Metraipolis › archive › dc_fd657384c395

Distilling the Shape of Uncertainty: What On-Policy Distillation Learns and How to Predict It — arXiv:2610.39020

Source: arXiv (en)
Original: https://arxiv.org/abs/2610.39020
License (as archived): CC0-1.0
Archived: 2026-10-31T06:13:02.443Z
SHA-256 of the archived text: d025d933d3226f236a606a8266ace3e47a78ac608ae158cd15381f860fca4c87
Stable URL of this page: https://metraipolis.com/documento/dc_fd657384c395

How to cite

Distilling the Shape of Uncertainty: What On-Policy Distillation Learns and How to Predict It — arXiv:2610.39020 — arXiv (en), https://arxiv.org/abs/2610.39020 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_fd657384c395 (Metraipolis archive dc_fd657384c395, sha256 d025d933d3226f236a606a8266ace3e47a78ac608ae158cd15381f860fca4c87)

BibTeX

@misc{dcfd657384c395,
  title = {Distilling the Shape of Uncertainty: What On-Policy Distillation Learns and How to Predict It — arXiv:2610.39020},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2610.39020},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_fd657384c395; sha256 d025d933d3226f236a606a8266ace3e47a78ac608ae158cd15381f860fca4c87},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.39020. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

On-policy distillation (OPD) is increasingly adopted in the post-training of large language models. However, a stronger teacher may fail to improve the student. Prior work attributes this to a lack of new knowledge, but what that knowledge is remains unclear. We show that OPD learns not only \textit{how to branch}, i.e., which candidate tokens to consider, but also \textit{where to branch}, i.e., where to keep reasoning paths open. The existing token-overlap metric captures only the former. We introduce the entropy shape, the sequence of next-token entropies, to characterize the latter, and propose the Entropy Shape Correlation (ESC) to measure student-teacher agreement in entropy shape. Tracking ESC throughout training and intervening on a teacher's entropy while preserving its token preferences, we empirically show that \textit{where to branch} causally drives OPD gains, independently of \textit{how to branch}. Finally, we formulate OPD outcome prediction as a teacher selection task and establish an evaluation protocol for it, under which pre-OPD ESC selects the best teacher for both students and outperforms existing predictors. Code is available at this https URL .

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/2610.39020 — 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 2026-10-31T06:07:08.690Z dalla pagina https://arxiv.org/list/cs.AI/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).