Metraipolis › archive › dc_019c4e72521c
Scaling Multilingual Safety and Reasoning Evaluation through Reasoning-Guided Translation — arXiv:2610.20221
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2610.20221
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-15T20:02:58.239Z
- SHA-256 of the archived text
37b20fa2c59bb665d92c9bbccba78f86be7bec5236832fac97dd656ec5a2d79a- Stable URL of this page
https://metraipolis.com/documento/dc_019c4e72521c
How to cite
Scaling Multilingual Safety and Reasoning Evaluation through Reasoning-Guided Translation — arXiv:2610.20221 — arXiv (en), https://arxiv.org/abs/2610.20221 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_019c4e72521c (Metraipolis archive dc_019c4e72521c, sha256 37b20fa2c59bb665d92c9bbccba78f86be7bec5236832fac97dd656ec5a2d79a)
BibTeX
@misc{dc019c4e72521c,
title = {Scaling Multilingual Safety and Reasoning Evaluation through Reasoning-Guided Translation — arXiv:2610.20221},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2610.20221},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_019c4e72521c; sha256 37b20fa2c59bb665d92c9bbccba78f86be7bec5236832fac97dd656ec5a2d79a},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.20221. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Large language models (LLMs) are increasingly deployed across languages, yet their safety and reasoning capabilities remain predominantly evaluated in English. Translating existing benchmarks offers a scalable approach to multilingual evaluation, but preserving task-specific meaning across languages presents significant challenges, particularly for harmful requests, mathematical reasoning, and conversational interactions. We introduce a reasoning-guided translation framework for constructing multilingual evaluations designed to preserve the structure and intended meaning of diverse benchmark tasks. Our approach separates translation strategy selection from large-scale generation, combining reasoning-guided configuration selection with task-aware validation, targeted repair, and independent quality screening. Using this framework, we develop a source-aligned evaluation collection spanning seven safety, conversational, and mathematical reasoning resources across 11 languages. The resulting collection achieves 99.94\% coverage of intended item--language pairs in the two subsets requiring repair. Evaluations of open-source LLMs reveal distinct language-dependent vulnerabilities, including inconsistent refusal behaviour and substantial reasoning degradation in lower-resource languages. These findings highlight the importance of task-sensitive benchmark construction and systematic quality assessment for understanding multilingual model behaviour beyond English. Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.AI, nuovi lavori del giorno). Categorie arXiv: Computation and Language (cs.CL) ; Artificial Intelligence (cs.AI). Pagina del lavoro: https://arxiv.org/abs/2610.20221 — 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 2026-10-15T19:35:48.410Z dalla pagina https://arxiv.org/list/cs.AI/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).