Metraipolis › archive › dc_41dfce4a5e08
Eliciting Weak-to-Strong Generalization with On-Policy Reverse Distillation — arXiv:2609.08798
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2609.08798
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-16T12:58:21.145Z
- SHA-256 of the archived text
0472aa647d330c3a71d8ad9eab1da7cb673740114f8eb6a986258e7055e26aee- Stable URL of this page
https://metraipolis.com/documento/dc_41dfce4a5e08
How to cite
Eliciting Weak-to-Strong Generalization with On-Policy Reverse Distillation — arXiv:2609.08798 — arXiv (en), https://arxiv.org/abs/2609.08798 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_41dfce4a5e08 (Metraipolis archive dc_41dfce4a5e08, sha256 0472aa647d330c3a71d8ad9eab1da7cb673740114f8eb6a986258e7055e26aee)
BibTeX
@misc{dc41dfce4a5e08,
title = {Eliciting Weak-to-Strong Generalization with On-Policy Reverse Distillation — arXiv:2609.08798},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2609.08798},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_41dfce4a5e08; sha256 0472aa647d330c3a71d8ad9eab1da7cb673740114f8eb6a986258e7055e26aee},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.08798. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Weak-to-strong generalization asks whether stronger models can learn from weaker supervisors and surpass them. This question is particularly important for successive model generations and multi-domain consolidation, where repeating frontier-scale post-training from scratch can be prohibitively expensive. Yet conventional distillation treats the weak teacher as an optimization target, potentially imposing its capacity ceiling on the student. We introduce On-Policy Reverse Distillation (OPRD), which evaluates the teacher's policy shift relative to its reference policy on student rollouts and amplifies the component of the student's verifier-driven policy gradient along that direction. By rescaling only verifier-supported updates, OPRD preserves the stationary points of policy optimization while accelerating learning beyond the teacher. In both successive model transfer and multi-teacher distillation, OPRD achieves higher performance with fewer student updates than existing RL and distillation approaches. Response-style analysis shows that OPRD students remain closer to models trained with verifier-based RL alone than to their weak teachers, suggesting that teacher guidance accelerates rather than redirects the student's own optimization. Results in conventional strong-to-weak distillation further demonstrate that OPRD effectively combines verifier-driven policy optimization with teacher guidance regardless of capacity ordering. Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-09 in cs.CL). Categorie arXiv: Machine Learning (cs.LG) ; Computation and Language (cs.CL). Pagina del lavoro: https://arxiv.org/abs/2609.08798 — 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 2026-10-16T12:30:25.366Z dalla pagina https://arxiv.org/catchup/cs.CL/2026-09-09?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).