Irish-BLiMP: A Linguistic Benchmark for Evaluating Human and Language Model Performance in a Low-Resource Setting
Fuente:
arXiv
Enregistré dans:
| Auteurs principaux: | , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Publié: |
2025
|
| Sujets: | |
| Accès en ligne: | |
| Tags: |
Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
|
| _version_ | 1866918169108545536 |
|---|---|
| author | McGiff, Josh Tran, Khanh-Tung Mulcahy, William Luinín, Dáibhidh Ó Dalzell, Jake Bhroin, Róisín Ní Burke, Adam O'Sullivan, Barry Nguyen, Hoang D. Nikolov, Nikola S. |
| author_facet | McGiff, Josh Tran, Khanh-Tung Mulcahy, William Luinín, Dáibhidh Ó Dalzell, Jake Bhroin, Róisín Ní Burke, Adam O'Sullivan, Barry Nguyen, Hoang D. Nikolov, Nikola S. |
| contents | We present Irish-BLiMP (Irish Benchmark of Linguistic Minimal Pairs), the first dataset and framework designed for fine-grained evaluation of linguistic competence in the Irish language, an endangered language. Drawing on a variety of linguistic literature and grammar reference works, we manually constructed and reviewed 1020 minimal pairs across a taxonomy of 11 linguistic features, through a team of fluent Irish speakers. We evaluate both existing Large Language Models (LLMs) and fluent human participants on their syntactic knowledge of Irish. Our findings show that humans outperform all models across all linguistic features, achieving 16.6% higher accuracy on average. Moreover, a substantial performance gap of 18.1% persists between open- and closed-source LLMs, with even the strongest model (gpt-5) reaching only 73.5% accuracy compared to 90.1% by human. Interestingly, human participants and models struggle on different aspects of Irish grammar, thus highlighting a difference in representation learned by the models. Overall, Irish-BLiMP provides the first systematic framework for evaluating the grammatical competence of LLMs in Irish and offers a valuable benchmark for advancing research on linguistic understanding in low-resource languages. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2510_20957 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Irish-BLiMP: A Linguistic Benchmark for Evaluating Human and Language Model Performance in a Low-Resource Setting McGiff, Josh Tran, Khanh-Tung Mulcahy, William Luinín, Dáibhidh Ó Dalzell, Jake Bhroin, Róisín Ní Burke, Adam O'Sullivan, Barry Nguyen, Hoang D. Nikolov, Nikola S. Computation and Language We present Irish-BLiMP (Irish Benchmark of Linguistic Minimal Pairs), the first dataset and framework designed for fine-grained evaluation of linguistic competence in the Irish language, an endangered language. Drawing on a variety of linguistic literature and grammar reference works, we manually constructed and reviewed 1020 minimal pairs across a taxonomy of 11 linguistic features, through a team of fluent Irish speakers. We evaluate both existing Large Language Models (LLMs) and fluent human participants on their syntactic knowledge of Irish. Our findings show that humans outperform all models across all linguistic features, achieving 16.6% higher accuracy on average. Moreover, a substantial performance gap of 18.1% persists between open- and closed-source LLMs, with even the strongest model (gpt-5) reaching only 73.5% accuracy compared to 90.1% by human. Interestingly, human participants and models struggle on different aspects of Irish grammar, thus highlighting a difference in representation learned by the models. Overall, Irish-BLiMP provides the first systematic framework for evaluating the grammatical competence of LLMs in Irish and offers a valuable benchmark for advancing research on linguistic understanding in low-resource languages. |
| title | Irish-BLiMP: A Linguistic Benchmark for Evaluating Human and Language Model Performance in a Low-Resource Setting |
| topic | Computation and Language |
| url | https://arxiv.org/abs/2510.20957 |