MUG-Eval: A Proxy Evaluation Framework for Multilingual Generation Capabilities in Any Language

Fuente: arXiv
Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Song, Seyoung, Jeong, Seogyeong, Kim, Eunsu, Jin, Jiho, Kim, Dongkwan, Shin, Jay, Oh, Alice
Format: Preprint
Veröffentlicht: 2025
Schlagworte:
Online-Zugang:
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866908639603720192
author Song, Seyoung
Jeong, Seogyeong
Kim, Eunsu
Jin, Jiho
Kim, Dongkwan
Shin, Jay
Oh, Alice
author_facet Song, Seyoung
Jeong, Seogyeong
Kim, Eunsu
Jin, Jiho
Kim, Dongkwan
Shin, Jay
Oh, Alice
contents Evaluating text generation capabilities of large language models (LLMs) is challenging, particularly for low-resource languages where methods for direct assessment are scarce. We propose MUG-Eval, a novel framework that evaluates LLMs' multilingual generation capabilities by transforming existing benchmarks into conversational tasks and measuring the LLMs' accuracies on those tasks. We specifically designed these conversational tasks to require effective communication in the target language. Then, we simply use task success rate as a proxy for successful conversation generation. Our approach offers two key advantages: it is independent of language-specific NLP tools or annotated datasets, which are limited for most languages, and it does not rely on LLMs-as-judges, whose evaluation quality degrades outside a few high-resource languages. We evaluate 8 LLMs across 30 languages spanning high, mid, and low-resource categories, and we find that MUG-Eval correlates strongly with established benchmarks ($r$ > 0.75) while enabling standardized comparisons across languages and models. Our framework provides a robust and resource-efficient solution for evaluating multilingual generation that can be extended to thousands of languages.
format Preprint
id arxiv_https___arxiv_org_abs_2505_14395
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle MUG-Eval: A Proxy Evaluation Framework for Multilingual Generation Capabilities in Any Language
Song, Seyoung
Jeong, Seogyeong
Kim, Eunsu
Jin, Jiho
Kim, Dongkwan
Shin, Jay
Oh, Alice
Computation and Language
Artificial Intelligence
Evaluating text generation capabilities of large language models (LLMs) is challenging, particularly for low-resource languages where methods for direct assessment are scarce. We propose MUG-Eval, a novel framework that evaluates LLMs' multilingual generation capabilities by transforming existing benchmarks into conversational tasks and measuring the LLMs' accuracies on those tasks. We specifically designed these conversational tasks to require effective communication in the target language. Then, we simply use task success rate as a proxy for successful conversation generation. Our approach offers two key advantages: it is independent of language-specific NLP tools or annotated datasets, which are limited for most languages, and it does not rely on LLMs-as-judges, whose evaluation quality degrades outside a few high-resource languages. We evaluate 8 LLMs across 30 languages spanning high, mid, and low-resource categories, and we find that MUG-Eval correlates strongly with established benchmarks ($r$ > 0.75) while enabling standardized comparisons across languages and models. Our framework provides a robust and resource-efficient solution for evaluating multilingual generation that can be extended to thousands of languages.
title MUG-Eval: A Proxy Evaluation Framework for Multilingual Generation Capabilities in Any Language
topic Computation and Language
Artificial Intelligence
url https://arxiv.org/abs/2505.14395