Metraipolis › archive › dc_3a2a13f2d787

Learn from Whoever Is Right: Answer-Verified Multi-Teacher Distillation for Multi-Domain LLMs — arXiv:2609.02548

Source
arXiv (en)
Original
https://arxiv.org/abs/2609.02548
License (as archived)
CC0-1.0
Archived
2026-10-17T03:38:21.400Z
SHA-256 of the archived text
eea33c1d6b680fca7f7849b5a7237d15011f4c31d05da24c681f4db69a61be34
Stable URL of this page
https://metraipolis.com/documento/dc_3a2a13f2d787

How to cite

Learn from Whoever Is Right: Answer-Verified Multi-Teacher Distillation for Multi-Domain LLMs — arXiv:2609.02548 — arXiv (en), https://arxiv.org/abs/2609.02548 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_3a2a13f2d787 (Metraipolis archive dc_3a2a13f2d787, sha256 eea33c1d6b680fca7f7849b5a7237d15011f4c31d05da24c681f4db69a61be34)
BibTeX
@misc{dc3a2a13f2d787,
  title = {Learn from Whoever Is Right: Answer-Verified Multi-Teacher Distillation for Multi-Domain LLMs — arXiv:2609.02548},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.02548},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_3a2a13f2d787; sha256 eea33c1d6b680fca7f7849b5a7237d15011f4c31d05da24c681f4db69a61be34},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.02548. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Modern large language models (LLMs) rely on reinforcement learning to build strong capabilities in individual domains, but integrating those capabilities into a single deployable model remains challenging. By routing each sample to the teacher whose domain matches it, existing approaches let a domain label decide which teacher provides supervision. However, domain expertise holds only on average: the matched teacher is not always correct on a given sample, while a teacher from another domain sometimes is. The reliable teacher therefore has to be identified per sample, not per domain. In this paper, we introduce Multi-Teacher Self-Distillation Policy Optimization (MT-SDPO), an on-policy distillation method that unifies several frozen teachers into one student model. MT-SDPO consists of three components: (1) self-anchors, where a rollout is supervised by a correct rollout from its own group; (2) answer-verified eligibility, where a teacher may supervise a sample only if its own answer passes a verifier; and (3) privileged distillation, which merges the anchor and all verified feedback into one context that an exponential moving average self-teacher reads and the student does not, thereby keeping one policy at deployment. Across five students from three model families, MT-SDPO lifts the weakest domain of Qwen3-8B by 14.79 points and narrows its domain gap by 74.7%, a better balance than serving one matched teacher per domain. Verified reliability, not domain membership, should decide who teaches. Code is available at this https URL .

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-03 in cs.AI).
Categorie arXiv: Machine Learning (cs.LG) ; Artificial Intelligence (cs.AI).
Pagina del lavoro: https://arxiv.org/abs/2609.02548 — 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 2026-10-17T03:13:12.559Z dalla pagina https://arxiv.org/catchup/cs.AI/2026-09-03?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).