@inproceedings{paniv-etal-2026-data,
    title = "Data-Efficient Adaptation of Multilingual {LLM}s to {U}krainian",
    author = "Paniv, Yurii  and
      Didenko, Bohdan  and
      Haltiuk, Mykola  and
      Humennyy, Vladyslav  and
      Kravchenko, Andrian  and
      Kyslyi, Roman  and
      Makovska, Viktoriia  and
      Orlovskyi, Artem  and
      Ruban, Bohdan  and
      Rudko, Maksym-Yurii  and
      Senyk, Anastasiia  and
      Drushchak, Nazarii  and
      Chaplynskyi, Dmytro  and
      Romanyshyn, Mariana",
    editor = "Romanyshyn, Mariana",
    booktitle = "Proceedings of the Fifth {U}krainian Natural Language Processing Conference ({UNLP} 2026)",
    month = may,
    year = "2026",
    address = "Lviv, Ukraine",
    publisher = "Association for Computational Linguistics",
    url = "https://preview.aclanthology.org/bulk-corrections-2026-07-02/2026.unlp-1.14/",
    pages = "155--168",
    ISBN = "979-8-89176-359-3",
    abstract = "Adapting large language models to low-resource languages presents three interconnected challenges: inefficient tokenization, scarcity of high-quality annotated data, and limited resources for instruction tuning. We present a reproducible approach that addresses each challenge using data-centric methods that primarily rely on unlabeled text corpora, parallel translation data, and a multilingual base model. Our approach combines (1) vocabulary surgery for tokenizer adaptation without full retraining, (2) cross-lingual transfer of quality classifiers via translation, enabling filtering without target-language annotations, and (3) generation of instruction data through translation, task conversion, and targeted synthesis. We validate this recipe by adapting Gemma-3-12B to Ukrainian. {\%}, producing Lapa-12BOur pretrained model achieves top performance on Ukrainian benchmarks, while our instruction-tuned variant demonstrates strong performance on translation (33 BLEU on FLORES), summarization, and question-answering tasks, while requiring 1.5x fewer tokens than the original model for the same text. We release all models, datasets, classifiers, and code to enable replication for other languages."
}