@inproceedings{anschutz-etal-2025-dis,
title = "(Dis)improved?! How Simplified Language Affects Large Language Model Performance across Languages",
author = {Ansch{\"u}tz, Miriam and
Damaratskaya, Anastasiya and
Lee, Chaeeun Joy and
Schmalz, Arthur and
Mosca, Edoardo and
Groh, Georg},
editor = "Arviv, Ofir and
Clinciu, Miruna and
Dhole, Kaustubh and
Dror, Rotem and
Gehrmann, Sebastian and
Habba, Eliya and
Itzhak, Itay and
Mille, Simon and
Perlitz, Yotam and
Santus, Enrico and
Sedoc, Jo{\~a}o and
Shmueli Scheuer, Michal and
Stanovsky, Gabriel and
Tafjord, Oyvind",
booktitle = "Proceedings of the Fourth Workshop on Generation, Evaluation and Metrics (GEM{\texttwosuperior})",
month = jul,
year = "2025",
address = "Vienna, Austria and virtual meeting",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2025.gem-1.70/",
pages = "847--861",
ISBN = "979-8-89176-261-9",
abstract = "Simplified language enhances the accessibility and human understanding of texts. However, whether it also benefits large language models (LLMs) remains underexplored. This paper extensively studies whether LLM performance improves on simplified data compared to its original counterpart. Our experiments span six datasets and nine automatic simplification systems across three languages. We show that English models, including GPT-4o-mini, show a weak generalization and exhibit a significant performance drop on simplified data. This introduces an intriguing paradox: simplified data is helpful for humans but not for LLMs. At the same time, the performance in non-English languages sometimes improves, depending on the task and quality of the simplifier. Our findings offer a comprehensive view of the impact of simplified language on LLM performance and uncover severe implications for people depending on simple language."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="anschutz-etal-2025-dis">
<titleInfo>
<title>(Dis)improved?! How Simplified Language Affects Large Language Model Performance across Languages</title>
</titleInfo>
<name type="personal">
<namePart type="given">Miriam</namePart>
<namePart type="family">Anschütz</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Anastasiya</namePart>
<namePart type="family">Damaratskaya</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Chaeeun</namePart>
<namePart type="given">Joy</namePart>
<namePart type="family">Lee</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Arthur</namePart>
<namePart type="family">Schmalz</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Edoardo</namePart>
<namePart type="family">Mosca</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Georg</namePart>
<namePart type="family">Groh</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2025-07</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of the Fourth Workshop on Generation, Evaluation and Metrics (GEM²)</title>
</titleInfo>
<name type="personal">
<namePart type="given">Ofir</namePart>
<namePart type="family">Arviv</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Miruna</namePart>
<namePart type="family">Clinciu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Kaustubh</namePart>
<namePart type="family">Dhole</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Rotem</namePart>
<namePart type="family">Dror</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Sebastian</namePart>
<namePart type="family">Gehrmann</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Eliya</namePart>
<namePart type="family">Habba</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Itay</namePart>
<namePart type="family">Itzhak</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Simon</namePart>
<namePart type="family">Mille</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yotam</namePart>
<namePart type="family">Perlitz</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Enrico</namePart>
<namePart type="family">Santus</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">João</namePart>
<namePart type="family">Sedoc</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Michal</namePart>
<namePart type="family">Shmueli Scheuer</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Gabriel</namePart>
<namePart type="family">Stanovsky</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Oyvind</namePart>
<namePart type="family">Tafjord</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Vienna, Austria and virtual meeting</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
<identifier type="isbn">979-8-89176-261-9</identifier>
</relatedItem>
<abstract>Simplified language enhances the accessibility and human understanding of texts. However, whether it also benefits large language models (LLMs) remains underexplored. This paper extensively studies whether LLM performance improves on simplified data compared to its original counterpart. Our experiments span six datasets and nine automatic simplification systems across three languages. We show that English models, including GPT-4o-mini, show a weak generalization and exhibit a significant performance drop on simplified data. This introduces an intriguing paradox: simplified data is helpful for humans but not for LLMs. At the same time, the performance in non-English languages sometimes improves, depending on the task and quality of the simplifier. Our findings offer a comprehensive view of the impact of simplified language on LLM performance and uncover severe implications for people depending on simple language.</abstract>
<identifier type="citekey">anschutz-etal-2025-dis</identifier>
<location>
<url>https://aclanthology.org/2025.gem-1.70/</url>
</location>
<part>
<date>2025-07</date>
<extent unit="page">
<start>847</start>
<end>861</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T (Dis)improved?! How Simplified Language Affects Large Language Model Performance across Languages
%A Anschütz, Miriam
%A Damaratskaya, Anastasiya
%A Lee, Chaeeun Joy
%A Schmalz, Arthur
%A Mosca, Edoardo
%A Groh, Georg
%Y Arviv, Ofir
%Y Clinciu, Miruna
%Y Dhole, Kaustubh
%Y Dror, Rotem
%Y Gehrmann, Sebastian
%Y Habba, Eliya
%Y Itzhak, Itay
%Y Mille, Simon
%Y Perlitz, Yotam
%Y Santus, Enrico
%Y Sedoc, João
%Y Shmueli Scheuer, Michal
%Y Stanovsky, Gabriel
%Y Tafjord, Oyvind
%S Proceedings of the Fourth Workshop on Generation, Evaluation and Metrics (GEM²)
%D 2025
%8 July
%I Association for Computational Linguistics
%C Vienna, Austria and virtual meeting
%@ 979-8-89176-261-9
%F anschutz-etal-2025-dis
%X Simplified language enhances the accessibility and human understanding of texts. However, whether it also benefits large language models (LLMs) remains underexplored. This paper extensively studies whether LLM performance improves on simplified data compared to its original counterpart. Our experiments span six datasets and nine automatic simplification systems across three languages. We show that English models, including GPT-4o-mini, show a weak generalization and exhibit a significant performance drop on simplified data. This introduces an intriguing paradox: simplified data is helpful for humans but not for LLMs. At the same time, the performance in non-English languages sometimes improves, depending on the task and quality of the simplifier. Our findings offer a comprehensive view of the impact of simplified language on LLM performance and uncover severe implications for people depending on simple language.
%U https://aclanthology.org/2025.gem-1.70/
%P 847-861
Markdown (Informal)
[(Dis)improved?! How Simplified Language Affects Large Language Model Performance across Languages](https://aclanthology.org/2025.gem-1.70/) (Anschütz et al., GEM 2025)
ACL