<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="rennard-etal-2025-bias">
    <titleInfo>
        <title>Bias in the Mirror : Are LLMs opinions robust to their own adversarial attacks</title>
    </titleInfo>
    <name type="personal">
        <namePart type="given">Virgile</namePart>
        <namePart type="family">Rennard</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Christos</namePart>
        <namePart type="family">Xypolopoulos</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Michalis</namePart>
        <namePart type="family">Vazirgiannis</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <originInfo>
        <dateIssued>2025-07</dateIssued>
    </originInfo>
    <typeOfResource>text</typeOfResource>
    <relatedItem type="host">
        <titleInfo>
            <title>Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)</title>
        </titleInfo>
        <name type="personal">
            <namePart type="given">Wanxiang</namePart>
            <namePart type="family">Che</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Joyce</namePart>
            <namePart type="family">Nabende</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Ekaterina</namePart>
            <namePart type="family">Shutova</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Mohammad</namePart>
            <namePart type="given">Taher</namePart>
            <namePart type="family">Pilehvar</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <originInfo>
            <publisher>Association for Computational Linguistics</publisher>
            <place>
                <placeTerm type="text">Vienna, Austria</placeTerm>
            </place>
        </originInfo>
        <genre authority="marcgt">conference publication</genre>
        <identifier type="isbn">979-8-89176-251-0</identifier>
    </relatedItem>
    <abstract>Large language models (LLMs) inherit biases from their training data and alignment processes, influencing their responses in subtle ways. While many studies have examined these biases, little work has explored their robustness during interactions. In this paper, we introduce a novel approach where two instances of an LLM engage in self-debate, arguing opposing viewpoints to persuade a neutral version of the model. Through this, we evaluate how firmly biases hold and whether models are susceptible to reinforcing misinformation or shifting to harmful viewpoints. Our experiments span multiple LLMs of varying sizes, origins, and languages, providing deeper insights into bias persistence and flexibility across linguistic and cultural contexts.</abstract>
    <identifier type="citekey">rennard-etal-2025-bias</identifier>
    <identifier type="doi">10.18653/v1/2025.acl-long.106</identifier>
    <location>
        <url>https://aclanthology.org/2025.acl-long.106/</url>
    </location>
    <part>
        <date>2025-07</date>
        <extent unit="page">
            <start>2128</start>
            <end>2143</end>
        </extent>
    </part>
</mods>
</modsCollection>

%0 Conference Proceedings
%T Bias in the Mirror : Are LLMs opinions robust to their own adversarial attacks
%A Rennard, Virgile
%A Xypolopoulos, Christos
%A Vazirgiannis, Michalis
%Y Che, Wanxiang
%Y Nabende, Joyce
%Y Shutova, Ekaterina
%Y Pilehvar, Mohammad Taher
%S Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)
%D 2025
%8 July
%I Association for Computational Linguistics
%C Vienna, Austria
%@ 979-8-89176-251-0
%F rennard-etal-2025-bias
%X Large language models (LLMs) inherit biases from their training data and alignment processes, influencing their responses in subtle ways. While many studies have examined these biases, little work has explored their robustness during interactions. In this paper, we introduce a novel approach where two instances of an LLM engage in self-debate, arguing opposing viewpoints to persuade a neutral version of the model. Through this, we evaluate how firmly biases hold and whether models are susceptible to reinforcing misinformation or shifting to harmful viewpoints. Our experiments span multiple LLMs of varying sizes, origins, and languages, providing deeper insights into bias persistence and flexibility across linguistic and cultural contexts.
%R 10.18653/v1/2025.acl-long.106
%U https://aclanthology.org/2025.acl-long.106/
%U https://doi.org/10.18653/v1/2025.acl-long.106
%P 2128-2143

Markdown (Informal)

[Bias in the Mirror : Are LLMs opinions robust to their own adversarial attacks](https://aclanthology.org/2025.acl-long.106/) (Rennard et al., ACL 2025)

• Bias in the Mirror : Are LLMs opinions robust to their own adversarial attacks (Rennard et al., ACL 2025)

ACL

• Virgile Rennard, Christos Xypolopoulos, and Michalis Vazirgiannis. 2025. Bias in the Mirror : Are LLMs opinions robust to their own adversarial attacks. In Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 2128–2143, Vienna, Austria. Association for Computational Linguistics.