@inproceedings{torres-etal-2026-citar,
title = "{CITAR}-{NER}-Br: A Unified {NER} Dataset for Fine-Grained Legal Citation Extraction in {B}razilian {P}ortuguese",
author = "Torres, Arthur Elwing and
Silva, Altigran da and
Albuquerque, Hidelberg Oliveira and
Pereira, Ellen Polliana Ramos Souza and
Albuquerque, H{\'e}ldon Jos{\'e} Oliveira and
Da Silva, N{\'a}dia F{\'e}lix Felipe and
Dias, M{\'a}rcio De Souza and
Vit{\'o}rio, Douglas and
Spritzer, Andre Suslik and
de Gouveia, Laura Pereira and
de Gouveia, Rafael Pereira and
Lima, Breno Gabriel De Melo",
editor = "Lee, Hung-Yi and
Dev, Sunipa and
Li, Jessy and
Elliott, Desmond",
booktitle = "Findings of the {A}ssociation for {C}omputational {L}inguistics: {EMNLP} 2026",
month = oct,
year = "2026",
address = "Budapest, Hungary",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2026.findings-emnlp.1410/",
pages = "30056--30069",
ISBN = "979-8-89176-462-0",
abstract = "We present CITAR-NER-Br, a unified NER dataset for direct extraction of fine-grained legal citation subcomponents in Brazilian Portuguese. CITAR-NER-Br annotates six citation subcomponents over 270 legal documents from LeNER-Br and Ulysses Tesem{\~o}, capturing the overlapping and discontinuous spans intrinsic to legal citations. To the best of our knowledge, this is the first unified NER dataset for the legal domain specifically designed to represent discontinuous entities, which account for 44.8{\%} of annotations, the highest proportion among existing legal NER corpora. We establish baselines using four configurations across three LLM-based and non-LLM NER methods to support future research. Our experiments demonstrate that, while non-LLM methods achieve strong performance, they remain heavily reliant on high-quality training data. Conversely, LLM-based methods achieve overall lower performance, but require only few-shot supervision. Moreover, the traditional exact-match NER evaluation approach significantly penalizes LLM-based methods even when their outputs preserve semantic accuracy."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="torres-etal-2026-citar">
<titleInfo>
<title>CITAR-NER-Br: A Unified NER Dataset for Fine-Grained Legal Citation Extraction in Brazilian Portuguese</title>
</titleInfo>
<name type="personal">
<namePart type="given">Arthur</namePart>
<namePart type="given">Elwing</namePart>
<namePart type="family">Torres</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Altigran</namePart>
<namePart type="given">da</namePart>
<namePart type="family">Silva</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Hidelberg</namePart>
<namePart type="given">Oliveira</namePart>
<namePart type="family">Albuquerque</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Ellen</namePart>
<namePart type="given">Polliana</namePart>
<namePart type="given">Ramos</namePart>
<namePart type="given">Souza</namePart>
<namePart type="family">Pereira</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Héldon</namePart>
<namePart type="given">José</namePart>
<namePart type="given">Oliveira</namePart>
<namePart type="family">Albuquerque</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Nádia</namePart>
<namePart type="given">Félix</namePart>
<namePart type="given">Felipe</namePart>
<namePart type="family">Da Silva</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Márcio</namePart>
<namePart type="given">De</namePart>
<namePart type="given">Souza</namePart>
<namePart type="family">Dias</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Douglas</namePart>
<namePart type="family">Vitório</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Andre</namePart>
<namePart type="given">Suslik</namePart>
<namePart type="family">Spritzer</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Laura</namePart>
<namePart type="given">Pereira</namePart>
<namePart type="family">de Gouveia</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Rafael</namePart>
<namePart type="given">Pereira</namePart>
<namePart type="family">de Gouveia</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Breno</namePart>
