Skip to content

Reward-Aligned Reweighting for On-Policy Distillation

Sep 2026 · arXiv.org · Vol abs/2609.35517 · 0 citations · 40 references
Computer Science

TL;DR

Reward-Aligned Reweighting for On-Policy Distillation (R$^{2}$-OPD), which uses outcome agreement and the magnitude of teacher--student disagreement to continuously reallocate teacher supervision, achieves the highest average accuracy among the compared training methods in both cross-size and same-size distillation.

Abstract

On-policy distillation (OPD) trains a student language model with dense feedback from a stronger teacher on student-generated trajectories. Yet standard OPD weights token-level distillation terms uniformly, implicitly treating local teacher preference as a proxy for correction utility. A decision's task value, however, depends on how the student completes the subsequent reasoning. This mismatch can cause imitation to suppress viable student strategies or reinforce paths the student cannot reliably execute. Verified trajectory outcomes provide complementary evidence about continuation quality, but do not directly identify the utility of individual decisions. We introduce Reward-Aligned Reweighting for On-Policy Distillation (R$^{2}$-OPD), which uses outcome agreement and the magnitude of teacher--student disagreement to continuously reallocate teacher supervision. It gives reward-aligned corrections greater relative influence while retaining dense feedback, moving beyond uniform imitation and hard filtering. Our analysis formalizes the mismatch between local teacher preference and student continuation value and establishes sufficient conditions for reallocation to improve first-order task progress over uniform OPD. Across seven mathematical reasoning benchmarks, R$^{2}$-OPD achieves the highest average accuracy among the compared training methods in both cross-size and same-size distillation. It outperforms standard OPD on all seven benchmarks, with average gains of 3.5 and 2.4 percentage points for 1.7B and 4B students, respectively. An extension to code generation yields an average gain of 1.6 percentage points over standard OPD. These results highlight outcome-guided supervision allocation as an effective way to translate dense teacher feedback into stronger student performance across model scales and task domains.

View source

Similar papers

Preprint Aug 2026

Beyond Imitation: Filtering On-Policy Distillation by Reasoning Progress

This work proposes Reasoning-Progress-Aware Reward Filtering for On-Policy Distillation (R2-OPD), which constructs two within-trajectory rankings of reasoning spans, one from teacher-derived rewards and the other from independently estimated progress reward.

Chen Yang, Hai-Yuan Wan, Rengrong Xiong et al. · 2 citations
#machine learning Preprint Oct 2026

DiffGate: Difficulty-Gated Teacher Guidance for On-Policy Distillation

DiffGate is introduced, an outcome-gated objective that combines GRPO with selective, bounded teacher guidance, and improves pass@8 across all four model--domain settings, demonstrating improved solution coverage under the evaluation protocol.

Karn Tiwari, V. Chordia, P. PrathoshA · 0 citations
#artificial intelligence Preprint Sep 2026

Teacher Should Think Ahead: Adaptive Continuations for Reliable On-Policy Distillation

This work identifies Teacher Uncertainty Contraction (TUC), a systematic phenomenon whereby the teacher's predictive uncertainty decreases as it continues from a student-generated prefix, and proposes Adaptive-Continuations On-Policy Distillation (AC-OPD), which augments informative states along student rollouts with t...

Jin-Gang Zhou, Yu-Yi Zhou, Hai-Yang Guo et al. · 3 citations
#artificial intelligence Preprint Aug 2026

When Teacher Guidance Misleads: Reward-Aligned On-Policy Distillation

RA-OPD selects more reliable trajectories to improve student model performance without requiring additional computational cost and is evaluated on math and code benchmarks using models from the Qwen3 family and the DeepSeek-R1 family.

Si-Yuan Gan, Yu-Hang Li, Xi-Ran Wang et al. · 2 citations · ⚡1
#artificial intelligence Preprint Oct 2026

Stochastic Teacher Intervention for Agentic On-Policy Distillation

STI-OPD, a stochastic teacher intervention framework for multi-turn agentic OPD, and an Importance-Weighted Reverse KL objective that corrects the token sampling mismatch between teacher-generated responses and the student policy to preserve the original OPD objective are introduced.

Jun-Nan Liu, Lin-Hao Luo, Zhijun Chen et al. · 0 citations
#artificial intelligence Preprint Oct 2026

Transfer-Stratified On-Policy Distillation for RL-Improved Reasoning Teachers

Reinforcement learning can substantially improve a reasoning teacher, but it is unclear which of those improvements survive when the teacher supervises a smaller on-policy student. We study this question in mathematical reasoning by comparing teacher lineages before and after GRPO, multiple student scales, direct GRPO,...

Xiao-Yu Chen, Bo Shao, Tian-Gang Zhu et al. · 0 citations

Related blog posts

Microsoft Research Blog Nov 16, 2026

TokenEconomizer: Learning from Experience to Make AI Coding Agents More Efficient

AI coding agents are increasingly help developers fix bugs and build features largely on their own, but every interaction with their models consumes tokens. In each session, an agent sends instructions, tool descriptions, and session history to its model. Most efforts to reduce this cost focus on individual tasks. Yet developers return to the same […] The post TokenEconomizer: Learning from Experience to Make AI Coding Agents More Efficient appeared first on Microsoft Research.

We use cookies to run the site and, with your consent, for analytics and to show ads. See our Cookie Policy.