Metraipolis › archive › dc_ebe5984f785e

Assessing the Impact of Language Disparity on Multilingual Linguistic Ability in Large Language Models — arXiv:2610.00540

Source
arXiv (en)
Original
https://arxiv.org/abs/2610.00540
License (as archived)
CC0-1.0
Archived
2026-10-06T13:54:01.402Z
SHA-256 of the archived text
bf064ba1fc8ed450a16d43651513c46a5daa452ae32cd6446f1818b4f4f5bd5e
Stable URL of this page
https://metraipolis.com/documento/dc_ebe5984f785e

How to cite

Assessing the Impact of Language Disparity on Multilingual Linguistic Ability in Large Language Models — arXiv:2610.00540 — arXiv (en), https://arxiv.org/abs/2610.00540 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_ebe5984f785e (Metraipolis archive dc_ebe5984f785e, sha256 bf064ba1fc8ed450a16d43651513c46a5daa452ae32cd6446f1818b4f4f5bd5e)
BibTeX
@misc{dcebe5984f785e,
  title = {Assessing the Impact of Language Disparity on Multilingual Linguistic Ability in Large Language Models — arXiv:2610.00540},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2610.00540},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_ebe5984f785e; sha256 bf064ba1fc8ed450a16d43651513c46a5daa452ae32cd6446f1818b4f4f5bd5e},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.00540. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Claims about the grammatical competence of multilingual language models vary sharply with how competence is measured, yet the interaction between evaluation paradigm, post-training, and language resource availability has not been systematically examined. We evaluate base and post-trained models from six families on MultiBLiMP, a syntactic minimal-pair benchmark covering 101 languages, using four evaluation methods. We report three principal findings. First, post-training degrades grammatical competence, but the magnitude of this effect is reduced unevenly by model scale, while low-resource languages bear the highest cost. Second, post-trained models retain grammatical knowledge they cannot articulate through explicit prompting, yet this is measurable only in high-resource languages, because near-chance baselines in low-resource settings leave little knowledge to hide. Third, native-language prompting recovers otherwise hidden competence on low-resource languages, demonstrating that only high-resource languages can be probed directly from unprompted probabilities. We conclude that multilingual grammatical evaluation must adopt language-informed, multi-paradigm protocols to avoid systematically underestimating low-resource abilities.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-10-02 in cs.CL).
Categorie arXiv: Computation and Language (cs.CL).
Pagina del lavoro: https://arxiv.org/abs/2610.00540 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito).
Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina.
Letto il 2026-10-06T12:15:34.623Z dalla pagina https://arxiv.org/catchup/cs.CL/2026-10-02?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).