<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="buchmann-gurevych-2026-citation">
    <titleInfo>
        <title>Citation Failure: Definition, Analysis and Efficient Mitigation</title>
    </titleInfo>
    <name type="personal">
        <namePart type="given">Jan</namePart>
        <namePart type="family">Buchmann</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Iryna</namePart>
        <namePart type="family">Gurevych</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <originInfo>
        <dateIssued>2026</dateIssued>
    </originInfo>
    <typeOfResource>text</typeOfResource>
    <genre authority="bibutilsgt">journal article</genre>
    <relatedItem type="host">
        <titleInfo>
            <title>Transactions of the Association for Computational Linguistics</title>
        </titleInfo>
        <originInfo>
            <issuance>continuing</issuance>
            <publisher>MIT Press</publisher>
            <place>
                <placeTerm type="text">Cambridge, MA</placeTerm>
            </place>
        </originInfo>
        <genre authority="marcgt">periodical</genre>
        <genre authority="bibutilsgt">academic journal</genre>
    </relatedItem>
    <abstract>Citations from LLM-based RAG systems are supposed to simplify response verification. However, this goal is undermined in cases of citation failure, where a model generates a helpful response, but fails to generate citations to complete evidence. In contrast to previous work, we propose to disentangle this from response failure, where the response itself is flawed, and citing complete evidence is impossible. To address citation failure, this work follows a two-step approach: (1) We study when citation failure occurs and (2) how it can be mitigated efficiently. For step 1, we extend prior work by investigating how the relation between response and evidence affects citation quality. We introduce CITE-CONTROL, a benchmark that systematically varies this relation to enable the analysis of failure modes. Experiments show that failures increase with relational complexity and suggest that combining citation methods could improve performance, motivating step 2. To study the efficient improvement of LLM citation, we propose CITENTION, a framework integrating generative, attention-based, and retrieval-based methods. Results demonstrate substantial citation improvements on CITECONTROL and in transfer settings. We make our data and code publicly available.1</abstract>
    <identifier type="citekey">buchmann-gurevych-2026-citation</identifier>
    <identifier type="doi">10.1162/tacl.a.736</identifier>
    <location>
        <url>https://aclanthology.org/2026.tacl-1.66/</url>
    </location>
    <part>
        <date>2026</date>
        <detail type="volume"><number>14</number></detail>
        <extent unit="page">
            <start>1453</start>
            <end>1481</end>
        </extent>
    </part>
</mods>
</modsCollection>

%0 Journal Article
%T Citation Failure: Definition, Analysis and Efficient Mitigation
%A Buchmann, Jan
%A Gurevych, Iryna
%J Transactions of the Association for Computational Linguistics
%D 2026
%V 14
%I MIT Press
%C Cambridge, MA
%F buchmann-gurevych-2026-citation
%X Citations from LLM-based RAG systems are supposed to simplify response verification. However, this goal is undermined in cases of citation failure, where a model generates a helpful response, but fails to generate citations to complete evidence. In contrast to previous work, we propose to disentangle this from response failure, where the response itself is flawed, and citing complete evidence is impossible. To address citation failure, this work follows a two-step approach: (1) We study when citation failure occurs and (2) how it can be mitigated efficiently. For step 1, we extend prior work by investigating how the relation between response and evidence affects citation quality. We introduce CITE-CONTROL, a benchmark that systematically varies this relation to enable the analysis of failure modes. Experiments show that failures increase with relational complexity and suggest that combining citation methods could improve performance, motivating step 2. To study the efficient improvement of LLM citation, we propose CITENTION, a framework integrating generative, attention-based, and retrieval-based methods. Results demonstrate substantial citation improvements on CITECONTROL and in transfer settings. We make our data and code publicly available.1
%R 10.1162/tacl.a.736
%U https://aclanthology.org/2026.tacl-1.66/
%U https://doi.org/10.1162/tacl.a.736
%P 1453-1481

Markdown (Informal)

[Citation Failure: Definition, Analysis and Efficient Mitigation](https://aclanthology.org/2026.tacl-1.66/) (Buchmann & Gurevych, TACL 2026)

• Citation Failure: Definition, Analysis and Efficient Mitigation (Buchmann & Gurevych, TACL 2026)

ACL

• Jan Buchmann and Iryna Gurevych. 2026. Citation Failure: Definition, Analysis and Efficient Mitigation. Transactions of the Association for Computational Linguistics, 14:1453–1481.