Pass the Baton: Trajectory-Relayed On-Policy Distillation
On-policy distillation (OPD) grounds token-level supervision in the student's own trajectory, yet suffers from prefix failure: once the student commits to a wrong reasoning direction, all subsequent generation builds on this deviation, producing misdirected continuations that elicit unreliable supervision and waste compute. We identify a teacher-student continuation asymmetry on failed prefixes, where the teacher tends to redirect while the student continues along the original direction, and con
Record details
Published: 28 July 2026
Source: arXiv cs.AI
Category: Research
Topics: Regulation · Children & education
Retrieved: 29 July 2026
Related evidence
These records share source-supplied organisations, an exact publisher byline, automatic topics or regions. The reason is shown on every link; related does not mean supporting, agreeing with or verifying this record.
Pictura: Perspective-View Self-Play at Scale for Driving
arXiv cs.AI · 28 July 2026
Weak-to-Strong On-Policy Distillation
arXiv cs.LG · 28 July 2026
A Framework for Developing University Policies on Generative AI Governance: A Cross-national Comparative Study
arXiv cs.CY · 28 July 2026
Weak-to-Strong On-Policy Distillation
HuggingFace Daily Papers · 27 July 2026
Pass the Baton: Trajectory-Relayed On-Policy Distillation
HuggingFace Daily Papers · 27 July 2026
Rethinking Classifier-Free Guidance in On-Policy Diffusion Distillation
arXiv cs.AI · 27 July 2026
How to cite this record
ethics.ai (28 July 2026), “Pass the Baton: Trajectory-Relayed On-Policy Distillation,” evidence record 14416, https://ethics.ai/record/14416 (originally published by arXiv cs.AI).
Use and limitations
This page is a stable index and citation surface for a source record. ethics.ai did not author the underlying report and has not independently verified every claim. Automatic topics may be imperfect. For consequential use, quote and cite the original publisher.