
        <url>https://aclanthology.org/2026.emnlp-main.388/</url>
    </location>
    <part>
        <date>2026-10</date>
        <extent unit="page">
            <start>8892</start>
            <end>8923</end>
        </extent>
    </part>
</mods>
</modsCollection>

%0 Conference Proceedings
%T Counter-Stereotypic Evidence Is Not Enough: Unequal Uptake in LLMs
%A Wu, Jinxuan
%A Li, Bin
%A Xue, Xiangyang
%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 wu-etal-2026-counter
%X Recent work shows that context can both amplify and mitigate social bias in LLMs. Yet a key question remains underexplored: when context provides counter-stereotypic evidence, is that evidence taken up equitably across demographic targets? Using agency–communality bias as a testbed, we introduce a sequential evaluation framework that holds total counter-evidence fixed while varying its evidence structure, adapting the bookkeeping–subtyping contrast from social cognition. Across three instruction-tuned LLMs evaluated in English, counter-evidence is taken up in explicit group-level judgement and transfers to explicit judgement about held-out individuals. The prior-category structure of this uptake, however, does not carry over to generation: communal-prior targets lose their correction advantage, absorb evidence more slowly, and the weakest uptake repeatedly appears for Asian and Hispanic female targets within our controlled target set. The tested prompt instruction and DPO fine-tuning variants yield partial improvements at prior-category level but do not reliably repair these target-level gaps. These findings show that fairness audits for context-rich LLM systems should go beyond checking whether corrective evidence is present and test whether models use that evidence equitably in downstream generation.
%U https://aclanthology.org/2026.emnlp-main.388/
%P 8892-8923

Markdown (Informal)

[Counter-Stereotypic Evidence Is Not Enough: Unequal Uptake in LLMs](https://aclanthology.org/2026.emnlp-main.388/) (Wu et al., EMNLP 2026)

• Counter-Stereotypic Evidence Is Not Enough: Unequal Uptake in LLMs (Wu et al., EMNLP 2026)

ACL

• Jinxuan Wu, Bin Li, and Xiangyang Xue. 2026. Counter-Stereotypic Evidence Is Not Enough: Unequal Uptake in LLMs. In Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing, pages 8892–8923, Budapest, Hungary. Association for Computational Linguistics.