Benchmarking Large Language Models for Lemmatization and Translation of Finnic Runosongs

Lidia Pivovarova, Kati Kallio, Antti Kanner, Jakob Lindström, Eetu Mäkelä, Liina Saarlo, Kaarel Veskis, Mari Väina


Abstract
We investigate the use of large language models (LLMs) for translation and annotation of Finnic runosongs—a highly variable multilingual poetic corpus with limited linguistic or NLP resources. We manually annotated a corpus of about 200 runosongs in a variety of languages, dialects and genres with lemmas and English translations. Using this manually annotated test set, we benchmark several large language models. We tested several prompt types and developed a collective prompt-writing methodology involving specialists from different backgrounds. Our results highlight both the potential and the limitations of current LLMs for cultural heritage NLP, and point towards strategies for prompt design, evaluation, and integration with linguistic expertise.
Anthology ID:
2025.iwclul-1.12
Volume:
Proceedings of the 10th International Workshop on Computational Linguistics for Uralic Languages
Month:
December
Year:
2025
Address:
Joensuu, Finland
Editors:
Mika Hämäläinen, Michael Rießler, Eiaki V. Morooka, Lev Kharlashkin
Venues:
IWCLUL | WS
SIG:
SIGUR
Publisher:
Association for Computational Linguistics
Note:
Pages:
87–105
Language:
URL:
https://aclanthology.org/2025.iwclul-1.12/
DOI:
Bibkey:
Cite (ACL):
Lidia Pivovarova, Kati Kallio, Antti Kanner, Jakob Lindström, Eetu Mäkelä, Liina Saarlo, Kaarel Veskis, and Mari Väina. 2025. Benchmarking Large Language Models for Lemmatization and Translation of Finnic Runosongs. In Proceedings of the 10th International Workshop on Computational Linguistics for Uralic Languages, pages 87–105, Joensuu, Finland. Association for Computational Linguistics.
Cite (Informal):
Benchmarking Large Language Models for Lemmatization and Translation of Finnic Runosongs (Pivovarova et al., IWCLUL 2025)
Copy Citation:
PDF:
https://aclanthology.org/2025.iwclul-1.12.pdf