@inproceedings{lorini-etal-2026-one,
title = "One Size Does Not Fit All: Why {EU} Legislative Translation Demands Domain-Specific Fine-Tuning of {LLM}s",
author = "Lorini, Valerio and
Vlaic, Paula and
Akbulut, Ulascan and
Marcoaldi, Daniele",
editor = "Shterionov, Dimitar and
Vanmassenhove, Eva and
De Sisto, Mirella and
Blain, Fred and
Pourmostafa Roshan Sharami, Javad and
Lepp, Lisa and
Manna, Chiara and
Rescigno, Argentina Anna and
Karakanta, Alina and
Rigouts Terryn, Ayla and
Lardelli, Manuel and
Resende, Natalia and
Murgolo, Elena and
Hackenbuchner, Jani{\c{c}}a and
Zaretskaya, Anna and
Espl{\`a}-Gomis, Miquel and
Etchegoyhen, Thierry and
Gromann, Dagmar and
Bawden, Rachel and
Haddow, Barry and
Szoc, Sara and
Forcada, Mikel and
Moniz, Helena",
booktitle = "Proceedings of the 26th Annual Conference of the {E}uropean Association for Machine Translation (Volume 1)",
month = jun,
year = "2026",
address = "Tilburg, The Netherlands",
publisher = "European Association for Machine Translation",
url = "https://aclanthology.org/2026.eamt-1.20/",
pages = "304--320",
ISBN = "9789403901411",
abstract = "EU legislation is equally authentic and legally binding in all 24 official languages, rendering high-quality translation a legal obligation rather than a mere choice. Therefore, high-quality language technology supporting translation processes in all EU languages are essential for language professionals at the European Parliament (EP). This paper investigates whether domain-specific fine-tuning of an open-weight Large Language Model (LLM) yields consistently larger quality gains on legislative text compared to generic text, in all 23 EU target languages from English. We evaluate ten experimental conditions: base model, in-domain and cross-domain fine-tuning, sequential generic-then-legislative fine-tuning, and zero-shot Claude Sonnet 4.6 as a proprietary reference. We analyse BLEU, chrF, TER, and COMET metrics on nearly 700,000 segments. Results confirm the hypothesis for all 23 languages: legislative fine-tuning enhances BLEU by +12.30 compared to +7.10 for generic fine-tuning, demonstrating a consistent advantage of +5.20 BLEU in all the metrics. The fine-tuned EuroLLM-22B decisively outperforms Claude Sonnet 4.6, Anthropic{'}s latest frontier model, on both domains, highlighting that targeted adaptation of a smaller open-weight model can surpass a state-of-the-art proprietary system. Cross-domain transfer within institutional domain is positive for all languages, with no catastrophic forgetting. Low-resource languages such as Irish and Maltese benefit the most from fine-tuning, while a divergence between BLEU and COMET rankings for some languages underlines the need of evaluation metrics alongside traditional measures."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="lorini-etal-2026-one">
<titleInfo>
<title>One Size Does Not Fit All: Why EU Legislative Translation Demands Domain-Specific Fine-Tuning of LLMs</title>
</titleInfo>
<name type="personal">
<namePart type="given">Valerio</namePart>
<namePart type="family">Lorini</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Paula</namePart>
<namePart type="family">Vlaic</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Ulascan</namePart>
<namePart type="family">Akbulut</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Daniele</namePart>
<namePart type="family">Marcoaldi</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2026-06</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of the 26th Annual Conference of the European Association for Machine Translation (Volume 1)</title>
</titleInfo>
<name type="personal">
<namePart type="given">Dimitar</namePart>
<namePart type="family">Shterionov</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Eva</namePart>
<namePart type="family">Vanmassenhove</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Mirella</namePart>
<namePart type="family">De Sisto</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Fred</namePart>
<namePart type="family">Blain</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Javad</namePart>
<namePart type="family">Pourmostafa Roshan Sharami</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Lisa</namePart>
<namePart type="family">Lepp</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Chiara</namePart>
<namePart type="family">Manna</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Argentina</namePart>
<namePart type="given">Anna</namePart>
<namePart type="family">Rescigno</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Alina</namePart>
<namePart type="family">Karakanta</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Ayla</namePart>
<namePart type="family">Rigouts Terryn</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Manuel</namePart>
<namePart type="family">Lardelli</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Natalia</namePart>
<namePart type="family">Resende</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Elena</namePart>
<namePart type="family">Murgolo</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Janiça</namePart>
<namePart type="family">Hackenbuchner</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Anna</namePart>
<namePart type="family">Zaretskaya</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Miquel</namePart>
<namePart type="family">Esplà-Gomis</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Thierry</namePart>
<namePart type="family">Etchegoyhen</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Dagmar</namePart>
<namePart type="family">Gromann</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Rachel</namePart>
<namePart type="family">Bawden</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Barry</namePart>
<namePart type="family">Haddow</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Sara</namePart>
<namePart type="family">Szoc</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Mikel</namePart>
<namePart type="family">Forcada</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Helena</namePart>
<namePart type="family">Moniz</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>European Association for Machine Translation</publisher>
<place>
<placeTerm type="text">Tilburg, The Netherlands</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
<identifier type="isbn">9789403901411</identifier>
</relatedItem>
<abstract>EU legislation is equally authentic and legally binding in all 24 official languages, rendering high-quality translation a legal obligation rather than a mere choice. Therefore, high-quality language technology supporting translation processes in all EU languages are essential for language professionals at the European Parliament (EP). This paper investigates whether domain-specific fine-tuning of an open-weight Large Language Model (LLM) yields consistently larger quality gains on legislative text compared to generic text, in all 23 EU target languages from English. We evaluate ten experimental conditions: base model, in-domain and cross-domain fine-tuning, sequential generic-then-legislative fine-tuning, and zero-shot Claude Sonnet 4.6 as a proprietary reference. We analyse BLEU, chrF, TER, and COMET metrics on nearly 700,000 segments. Results confirm the hypothesis for all 23 languages: legislative fine-tuning enhances BLEU by +12.30 compared to +7.10 for generic fine-tuning, demonstrating a consistent advantage of +5.20 BLEU in all the metrics. The fine-tuned EuroLLM-22B decisively outperforms Claude Sonnet 4.6, Anthropic’s latest frontier model, on both domains, highlighting that targeted adaptation of a smaller open-weight model can surpass a state-of-the-art proprietary system. Cross-domain transfer within institutional domain is positive for all languages, with no catastrophic forgetting. Low-resource languages such as Irish and Maltese benefit the most from fine-tuning, while a divergence between BLEU and COMET rankings for some languages underlines the need of evaluation metrics alongside traditional measures.</abstract>
<identifier type="citekey">lorini-etal-2026-one</identifier>
<location>
<url>https://aclanthology.org/2026.eamt-1.20/</url>
</location>
<part>
<date>2026-06</date>
<extent unit="page">
<start>304</start>
<end>320</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T One Size Does Not Fit All: Why EU Legislative Translation Demands Domain-Specific Fine-Tuning of LLMs
%A Lorini, Valerio
%A Vlaic, Paula
%A Akbulut, Ulascan
%A Marcoaldi, Daniele
%Y Shterionov, Dimitar
%Y Vanmassenhove, Eva
%Y De Sisto, Mirella
%Y Blain, Fred
%Y Pourmostafa Roshan Sharami, Javad
%Y Lepp, Lisa
%Y Manna, Chiara
%Y Rescigno, Argentina Anna
%Y Karakanta, Alina
%Y Rigouts Terryn, Ayla
%Y Lardelli, Manuel
%Y Resende, Natalia
%Y Murgolo, Elena
%Y Hackenbuchner, Janiça
%Y Zaretskaya, Anna
%Y Esplà-Gomis, Miquel
%Y Etchegoyhen, Thierry
%Y Gromann, Dagmar
%Y Bawden, Rachel
%Y Haddow, Barry
%Y Szoc, Sara
%Y Forcada, Mikel
%Y Moniz, Helena
%S Proceedings of the 26th Annual Conference of the European Association for Machine Translation (Volume 1)
%D 2026
%8 June
%I European Association for Machine Translation
%C Tilburg, The Netherlands
%@ 9789403901411
%F lorini-etal-2026-one
%X EU legislation is equally authentic and legally binding in all 24 official languages, rendering high-quality translation a legal obligation rather than a mere choice. Therefore, high-quality language technology supporting translation processes in all EU languages are essential for language professionals at the European Parliament (EP). This paper investigates whether domain-specific fine-tuning of an open-weight Large Language Model (LLM) yields consistently larger quality gains on legislative text compared to generic text, in all 23 EU target languages from English. We evaluate ten experimental conditions: base model, in-domain and cross-domain fine-tuning, sequential generic-then-legislative fine-tuning, and zero-shot Claude Sonnet 4.6 as a proprietary reference. We analyse BLEU, chrF, TER, and COMET metrics on nearly 700,000 segments. Results confirm the hypothesis for all 23 languages: legislative fine-tuning enhances BLEU by +12.30 compared to +7.10 for generic fine-tuning, demonstrating a consistent advantage of +5.20 BLEU in all the metrics. The fine-tuned EuroLLM-22B decisively outperforms Claude Sonnet 4.6, Anthropic’s latest frontier model, on both domains, highlighting that targeted adaptation of a smaller open-weight model can surpass a state-of-the-art proprietary system. Cross-domain transfer within institutional domain is positive for all languages, with no catastrophic forgetting. Low-resource languages such as Irish and Maltese benefit the most from fine-tuning, while a divergence between BLEU and COMET rankings for some languages underlines the need of evaluation metrics alongside traditional measures.
%U https://aclanthology.org/2026.eamt-1.20/
%P 304-320
Markdown (Informal)
[One Size Does Not Fit All: Why EU Legislative Translation Demands Domain-Specific Fine-Tuning of LLMs](https://aclanthology.org/2026.eamt-1.20/) (Lorini et al., EAMT 2026)
ACL