SOD: Step-wise On-policy Distillation for Small Language Model Agents proposes a new framework that adaptively reweights distillation strength at each reasoning step based on step-level divergence. This approach mitigates cascading errors in tool-integrated reasoning by attenuating misleading teacher signals in high-divergence regions while preserving dense guidance where student and teacher align. Experiments on math, science, and code benchmarks show up to 20.86% improvement over the second-best baseline, with a 0.6B student scoring 26.13% on AIME 2025.
By Qiyong Zhong, Mao Zheng, Mingyang Song, Xin Lin, Jie Sun, Houcheng Jiang, Xiang Wang, Junfeng Fang
arXiv:2607. 19450v1 Announce Type: cross Abstract: Large-scale online reinforcement learning (RL) is the predominant means of eliciting advanced abilities including long-term reasoning and agentic tool use in large language models (LLMs).
By Yunjie Chen, Xiaoxin Chen, Fang Wang
Large-scale online reinforcement learning (RL) is the predominant means of eliciting advanced abilities including long-term reasoning and agentic tool use in large language models (LLMs). However, continuing to scale it across vast task domains of interest remains challenging in both computational infrastructure and cost, especially when considering RL as merely a one-off learning stage.
arXiv:2609.36546v1 Announce Type: cross
Abstract: On-policy distillation (OPD) trains a student model on its self-generated trajectories with dense token-level teacher feedback. However, naive OPD ma...
By Shutong Wu, Xiwen Chen, Brendan Rappazzo, Daiheng Zhang, Anderson Schneider, Yuriy Nevmyvaka, Jiawei Zhang
arXiv:2607. 28582v1 Announce Type: new Abstract: On-policy self-distillation (OPSD) is a promising approach to improve reasoning language models, but it remains brittle in practice: making it work reliably often requires substantial engineering effort.
By Jiawei Xu, Minghui Liu, Juzheng Zhang, Tom Goldstein, Furong Huang
SIPO (Self‑Instructing Policy Optimization) unifies reinforcement learning with on‑policy self‑distillation by using a contrastive self‑teacher to generate token‑level credit signals. The method samples multiple rollouts per prompt, pairs each with a reference answer and its mistakes, and uses the difference in teacher log‑probabilities to provide dense feedback while still respecting the overall task reward. Experiments on reasoning and code‑generation benchmarks show that SIPO outperforms both RLVR and OPSD baselines without requiring an external teacher or extra generation steps.
By Zhenrui Yue, Huimin Zeng, Yueqi Wang, Yaokun Liu, Fengran Mo, Jinghan Zhang, Mung Yao Jia, Gyuseok Lee, Yang Zhang, Na Wei, Dong Wang