Benchmarking Multilingual LLM Translation Accuracy for Fuzhounese

Sue Zheng, Jelke Bloem


Abstract
Multilingual large language models are known to perform very well on high-resource languages, while their ability to process severely under-resourced languages remains underexplored. We investigate multilingual LLM translation performance on Fuzhounese, an under-resourced Sinitic language without a standardized orthography and almost no digital presence. Having adopted some methodological insights from the HKCanto-Eval benchmark, this paper presents a bidirectional translation framework based on a dataset of 305 sentences (300 constructed English sentences and 5 additional reference translations), that assesses the comprehension and generation of Fuzhounese, evaluated using automatic metrics and human Likert-scale judgments. The results reveal poor performance on Fuzhounese in both translation directions: BERTScore and chrF++ values consistently stay low when models are faced with comprehension tasks, while for generation tasks, scores are generally more than twofold lower than those for Mandarin or Cantonese. These findings highlight structural biases in multilingual LLMs toward high-resource languages and stress the need for resource-aware modeling and evaluation approaches in multilingual NLP systems.
Anthology ID:
2026.sigul-1.20
Volume:
Proceedings of the SIGUL 2026 Joint Workshop with ELE, EURALI, and DCLRL: Towards Inclusivity and Equality: Language Resources and Technologies for Under-Resourced and Endangered Languages
Month:
May
Year:
2026
Address:
Palma, Mallorca, Spain
Editors:
Atul Kr. Ojha, Sakriani Sakti, Claudia Soria, Maite Melero, John P. McCrae, Constantine Lignos, Chao-Hong Liu, German Rigau Claramunt, Georg Rehm
Venues:
SIGUL | EURALI | DCLRL | WS
SIG:
Publisher:
ELRA Language Resources Association (ELRA)
Note:
Pages:
198–209
Language:
External URL:
https://lrec.elra.info/lrec2026-ws-sigul-20
DOI:
10.63317/4mm9bs8yy4ie
Bibkey:
Cite (ACL):
Sue Zheng and Jelke Bloem. 2026. Benchmarking Multilingual LLM Translation Accuracy for Fuzhounese. In Proceedings of the SIGUL 2026 Joint Workshop with ELE, EURALI, and DCLRL: Towards Inclusivity and Equality: Language Resources and Technologies for Under-Resourced and Endangered Languages, pages 198–209, Palma, Mallorca, Spain. ELRA Language Resources Association (ELRA).
Cite (Informal):
Benchmarking Multilingual LLM Translation Accuracy for Fuzhounese (Zheng & Bloem, SIGUL-EURALI-DCLRL 2026)
Copy Citation: