@article{toyin-etal-2026-llms,
    title = "Are {LLM}s Good Text Diacritizers? An {A}rabic and {Y}oruba Case Study",
    author = "Toyin, Hawau Olamide  and
      Mohamed Magdy, Samar  and
      Aldarmaki, Hanan",
    editor = "Piperidis, Stelios  and
      Bel, N{\'u}ria  and
      van den Heuvel, Henk  and
      Ide, Nancy  and
      Krek, Simon  and
      Toral, Antonio",
    journal = "International Conference on Language Resources and Evaluation",
    volume = "main",
    month = may,
    year = "2026",
    address = "Palma de Mallorca, Spain",
    publisher = "ELRA Language Resource Association",
    url = "https://preview.aclanthology.org/ingest-lrec/2026.lrec-main.40/",
    pages = "580--589",
    abstract = "We investigate the effectiveness of large language models (LLMs) for text diacritization in two typologically distinct languages: Arabic and Yoruba. To enable a rigorous evaluation, we introduce a novel multilingual dataset MultiDiac, with diverse samples that capture a range of diacritic ambiguities. We evaluate 12 LLMs varying in size, accessibility, and language coverage, and benchmark them against 4 specialized diacritization models. Additionally, we fine-tune four small open-source models using LoRA for Yoruba. Our results show that many off-the-shelf LLMs outperform specialized diacritization models for both Arabic and Yoruba, but smaller models suffer from hallucinations. We find that fine-tuning on a small dataset can help improve diacritization performance and reduce hallucination rates for Yoruba."
}