<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="zhong-etal-2026-evidence">
    <titleInfo>
        <title>More Evidence Is Not Always More Support: Controlled Diagnostics for Evidence Aggregation in Fact Verification</title>
    </titleInfo>
    <name type="personal">
        <namePart type="given">Shaoxin</namePart>
        <namePart type="family">Zhong</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Yuchen</namePart>
        <namePart type="family">Su</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Qian</namePart>
        <namePart type="family">Liu</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Michael</namePart>
        <namePart type="given">J</namePart>
        <namePart type="family">Witbrock</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>Final-context accuracy in multi-passage fact verification can hide whether an added passage contributes a new verdict-relevant fact or merely repeats evidence already present. We introduce a controlled same-prefix diagnostic derived from HoVer supporting-fact annotations to measure marginal evidence contribution. For each claim, a shared prefix covers one fact’s evidence while leaving another required fact absent; the next passage either repeats the covered fact or supplies the missing fact. The diagnostic contains 1,180 same-prefix pairs and an exact-repetition ladder testing whether repeated copies of one evidence sentence can change a verifier’s label. Supervised verifiers with similar clean F1 scores, from 0.6324 to 0.6528, differ sharply in eight-repeat flip rate, from 8.5% to 30.3%. Conservative pooling and natural-language-inference (NLI) scoring reduce duplicate sensitivity but give near-zero missing-fact gain, while gain-oriented systems improve gain but remain repetition-sensitive. Latent-Gated Evidence Dynamics (LGED), a compact stateful reference model, shows that low duplicate sensitivity and positive missing-fact gain can coexist. Surface controls and decision-level analyses support this diagnostic interpretation. Redundancy and reader tests extend the duplicate-sensitivity finding beyond HoVer, while prompted open-weight LLMs reveal a gap between role recognition and verdict updating.</abstract>
    <identifier type="citekey">zhong-etal-2026-evidence</identifier>
    <location>
        <url>https://aclanthology.org/2026.findings-emnlp.747/</url>
    </location>
    <part>
        <date>2026-10</date>
        <extent unit="page">
            <start>15724</start>
            <end>15747</end>
        </extent>
    </part>
</mods>
</modsCollection>

%0 Conference Proceedings
%T More Evidence Is Not Always More Support: Controlled Diagnostics for Evidence Aggregation in Fact Verification
%A Zhong, Shaoxin
%A Su, Yuchen
%A Liu, Qian
%A Witbrock, Michael J.
%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 zhong-etal-2026-evidence
%X Final-context accuracy in multi-passage fact verification can hide whether an added passage contributes a new verdict-relevant fact or merely repeats evidence already present. We introduce a controlled same-prefix diagnostic derived from HoVer supporting-fact annotations to measure marginal evidence contribution. For each claim, a shared prefix covers one fact’s evidence while leaving another required fact absent; the next passage either repeats the covered fact or supplies the missing fact. The diagnostic contains 1,180 same-prefix pairs and an exact-repetition ladder testing whether repeated copies of one evidence sentence can change a verifier’s label. Supervised verifiers with similar clean F1 scores, from 0.6324 to 0.6528, differ sharply in eight-repeat flip rate, from 8.5% to 30.3%. Conservative pooling and natural-language-inference (NLI) scoring reduce duplicate sensitivity but give near-zero missing-fact gain, while gain-oriented systems improve gain but remain repetition-sensitive. Latent-Gated Evidence Dynamics (LGED), a compact stateful reference model, shows that low duplicate sensitivity and positive missing-fact gain can coexist. Surface controls and decision-level analyses support this diagnostic interpretation. Redundancy and reader tests extend the duplicate-sensitivity finding beyond HoVer, while prompted open-weight LLMs reveal a gap between role recognition and verdict updating.
%U https://aclanthology.org/2026.findings-emnlp.747/
%P 15724-15747

Markdown (Informal)

[More Evidence Is Not Always More Support: Controlled Diagnostics for Evidence Aggregation in Fact Verification](https://aclanthology.org/2026.findings-emnlp.747/) (Zhong et al., Findings 2026)

• More Evidence Is Not Always More Support: Controlled Diagnostics for Evidence Aggregation in Fact Verification (Zhong et al., Findings 2026)

ACL

• Shaoxin Zhong, Yuchen Su, Qian Liu, and Michael J. Witbrock. 2026. More Evidence Is Not Always More Support: Controlled Diagnostics for Evidence Aggregation in Fact Verification. In Findings of the Association for Computational Linguistics: EMNLP 2026, pages 15724–15747, Budapest, Hungary. Association for Computational Linguistics.