Saved in:
Bibliographic Details
Main Authors: Jiang, Yuxuan, Ferraro, Francis
Format: Preprint
Published: 2026
Subjects:
Online Access:https://arxiv.org/abs/2606.00305
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866914619583365120
author Jiang, Yuxuan
Ferraro, Francis
author_facet Jiang, Yuxuan
Ferraro, Francis
contents On-Policy Distillation (OPD) improves large language model reasoning by training a student model on trajectories sampled from its own policy under teacher supervision. Although OPD operates on trajectories, its learning signal remains token-level: it identifies deviations through high-loss tokens and repairs them through local reverse-KL correction. We show that this "trajectory-sampled but token-learned" mechanism cannot reliably bridge student trajectories toward teacher trajectories. About 30% of high-loss tokens fall into the low-divergence regime, indicating that many are surface-form mismatches rather than real reasoning forks. Moreover, even truly divergent tokens are difficult to repair with isolated token-level supervision, since reasoning failures often unfold as short-horizon distributional drift. We propose Trajectory-aware OPD (TOPD), which uses near-future trajectory information to identify real divergent states and distribute guidance across multiple future tokens. Experiments show that suppressing non-divergent high-loss tokens improves standard OPD from 47.8% to 48.2% average accuracy, while TOPD further improves performance to 52.2%, with gains on AIME24 from 60.0% to 63.3% and AIME25 from 46.7% to 53.3%.
format Preprint
id arxiv_https___arxiv_org_abs_2606_00305
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Bridging Reasoning Trajectories in On-Policy Distillation via Near-Future Guidance
Jiang, Yuxuan
Ferraro, Francis
Computation and Language
Artificial Intelligence
On-Policy Distillation (OPD) improves large language model reasoning by training a student model on trajectories sampled from its own policy under teacher supervision. Although OPD operates on trajectories, its learning signal remains token-level: it identifies deviations through high-loss tokens and repairs them through local reverse-KL correction. We show that this "trajectory-sampled but token-learned" mechanism cannot reliably bridge student trajectories toward teacher trajectories. About 30% of high-loss tokens fall into the low-divergence regime, indicating that many are surface-form mismatches rather than real reasoning forks. Moreover, even truly divergent tokens are difficult to repair with isolated token-level supervision, since reasoning failures often unfold as short-horizon distributional drift. We propose Trajectory-aware OPD (TOPD), which uses near-future trajectory information to identify real divergent states and distribute guidance across multiple future tokens. Experiments show that suppressing non-divergent high-loss tokens improves standard OPD from 47.8% to 48.2% average accuracy, while TOPD further improves performance to 52.2%, with gains on AIME24 from 60.0% to 63.3% and AIME25 from 46.7% to 53.3%.
title Bridging Reasoning Trajectories in On-Policy Distillation via Near-Future Guidance
topic Computation and Language
Artificial Intelligence
url https://arxiv.org/abs/2606.00305