<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="zhang-etal-2026-identity">
    <titleInfo>
        <title>Identity-Robust Language Model Generation via Content Integrity Preservation</title>
    </titleInfo>
    <name type="personal">
        <namePart type="given">Miao</namePart>
        <namePart type="family">Zhang</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Kelly</namePart>
        <namePart type="family">Chen</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Mehrab</namePart>
        <namePart type="family">Tanjim</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Rumi</namePart>
        <namePart type="family">Chunara</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <originInfo>
        <dateIssued>2026-07</dateIssued>
    </originInfo>
    <typeOfResource>text</typeOfResource>
    <relatedItem type="host">
        <titleInfo>
            <title>Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)</title>
        </titleInfo>
        <name type="personal">
            <namePart type="given">Maria</namePart>
            <namePart type="family">Liakata</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Viviane</namePart>
            <namePart type="given">P</namePart>
            <namePart type="family">Moreira</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Jiajun</namePart>
            <namePart type="family">Zhang</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">David</namePart>
            <namePart type="family">Jurgens</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <originInfo>
            <publisher>Association for Computational Linguistics</publisher>
            <place>
                <placeTerm type="text">San Diego, California, United States</placeTerm>
            </place>
        </originInfo>
        <genre authority="marcgt">conference publication</genre>
        <identifier type="isbn">979-8-89176-390-6</identifier>
    </relatedItem>
    <abstract>Large Language Model (LLM) outputs often vary across user sociodemographic attributes, leading to disparities in factual accuracy, utility, and safety, even for objective questions where demographic information is irrelevant. Unlike prior work on stereotypical or representational bias, this paper studies identity-dependent degradation of core response quality. We show empirically that such degradation arises from biased generation behavior, despite factual knowledge being robustly encoded across identities. Motivated by this mismatch, we propose a lightweight, training-free framework for identity-robust generation that selectively neutralizes non-critical identity information while preserving semantically essential attributes, thus maintaining output content integrity. Experiments across four benchmarks and 18 sociodemographic identities demonstrate an average 66.3% reduction in identity-dependent bias compared to vanilla prompting and outperforms existing prompt-based defenses. Our work addresses a critical gap in mitigating the impact of user identity cues in prompts on core generation quality.</abstract>
    <identifier type="citekey">zhang-etal-2026-identity</identifier>
    <location>
        <url>https://aclanthology.org/2026.acl-long.884/</url>
    </location>
    <part>
        <date>2026-07</date>
        <extent unit="page">
            <start>19356</start>
            <end>19370</end>
        </extent>
    </part>
</mods>
</modsCollection>

%0 Conference Proceedings
%T Identity-Robust Language Model Generation via Content Integrity Preservation
%A Zhang, Miao
%A Chen, Kelly
%A Tanjim, Mehrab
%A Chunara, Rumi
%Y Liakata, Maria
%Y Moreira, Viviane P.
%Y Zhang, Jiajun
%Y Jurgens, David
%S Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)
%D 2026
%8 July
%I Association for Computational Linguistics
%C San Diego, California, United States
%@ 979-8-89176-390-6
%F zhang-etal-2026-identity
%X Large Language Model (LLM) outputs often vary across user sociodemographic attributes, leading to disparities in factual accuracy, utility, and safety, even for objective questions where demographic information is irrelevant. Unlike prior work on stereotypical or representational bias, this paper studies identity-dependent degradation of core response quality. We show empirically that such degradation arises from biased generation behavior, despite factual knowledge being robustly encoded across identities. Motivated by this mismatch, we propose a lightweight, training-free framework for identity-robust generation that selectively neutralizes non-critical identity information while preserving semantically essential attributes, thus maintaining output content integrity. Experiments across four benchmarks and 18 sociodemographic identities demonstrate an average 66.3% reduction in identity-dependent bias compared to vanilla prompting and outperforms existing prompt-based defenses. Our work addresses a critical gap in mitigating the impact of user identity cues in prompts on core generation quality.
%U https://aclanthology.org/2026.acl-long.884/
%P 19356-19370

Markdown (Informal)

[Identity-Robust Language Model Generation via Content Integrity Preservation](https://aclanthology.org/2026.acl-long.884/) (Zhang et al., ACL 2026)

• Identity-Robust Language Model Generation via Content Integrity Preservation (Zhang et al., ACL 2026)

ACL

• Miao Zhang, Kelly Chen, Mehrab Tanjim, and Rumi Chunara. 2026. Identity-Robust Language Model Generation via Content Integrity Preservation. In Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 19356–19370, San Diego, California, United States. Association for Computational Linguistics.