Metraipolis › archive › dc_cf3ffb58c66d
Ask to Distill: Learning to Synthesize Data for On-Policy Distillation — arXiv:2610.27450
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2610.27450
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-21T07:06:36.813Z
- SHA-256 of the archived text
db0cd193e3e99c23f6043fe24d755304f2dcc5740f95caf7ff4a2915bfcdc71e- Stable URL of this page
https://metraipolis.com/documento/dc_cf3ffb58c66d
How to cite
Ask to Distill: Learning to Synthesize Data for On-Policy Distillation — arXiv:2610.27450 — arXiv (en), https://arxiv.org/abs/2610.27450 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_cf3ffb58c66d (Metraipolis archive dc_cf3ffb58c66d, sha256 db0cd193e3e99c23f6043fe24d755304f2dcc5740f95caf7ff4a2915bfcdc71e)
BibTeX
@misc{dccf3ffb58c66d,
title = {Ask to Distill: Learning to Synthesize Data for On-Policy Distillation — arXiv:2610.27450},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2610.27450},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_cf3ffb58c66d; sha256 db0cd193e3e99c23f6043fe24d755304f2dcc5740f95caf7ff4a2915bfcdc71e},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.27450. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
On-policy distillation (OPD) can succeed with surprisingly few training questions, motivating a view of questions as carriers of distributional knowledge transfer. How can we assess whether a newly generated question is valuable for distillation, and learn to generate better questions---without access to its correct answer? We introduce Ask-to-Distill (\method), a framework that treats synthetic questions as distribution elicitors and optimizes the teacher supervision they expose along student-generated trajectories. A student-initialized synthesizer is optimized using a distributional surrogate for question value, computed from a frozen student--teacher pair without reference answers, correctness rewards, or per-candidate student updates. The reward combines teacher--student disagreement with teacher probability mass on the student's candidate support, together with length and diversity controls. The generated questions are filtered and deduplicated into a prompt-only bank for standard OPD. Across mathematical reasoning, code generation, and agentic tasks, our method improves average performance and teacher-gap recovery over unoptimized student-generated prompts and KL-filtered real questions, while matching or exceeding real-bank OPD in the evaluated settings. Signal analyses further show that optimized questions can elicit denser supervision and more total signal from shorter student responses, supporting elicited distributional supervision as an objective for synthetic distillation data. Our code, data and models are publicly available at this https URL . Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.AI, nuovi lavori del giorno). Categorie arXiv: Artificial Intelligence (cs.AI) ; Computation and Language (cs.CL). Pagina del lavoro: https://arxiv.org/abs/2610.27450 — 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 2026-10-21T07:05:43.410Z dalla pagina https://arxiv.org/list/cs.AI/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).