Skip to content
AI News HubLIVE
Original source2 min read

Verify Before You Distill: Prompt-Level Teacher Gating for On-Policy Distillation

Summary

The paper introduces TGOPD, which verifies teacher reliability per prompt before admitting dense on-policy distillation supervision, routing unreliable prompts to verifier-grounded GRPO instead. It outperforms vanilla OPD across 4B and 35B students in math, code, and instruction following, and raises teacher-node GPU utilization from 9.8% to 78.9%.

SourcearXiv Machine LearningAuthor: Zhiwei Zhang, Zechen Sun, Fei Zhao, Kang Peng, Bin Liang, Huayu Deng, Yao Hu, Kam-Fai Wong, Mu Chuan
Verify Before You Distill: Prompt-Level Teacher Gating for On-Policy Distillation
Report an error

The correction channel is not available yet. You can copy the article reference below for later.

Correction instructions
Read article

[Submitted on 2 Sep 2026]

Title:Verify Before You Distill: Prompt-Level Teacher Gating for On-Policy Distillation

View a PDF of the paper titled Verify Before You Distill: Prompt-Level Teacher Gating for On-Policy Distillation, by Zhiwei Zhang and 8 other authors

View PDF HTML (experimental)

Abstract:On-policy distillation (OPD) accelerates post-training by providing dense token-level supervision from a frozen teacher on the student's own rollouts. Vanilla OPD applies this supervision uniformly across prompts, without checking whether the teacher is reliable for each prompt. Because reverse KL is mode-seeking, a confidently wrong teacher can induce a strong yet misleading update. Distributional proxies, such as entropy or teacher-student likelihood agreement, measure uncertainty or agreement but do not directly verify outcome correctness. We introduce Teacher-Gated On-Policy Distillation (TGOPD), built on the principle that teacher reliability should be verified at the prompt level before dense supervision is admitted. TGOPD estimates reliability from a small set of verifier-scored teacher probes and routes each prompt exclusively to dense OPD when the reliability check passes or to verifier-grounded GRPO otherwise. Across 4B and 35B students in mathematics, code, and instruction following, TGOPD outperforms Vanilla OPD in all six single-domain settings and achieves higher seven-benchmark averages at both scales under multi-domain training. By using otherwise-idle teacher capacity for reliability estimation, TGOPD also reduces teacher-side compute waste in asynchronous OPD, increasing teacher-node GPU utilization from 9.8% to 78.9% in the measured 4B single-domain run.

Comments: 17 pages, 6 figures, 7 tables

Subjects:

Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Computation and Language (cs.CL)

Cite as: arXiv:2609.02998 [cs.LG]

(or arXiv:2609.02998v1 [cs.LG] for this version)

https://doi.org/10.48550/arXiv.2609.02998

arXiv-issued DOI via DataCite (pending registration)

Submission history

From: Zhiwei Zhang [view email] [v1] Wed, 2 Sep 2026 17:54:09 UTC (6,047 KB)

Full-text links:

Access Paper:

View a PDF of the paper titled Verify Before You Distill: Prompt-Level Teacher Gating for On-Policy Distillation, by Zhiwei Zhang and 8 other authors

View PDF

HTML (experimental)

TeX Source

view license

Current browse context:

cs.LG

new | recent | 2026-09

Change to browse by:

cs cs.AI cs.CL

References & Citations

NASA ADS

Google Scholar

Semantic Scholar

Loading...

Data provided by:

Bibliographic Tools

Bibliographic and Citation Tools

Bibliographic Explorer Toggle

Bibliographic Explorer (What is the Explorer?)

Connected Papers Toggle

Connected Papers (What is Connected Papers?)

Litmaps Toggle

Litmaps (What is Litmaps?)

scite.ai Toggle

scite Smart Citations (What are Smart Citations?)

Code, Data, Media

Code, Data and Media Associated with this Article

alphaXiv Toggle

alphaXiv (What is alphaXiv?)

Links to Code Toggle

CatalyzeX Code Finder for Papers (What is CatalyzeX?)

DagsHub Toggle

DagsHub (What is DagsHub?)

GotitPub Toggle

Gotit.pub (What is GotitPub?)

Huggingface Toggle

Hugging Face (What is Huggingface?)

ScienceCast Toggle

ScienceCast (What is ScienceCast?)

Demos

Demos

Replicate Toggle

Replicate (What is Replicate?)

Spaces Toggle

Hugging Face Spaces (What is Spaces?)

Spaces Toggle

TXYZ.AI (What is TXYZ.AI?)

Related Papers

Recommenders and Search Tools

Link to Influence Flower

Influence Flower (What are Influence Flowers?)

Core recommender toggle

CORE Recommender (What is CORE?)

IArxiv recommender toggle

IArxiv Recommender (What is IArxiv?)

Author

Venue

Institution

Topic

About arXivLabs

arXivLabs: experimental projects with community collaborators

arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.

Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.

Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.

Which authors of this paper are endorsers? | Disable MathJax (What is MathJax?)

Key points and analysis

Article intelligence

EngineersAdvanced

Key points

  • TGOPD gates dense OPD supervision at the prompt level using verifier-scored teacher probes.
  • Reliable prompts use dense OPD; unreliable prompts fall back to verifier-grounded GRPO.
  • It beats vanilla OPD in all six single-domain settings and improves multi-domain averages at both scales.
  • Teacher-node GPU utilization rises from 9.8% to 78.9% in the measured 4B single-domain run.

Highlights and analysis are generated automatically and may contain errors. Check the original source.