Metraipolis › archive › dc_32e94dbfc8fd

Multi-SFT: Permissively-Licensed Multilingual Instruction-Following — arXiv:2610.22086

Source
arXiv (en)
Original
https://arxiv.org/abs/2610.22086
License (as archived)
CC0-1.0
Archived
2026-10-16T09:30:56.440Z
SHA-256 of the archived text
4854b32c5f76d650ed99d7473399cad50ea9388dd2c76217405ff90f4ef78193
Stable URL of this page
https://metraipolis.com/documento/dc_32e94dbfc8fd

How to cite

Multi-SFT: Permissively-Licensed Multilingual Instruction-Following — arXiv:2610.22086 — arXiv (en), https://arxiv.org/abs/2610.22086 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_32e94dbfc8fd (Metraipolis archive dc_32e94dbfc8fd, sha256 4854b32c5f76d650ed99d7473399cad50ea9388dd2c76217405ff90f4ef78193)
BibTeX
@misc{dc32e94dbfc8fd,
  title = {Multi-SFT: Permissively-Licensed Multilingual Instruction-Following — arXiv:2610.22086},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2610.22086},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_32e94dbfc8fd; sha256 4854b32c5f76d650ed99d7473399cad50ea9388dd2c76217405ff90f4ef78193},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.22086. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Supervised fine-tuning datasets remain overwhelmingly English-centric and commercially restricted, as their content is frequently generated by proprietary models whose terms of service forbid training competing systems. We introduce Multi-SFT, a commercially permissive collection of 502,818 instruction-response pairs across five European languages, built via translation and regeneration pipelines for tasks where translation is insufficient. Human evaluation yields inter-annotator agreement of at least 0.87 across all datasets and languages. Fine-tuning nine models from three different families (EuroLLM, OLMo 2, Qwen3; 1B-32B) and evaluating them on 13 benchmarks shows that Multi-SFT alone is competitive with larger augmented mixes and official instruction-tuned releases, with decisive advantages on safety tasks.

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/2610.22086 — 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 2026-10-16T08:44:18.025Z dalla pagina https://arxiv.org/list/cs.AI/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).