Metraipolis › archive › dc_ccc02117c623

On-Policy Delta Distillation for Multilingual Math Reasoning — arXiv:2608.05802

Source
arXiv (en)
Original
https://arxiv.org/abs/2608.05802
License (as archived)
CC0-1.0
Archived
2026-10-25T21:27:13.739Z
SHA-256 of the archived text
4c1d03991984a75d8f02e7647debee71d2309ebc90ef5a051282e67bf7e19863
Stable URL of this page
https://metraipolis.com/documento/dc_ccc02117c623

How to cite

On-Policy Delta Distillation for Multilingual Math Reasoning — arXiv:2608.05802 — arXiv (en), https://arxiv.org/abs/2608.05802 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_ccc02117c623 (Metraipolis archive dc_ccc02117c623, sha256 4c1d03991984a75d8f02e7647debee71d2309ebc90ef5a051282e67bf7e19863)
BibTeX
@misc{dcccc02117c623,
  title = {On-Policy Delta Distillation for Multilingual Math Reasoning — arXiv:2608.05802},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2608.05802},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_ccc02117c623; sha256 4c1d03991984a75d8f02e7647debee71d2309ebc90ef5a051282e67bf7e19863},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2608.05802. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

On-Policy Distillation (OPD) is emerging as a promising alternative to reinforcement learning for LLM post-training, yet its effectiveness in multilingual settings remains underexplored. We study OPD and its advanced variant, On-Policy Delta Distillation (OPD$^2$), for mathematical reasoning in English, Korean, and Japanese. OPD$^2$ improves OPD by using the probability gap between a post-trained teacher and its base model as the learning signal. Experiments with Qwen3 show that OPD$^2$ consistently outperforms the original OPD, with particularly strong improvements in Korean and Japanese, and generally narrows the English-Korean performance gap. We further find that English-only OPD can also increase performance for Korean and Japanese, but often shifts the responses toward English, highlighting the importance of multilingual data to preserving target-language responses.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-08-07 in cs.CL).
Categorie arXiv: Computation and Language (cs.CL) ; Machine Learning (cs.LG).
Pagina del lavoro: https://arxiv.org/abs/2608.05802 — 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 2026-10-25T21:24:38.206Z dalla pagina https://arxiv.org/catchup/cs.CL/2026-08-07?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).