<namePart type="given">Gabriel</namePart>
<namePart type="given">De</namePart>
<namePart type="given">Melo</namePart>
<namePart type="family">Lima</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>We present CITAR-NER-Br, a unified NER dataset for direct extraction of fine-grained legal citation subcomponents in Brazilian Portuguese. CITAR-NER-Br annotates six citation subcomponents over 270 legal documents from LeNER-Br and Ulysses Tesemõ, capturing the overlapping and discontinuous spans intrinsic to legal citations. To the best of our knowledge, this is the first unified NER dataset for the legal domain specifically designed to represent discontinuous entities, which account for 44.8% of annotations, the highest proportion among existing legal NER corpora. We establish baselines using four configurations across three LLM-based and non-LLM NER methods to support future research. Our experiments demonstrate that, while non-LLM methods achieve strong performance, they remain heavily reliant on high-quality training data. Conversely, LLM-based methods achieve overall lower performance, but require only few-shot supervision. Moreover, the traditional exact-match NER evaluation approach significantly penalizes LLM-based methods even when their outputs preserve semantic accuracy.</abstract>
<identifier type="citekey">torres-etal-2026-citar</identifier>
<location>
<url>https://aclanthology.org/2026.findings-emnlp.1410/</url>
</location>
<part>
<date>2026-10</date>
<extent unit="page">
<start>30056</start>
<end>30069</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T CITAR-NER-Br: A Unified NER Dataset for Fine-Grained Legal Citation Extraction in Brazilian Portuguese
%A Torres, Arthur Elwing
%A Silva, Altigran da
%A Albuquerque, Hidelberg Oliveira
%A Pereira, Ellen Polliana Ramos Souza
%A Albuquerque, Héldon José Oliveira
%A Da Silva, Nádia Félix Felipe
%A Dias, Márcio De Souza
%A Vitório, Douglas
%A Spritzer, Andre Suslik
%A de Gouveia, Laura Pereira
%A de Gouveia, Rafael Pereira
%A Lima, Breno Gabriel De Melo
%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 torres-etal-2026-citar
%X We present CITAR-NER-Br, a unified NER dataset for direct extraction of fine-grained legal citation subcomponents in Brazilian Portuguese. CITAR-NER-Br annotates six citation subcomponents over 270 legal documents from LeNER-Br and Ulysses Tesemõ, capturing the overlapping and discontinuous spans intrinsic to legal citations. To the best of our knowledge, this is the first unified NER dataset for the legal domain specifically designed to represent discontinuous entities, which account for 44.8% of annotations, the highest proportion among existing legal NER corpora. We establish baselines using four configurations across three LLM-based and non-LLM NER methods to support future research. Our experiments demonstrate that, while non-LLM methods achieve strong performance, they remain heavily reliant on high-quality training data. Conversely, LLM-based methods achieve overall lower performance, but require only few-shot supervision. Moreover, the traditional exact-match NER evaluation approach significantly penalizes LLM-based methods even when their outputs preserve semantic accuracy.
%U https://aclanthology.org/2026.findings-emnlp.1410/
%P 30056-30069
Markdown (Informal)
[CITAR-NER-Br: A Unified NER Dataset for Fine-Grained Legal Citation Extraction in Brazilian Portuguese](https://aclanthology.org/2026.findings-emnlp.1410/) (Torres et al., Findings 2026)
ACL
- Arthur Elwing Torres, Altigran da Silva, Hidelberg Oliveira Albuquerque, Ellen Polliana Ramos Souza Pereira, Héldon José Oliveira Albuquerque, Nádia Félix Felipe Da Silva, Márcio De Souza Dias, Douglas Vitório, Andre Suslik Spritzer, Laura Pereira de Gouveia, Rafael Pereira de Gouveia, and Breno Gabriel De Melo Lima. 2026. CITAR-NER-Br: A Unified NER Dataset for Fine-Grained Legal Citation Extraction in Brazilian Portuguese. In Findings of the Association for Computational Linguistics: EMNLP 2026, pages 30056–30069, Budapest, Hungary. Association for Computational Linguistics.