@inproceedings{huang-etal-2024-lost,
title = "Lost in the Source Language: How Large Language Models Evaluate the Quality of Machine Translation",
author = "Huang, Xu and
Zhang, Zhirui and
Geng, Xiang and
Du, Yichao and
Chen, Jiajun and
Huang, Shujian",
editor = "Ku, Lun-Wei and
Martins, Andre and
Srikumar, Vivek",
booktitle = "Findings of the Association for Computational Linguistics: ACL 2024",
month = aug,
year = "2024",
address = "Bangkok, Thailand",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2024.findings-acl.211",
doi = "10.18653/v1/2024.findings-acl.211",
pages = "3546--3562",
abstract = "This study investigates how Large Language Models (LLMs) leverage source and reference data in machine translation evaluation task, aiming to better understand the mechanisms behind their remarkable performance in this task.We design the controlled experiments across various input modes and model types, and employ both coarse-grained and fine-grained prompts to discern the utility of source versus reference information.We find that reference information significantly enhances the evaluation accuracy, while surprisingly, source information sometimes is counterproductive, indicating LLMs{'} inability to fully leverage the cross-lingual capability when evaluating translations.Further analysis of the fine-grained evaluation and fine-tuning experiments show similar results.These findings also suggest a potential research direction for LLMs that fully exploits the cross-lingual capability of LLMs to achieve better performance in machine translation evaluation tasks.",
}
<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="huang-etal-2024-lost">
<titleInfo>
<title>Lost in the Source Language: How Large Language Models Evaluate the Quality of Machine Translation</title>
</titleInfo>
<name type="personal">
<namePart type="given">Xu</namePart>
<namePart type="family">Huang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Zhirui</namePart>
<namePart type="family">Zhang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Xiang</namePart>
<namePart type="family">Geng</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yichao</namePart>
<namePart type="family">Du</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jiajun</namePart>
<namePart type="family">Chen</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Shujian</namePart>
<namePart type="family">Huang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2024-08</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Findings of the Association for Computational Linguistics: ACL 2024</title>
</titleInfo>
<name type="personal">
<namePart type="given">Lun-Wei</namePart>
<namePart type="family">Ku</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Andre</namePart>
<namePart type="family">Martins</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Vivek</namePart>
<namePart type="family">Srikumar</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Bangkok, Thailand</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
</relatedItem>
<abstract>This study investigates how Large Language Models (LLMs) leverage source and reference data in machine translation evaluation task, aiming to better understand the mechanisms behind their remarkable performance in this task.We design the controlled experiments across various input modes and model types, and employ both coarse-grained and fine-grained prompts to discern the utility of source versus reference information.We find that reference information significantly enhances the evaluation accuracy, while surprisingly, source information sometimes is counterproductive, indicating LLMs’ inability to fully leverage the cross-lingual capability when evaluating translations.Further analysis of the fine-grained evaluation and fine-tuning experiments show similar results.These findings also suggest a potential research direction for LLMs that fully exploits the cross-lingual capability of LLMs to achieve better performance in machine translation evaluation tasks.</abstract>
<identifier type="citekey">huang-etal-2024-lost</identifier>
<identifier type="doi">10.18653/v1/2024.findings-acl.211</identifier>
<location>
<url>https://aclanthology.org/2024.findings-acl.211</url>
</location>
<part>
<date>2024-08</date>
<extent unit="page">
<start>3546</start>
<end>3562</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Lost in the Source Language: How Large Language Models Evaluate the Quality of Machine Translation
%A Huang, Xu
%A Zhang, Zhirui
%A Geng, Xiang
%A Du, Yichao
%A Chen, Jiajun
%A Huang, Shujian
%Y Ku, Lun-Wei
%Y Martins, Andre
%Y Srikumar, Vivek
%S Findings of the Association for Computational Linguistics: ACL 2024
%D 2024
%8 August
%I Association for Computational Linguistics
%C Bangkok, Thailand
%F huang-etal-2024-lost
%X This study investigates how Large Language Models (LLMs) leverage source and reference data in machine translation evaluation task, aiming to better understand the mechanisms behind their remarkable performance in this task.We design the controlled experiments across various input modes and model types, and employ both coarse-grained and fine-grained prompts to discern the utility of source versus reference information.We find that reference information significantly enhances the evaluation accuracy, while surprisingly, source information sometimes is counterproductive, indicating LLMs’ inability to fully leverage the cross-lingual capability when evaluating translations.Further analysis of the fine-grained evaluation and fine-tuning experiments show similar results.These findings also suggest a potential research direction for LLMs that fully exploits the cross-lingual capability of LLMs to achieve better performance in machine translation evaluation tasks.
%R 10.18653/v1/2024.findings-acl.211
%U https://aclanthology.org/2024.findings-acl.211
%U https://doi.org/10.18653/v1/2024.findings-acl.211
%P 3546-3562
Markdown (Informal)
[Lost in the Source Language: How Large Language Models Evaluate the Quality of Machine Translation](https://aclanthology.org/2024.findings-acl.211) (Huang et al., Findings 2024)
ACL