<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="yagubyan-2026-coin">
    <titleInfo>
        <title>The Coin Flip Judge? Reliability and Bias in LLM-as-a-Judge Evaluation</title>
    </titleInfo>
    <name type="personal">
        <namePart type="given">Abel</namePart>
        <namePart type="family">Yagubyan</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 1st Workshop on Grounding Language Models: Learning Faithfully and Efficiently (GroundLM 2026)</title>
        </titleInfo>
        <name type="personal">
            <namePart type="given">Yimu</namePart>
            <namePart type="family">Wang</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Yee</namePart>
            <namePart type="given">Man</namePart>
            <namePart type="family">Choi</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Di</namePart>
            <namePart type="family">Wu</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Siqi</namePart>
            <namePart type="family">Ouyang</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Mozhgan</namePart>
            <namePart type="given">Nasr</namePart>
            <namePart type="family">Azadani</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Yi</namePart>
            <namePart type="given">R</namePart>
            <namePart type="given">(May)</namePart>
            <namePart type="family">Fung</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">979-8-89176-478-1</identifier>
    </relatedItem>
    <abstract>LLM-as-a-Judge is widely used to rank model outputs and populate leaderboards, yet its run-to-run reliability is under-characterized. We study repeated identical evaluations across six judges from three providers spanning three model generations (2024-2026), with 50 pairwise and 50 pointwise trials per question on 29 tasks. Three findings stand out. First, pairwise verdicts are unstable, with an 11-14% mean flip rate on identical re-runs, and this has not improved across generations: the newest, cheapest deployable judges flip just as often as 2024 models. Second, a dedicated swap-order experiment (8,700 judgments) shows the fixed-order asymmetry is order bias, not self-preference, and that order bias has not been fixed either: the 2026 judges reverse its direction, favouring the second slot where earlier judges favoured the first, so a first-position check records them as unbiased while they still decide 21-41% of matched comparisons on slot rather than content. Third, forced-choice verdicts and scalar scores come apart, and ~11 repeated trials are needed to stabilize a verdict. We argue that multi-trial aggregation, symmetric order randomization, and uncertainty reporting should be standard practice.</abstract>
    <identifier type="citekey">yagubyan-2026-coin</identifier>
    <location>
        <url>https://aclanthology.org/2026.groundlm-1.64/</url>
    </location>
    <part>
        <date>2026-10</date>
        <extent unit="page">
            <start>882</start>
            <end>890</end>
        </extent>
    </part>
</mods>
</modsCollection>

%0 Conference Proceedings
%T The Coin Flip Judge? Reliability and Bias in LLM-as-a-Judge Evaluation
%A Yagubyan, Abel
%Y Wang, Yimu
%Y Choi, Yee Man
%Y Wu, Di
%Y Ouyang, Siqi
%Y Azadani, Mozhgan Nasr
%Y Fung, Yi R. (May)
%S Proceedings of the 1st Workshop on Grounding Language Models: Learning Faithfully and Efficiently (GroundLM 2026)
%D 2026
%8 October
%I Association for Computational Linguistics
%C Budapest, Hungary
%@ 979-8-89176-478-1
%F yagubyan-2026-coin
%X LLM-as-a-Judge is widely used to rank model outputs and populate leaderboards, yet its run-to-run reliability is under-characterized. We study repeated identical evaluations across six judges from three providers spanning three model generations (2024-2026), with 50 pairwise and 50 pointwise trials per question on 29 tasks. Three findings stand out. First, pairwise verdicts are unstable, with an 11-14% mean flip rate on identical re-runs, and this has not improved across generations: the newest, cheapest deployable judges flip just as often as 2024 models. Second, a dedicated swap-order experiment (8,700 judgments) shows the fixed-order asymmetry is order bias, not self-preference, and that order bias has not been fixed either: the 2026 judges reverse its direction, favouring the second slot where earlier judges favoured the first, so a first-position check records them as unbiased while they still decide 21-41% of matched comparisons on slot rather than content. Third, forced-choice verdicts and scalar scores come apart, and ~11 repeated trials are needed to stabilize a verdict. We argue that multi-trial aggregation, symmetric order randomization, and uncertainty reporting should be standard practice.
%U https://aclanthology.org/2026.groundlm-1.64/
%P 882-890

Markdown (Informal)

[The Coin Flip Judge? Reliability and Bias in LLM-as-a-Judge Evaluation](https://aclanthology.org/2026.groundlm-1.64/) (Yagubyan, GroundLM 2026)

• The Coin Flip Judge? Reliability and Bias in LLM-as-a-Judge Evaluation (Yagubyan, GroundLM 2026)

ACL

• Abel Yagubyan. 2026. The Coin Flip Judge? Reliability and Bias in LLM-as-a-Judge Evaluation. In Proceedings of the 1st Workshop on Grounding Language Models: Learning Faithfully and Efficiently (GroundLM 2026), pages 882–890, Budapest, Hungary. Association for Computational Linguistics.