<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="hoang-ha-etal-2026-web">
    <titleInfo>
        <title>Web Search Is Not Enough: LLMs Fail to Account for Paper Retraction Status</title>
    </titleInfo>
    <name type="personal">
        <namePart type="given">Huy</namePart>
        <namePart type="family">Hoang Ha</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Benoit</namePart>
        <namePart type="family">Favre</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Cyril</namePart>
        <namePart type="family">Labbe</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">François</namePart>
        <namePart type="family">Portet</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>Large language models (LLMs) learn scientific knowledge from static training corpora, while web search can provide updated information such as paper retractions. This matters for LLM-based literature systems because a model may have seen a paper as valid evidence, while current web results report that it has been retracted. If an LLM does not use the updated retraction status, it may cite a retracted paper as valid evidence. We analyze eleven open-weight models on the retraction-status task and ten on the citation-recommendation task. With web-search results provided, LLMs increase correct status answers for retracted papers from 11.0% to 94.0%. But the same models do not use the retrieved context as reliably when no retraction signal is present or when they recommend citations. On non-retracted papers under web search, models answer “I don’t know” in 52.9% of answers in the retraction-status task. In the citation-recommendation task, even when search results state the retraction and the prompt asks for warnings, models recommend retracted papers without warning in 10.2% of answers. A separate pattern appears for highly cited retractions. Models give fewer correct status answers for high-citation than low-citation retractions (89.3% versus 96.4%). OLMo’s training data suggests one reason. Highly cited retractions appear more often in training data, mostly without any retraction signal, so retrieved retraction information may compete with outdated training knowledge. Overall, access to web search does not guarantee that LLMs will correctly use a paper’s current status when answering questions or producing citations, highlighting the broader challenge of knowledge updating in LLM systems.</abstract>
    <identifier type="citekey">hoang-ha-etal-2026-web</identifier>
    <location>
        <url>https://aclanthology.org/2026.findings-emnlp.2070/</url>
    </location>
    <part>
        <date>2026-10</date>
        <extent unit="page">
            <start>44052</start>
            <end>44077</end>
        </extent>
    </part>
</mods>
</modsCollection>

%0 Conference Proceedings
%T Web Search Is Not Enough: LLMs Fail to Account for Paper Retraction Status
%A Hoang Ha, Huy
%A Favre, Benoit
%A Labbe, Cyril
%A Portet, François
%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 hoang-ha-etal-2026-web
%X Large language models (LLMs) learn scientific knowledge from static training corpora, while web search can provide updated information such as paper retractions. This matters for LLM-based literature systems because a model may have seen a paper as valid evidence, while current web results report that it has been retracted. If an LLM does not use the updated retraction status, it may cite a retracted paper as valid evidence. We analyze eleven open-weight models on the retraction-status task and ten on the citation-recommendation task. With web-search results provided, LLMs increase correct status answers for retracted papers from 11.0% to 94.0%. But the same models do not use the retrieved context as reliably when no retraction signal is present or when they recommend citations. On non-retracted papers under web search, models answer “I don’t know” in 52.9% of answers in the retraction-status task. In the citation-recommendation task, even when search results state the retraction and the prompt asks for warnings, models recommend retracted papers without warning in 10.2% of answers. A separate pattern appears for highly cited retractions. Models give fewer correct status answers for high-citation than low-citation retractions (89.3% versus 96.4%). OLMo’s training data suggests one reason. Highly cited retractions appear more often in training data, mostly without any retraction signal, so retrieved retraction information may compete with outdated training knowledge. Overall, access to web search does not guarantee that LLMs will correctly use a paper’s current status when answering questions or producing citations, highlighting the broader challenge of knowledge updating in LLM systems.
%U https://aclanthology.org/2026.findings-emnlp.2070/
%P 44052-44077

Markdown (Informal)

[Web Search Is Not Enough: LLMs Fail to Account for Paper Retraction Status](https://aclanthology.org/2026.findings-emnlp.2070/) (Hoang Ha et al., Findings 2026)

• Web Search Is Not Enough: LLMs Fail to Account for Paper Retraction Status (Hoang Ha et al., Findings 2026)

ACL

• Huy Hoang Ha, Benoit Favre, Cyril Labbe, and François Portet. 2026. Web Search Is Not Enough: LLMs Fail to Account for Paper Retraction Status. In Findings of the Association for Computational Linguistics: EMNLP 2026, pages 44052–44077, Budapest, Hungary. Association for Computational Linguistics.