Metraipolis › archive › dc_4269bfebaa9c
EuroAlpaca: Task-Preserving Localisation of Instruction Data for European Languages — arXiv:2609.05043
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2609.05043
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-16T22:35:24.218Z
- SHA-256 of the archived text
e9691a0a2229a9585fe54b93e9a27e367080bf667e690f3e7dddc59dce623a25- Stable URL of this page
https://metraipolis.com/documento/dc_4269bfebaa9c
How to cite
EuroAlpaca: Task-Preserving Localisation of Instruction Data for European Languages — arXiv:2609.05043 — arXiv (en), https://arxiv.org/abs/2609.05043 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_4269bfebaa9c (Metraipolis archive dc_4269bfebaa9c, sha256 e9691a0a2229a9585fe54b93e9a27e367080bf667e690f3e7dddc59dce623a25)
BibTeX
@misc{dc4269bfebaa9c,
title = {EuroAlpaca: Task-Preserving Localisation of Instruction Data for European Languages — arXiv:2609.05043},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2609.05043},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_4269bfebaa9c; sha256 e9691a0a2229a9585fe54b93e9a27e367080bf667e690f3e7dddc59dce623a25},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.05043. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Machine translation (MT) offers a scalable way to extend English instruction-tuning data to multiple languages, but it can distort task-critical constraints and required outputs, creating corrupted training examples and degrading models trained on such data. We introduce EuroAlpaca, a task-preserving localisation pipeline and near-parallel resource covering 50 European languages and regional varieties, together with European-IFEval, a multilingual benchmark for verifiable instruction following. Depending on the example, our pipeline applies field-wise MT while preserving task-critical content or reconstructs a task-equivalent target-language instance, followed by validation of cross-field coherence and target-language consistency. Across LoRA experiments with four LLMs, training on directly translated data improves ROUGE-L and F-BERT on the Aya Evaluation Suite, but reduces accuracy on European-IFEval by 29.8% relative to the unadapted baseline. In contrast, adaptation with EuroAlpaca improves accuracy by 12.9% over the same baseline, reversing the degradation caused by direct MT, while also achieving the highest ROUGE-L and F-BERT scores on Aya. These results show that preserving task semantics is essential for multilingual instruction tuning. Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-07 in cs.CL). Categorie arXiv: Computation and Language (cs.CL). Pagina del lavoro: https://arxiv.org/abs/2609.05043 — 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 2026-10-16T22:21:16.292Z dalla pagina https://arxiv.org/catchup/cs.CL/2026-09-07?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).