<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="nikeghbal-etal-2026-flips">
    <titleInfo>
        <title>Who Flips? Self- and Cross-Model Counterarguments Reveal Answer Instability in LLMs</title>
    </titleInfo>
    <name type="personal">
        <namePart type="given">Nafiseh</namePart>
        <namePart type="family">Nikeghbal</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Amir</namePart>
        <namePart type="given">Hossein</namePart>
        <namePart type="family">Kargaran</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Shaghayegh</namePart>
        <namePart type="family">Kolli</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Jana</namePart>
        <namePart type="family">Diesner</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>Findings of the Association for Computational Linguistics: EMNLP 2026</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">979-8-89176-462-0</identifier>
    </relatedItem>
    <abstract>Standard accuracy benchmarks evaluate whether large language models (LLMs) reach correct answers. However, they do not test whether models maintain that answer when challenged by a plausible counter-argument. We introduce a controlled protocol for evaluating answer stability: after a model answers a multiple-choice question correctly, we challenge the model’s answer with a coherent argument for an incorrect option and measure whether the model flips. The setup a) isolates argumentative content from overt social pressure and b) varies argument length, self-attribution, and cross-model source. Across seven frontier models and 57 MMLU subjects, flip rates range from 17.5% to 97.3%, revealing large differences in stability that are not captured by accuracy metrics alone. We find that self-attribution consistently increases flip rates (mean +7.1pp, up to +18.7pp). Furthermore, pooling wrong-answer arguments across models and selecting the most effective one per question yields stronger adversarial challenges than relying on any single source model. From this cross-model pool, we construct MaxFlip, a curated benchmark that amplifies answer flips by up to +23.6pp over self-generated challenges. We release the protocol, challenge records, and MaxFlip to support stability evaluation alongside standard accuracy benchmarks. Materials are available at https://github.com/nafisenik/WhoFlips, https://hf.co/datasets/nafisehNik/WhoFlips.</abstract>
    <identifier type="citekey">nikeghbal-etal-2026-flips</identifier>
    <location>
        <url>https://aclanthology.org/2026.findings-emnlp.1822/</url>
    </location>
    <part>
        <date>2026-10</date>
        <extent unit="page">
            <start>38928</start>
            <end>38942</end>
        </extent>
    </part>
</mods>
</modsCollection>

%0 Conference Proceedings
%T Who Flips? Self- and Cross-Model Counterarguments Reveal Answer Instability in LLMs
%A Nikeghbal, Nafiseh
%A Kargaran, Amir Hossein
%A Kolli, Shaghayegh
%A Diesner, Jana
%Y Lee, Hung-Yi
%Y Dev, Sunipa
%Y Li, Jessy
%Y Elliott, Desmond
%S Findings of the Association for Computational Linguistics: EMNLP 2026
%D 2026
%8 October
%I Association for Computational Linguistics
%C Budapest, Hungary
%@ 979-8-89176-462-0
%F nikeghbal-etal-2026-flips
%X Standard accuracy benchmarks evaluate whether large language models (LLMs) reach correct answers. However, they do not test whether models maintain that answer when challenged by a plausible counter-argument. We introduce a controlled protocol for evaluating answer stability: after a model answers a multiple-choice question correctly, we challenge the model’s answer with a coherent argument for an incorrect option and measure whether the model flips. The setup a) isolates argumentative content from overt social pressure and b) varies argument length, self-attribution, and cross-model source. Across seven frontier models and 57 MMLU subjects, flip rates range from 17.5% to 97.3%, revealing large differences in stability that are not captured by accuracy metrics alone. We find that self-attribution consistently increases flip rates (mean +7.1pp, up to +18.7pp). Furthermore, pooling wrong-answer arguments across models and selecting the most effective one per question yields stronger adversarial challenges than relying on any single source model. From this cross-model pool, we construct MaxFlip, a curated benchmark that amplifies answer flips by up to +23.6pp over self-generated challenges. We release the protocol, challenge records, and MaxFlip to support stability evaluation alongside standard accuracy benchmarks. Materials are available at https://github.com/nafisenik/WhoFlips, https://hf.co/datasets/nafisehNik/WhoFlips.
%U https://aclanthology.org/2026.findings-emnlp.1822/
%P 38928-38942

Markdown (Informal)

[Who Flips? Self- and Cross-Model Counterarguments Reveal Answer Instability in LLMs](https://aclanthology.org/2026.findings-emnlp.1822/) (Nikeghbal et al., Findings 2026)

• Who Flips? Self- and Cross-Model Counterarguments Reveal Answer Instability in LLMs (Nikeghbal et al., Findings 2026)

ACL

• Nafiseh Nikeghbal, Amir Hossein Kargaran, Shaghayegh Kolli, and Jana Diesner. 2026. Who Flips? Self- and Cross-Model Counterarguments Reveal Answer Instability in LLMs. In Findings of the Association for Computational Linguistics: EMNLP 2026, pages 38928–38942, Budapest, Hungary. Association for Computational Linguistics.