The paper reports that in on‑policy distillation for large language models, reasoning performance can be improved by supervising only a tiny fraction of generated tokens—sometimes just one or two tokens per reasoning trajectory, about 0.05% of all tokens. This sparse supervision consistently matches or exceeds full‑token training across nine teacher‑student setups on mathematical reasoning, and is also validated on coding reasoning, Llama models, and PPO‑based reinforcement learning with verifiable reward. The findings suggest that effective post‑training does not require token‑intensive supervision and may align more closely with natural learning processes that focus on critical reasoning steps.
By Zhishuai Liu, Xingzi Xu, Mehmet Saygin Seyfioglu, Pan Xu, Karim Bouyarmane
arXiv:2605. 15532v3 Announce Type: replace-cross Abstract: Distillation enables compact Vision-Language Models (VLMs) to obtain strong reasoning capabilities, yet the prompts driving this process are typically chosen via simple heuristics or aggregated from off-the-shelf datasets.
By Jaehun Jung, Hyunwoo Kim, Brandon Cui, Ximing Lu, David Acuna, Prithviraj Ammanabrolu, Yejin Choi
arXiv:2602. 22495v3 Announce Type: replace-cross Abstract: Reinforcement learning (RL) post-training has recently driven major gains in long chain-of-thought reasoning large language models (LLMs), but the high inference cost of such models motivates distillation into smaller students.
By Zhaoyang Zhang, Shuli Jiang, Yantao Shen, Yuting Zhang, Dhananjay Ram, Shuo Yang, Zhuowen Tu, Wei Xia, Stefano Soatto
The paper introduces Echo-GRPO, a method that rewrites privileged reasoning traces into a model’s own idiolect to align off‑policy supervision with the student policy’s vocabulary. By preserving semantics through Dual‑Reference Decoding, Echo‑GRPO mitigates gradient clipping on critical reasoning tokens and improves reasoning distillation. The approach is instantiated as VideoEcho‑R1 for video reasoning, yielding consistent gains across multiple multimodal LLM backbones and benchmarks, and it can be applied as a plug‑in to both RL and supervised fine‑tuning frameworks.
By Ji Soo Lee, Jinyoung Park, Seohyun Lee, Jongha Kim, Joonmyung Choi, Jinsung Yoon, Hyunwoo J. Kim
arXiv:2608.29623v1 Announce Type: cross
Abstract: Recent advances in large reasoning models (LRMs) have shown strong performance on complex problems through long chain-of-thought (Long CoT) reasoning...
By Yangsong Lan, Renkai Hu, HongKai Zheng, Bo Zhang, Renzhi Wang, Hongliang Dai, Piji Li
OBC‑Prune introduces an outcome‑based calibration approach for pruning large reasoning models, focusing on the causal importance of each reasoning sentence rather than uniform activation salience. By pairing correct and incorrect rollouts and using intervention‑based analysis, it assigns per‑token weights that guide one‑shot pruning methods such as SparseGPT, Wanda, and ALPS. Experiments on DeepSeek‑R1‑Distill‑Qwen models show consistent accuracy gains and shorter reasoning traces across multiple benchmarks at 40‑50% sparsity.
By Ha Lan Nguyen, Huy Hoang Tran, Trac-Duy Tran, Dung D. Le
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
Ladders-of-Thought (LoT) is a framework that enhances reasoning in small- to mid-scale large language models by automatically generating easier variants of reasoning problems and organizing them into difficulty buckets. It uses a self‑evolving bandit scheduler to adaptively allocate training, improving performance across math and multi‑hop reasoning tasks on 1–8 B models. LoT achieves significant gains (e.g., +32 pp on AddSub, +16 pp on QASC) and converges faster than staged curricula.
By Minghui Liu, Thomas Magelinski, Dehao Yuan, Qi Yu, Furong Huang
arXiv:2602. 17686v5 Announce Type: replace-cross Abstract: Chain-of-Thought (CoT) reasoning has significantly improved LLMs' mathematical problem-solving capabilities, but distilling such capabilities into smaller models remains challenging due to the capacity mismatch between verbose teachers and compact students.
By Bowen Yu, Sheng Zhang, Binhao Wang, Yi Wen, Jingtong Gao, Bowen Liu, Zimo Zhao, Shanshan Ye, Wanyu Wang, Maolin Wang, Xiangyu Zhao
The paper investigates data efficiency and selection in On‑Policy Distillation (OPD) for large language models. It shows that 1‑shot OPD—training on a single example—consistently improves performance, especially when the example is hard, and that longer chain‑of‑thought (CoT) paths drive the gains rather than token entropy. Based on these findings, the authors propose a simple hard‑example selection strategy that, using only eight carefully chosen hard examples, matches the performance of a 17,000‑example baseline across models from 1.5B to 7B parameters.
By Zhinan Hou, Jiaqi Zhang, Xunliang Cai, Keyou You
arXiv:2610. 02191v1 Announce Type: new Abstract: While Large Language Models (LLMs) have demonstrated striking capabilities on frontier mathematical problems, it remains unclear whether they possess the structural mathematical understanding underlying their solutions.
By Shuo Xing, Zilin Dai, Chengyuan Qian, Fangzhou Lin, Wenjing Chen, Ping He, Pan Lu, Alvaro Velasquez, Mohit Bansal, Zhengzhong Tu
Contrastive On-Policy Distillation (COPD) is a framework that improves on-policy distillation by using a frozen teacher to evaluate student states under two contrasting prompts—one encouraging low reasoning effort and one encouraging high effort. The difference in log‑probabilities between these prompts provides a token‑level advantage signal that guides the student toward more concise and efficient reasoning strategies. Experiments on nine multimodal benchmarks show that COPD reduces reasoning length while maintaining task performance, and the contrastive approach can also be applied to on‑policy self‑distillation, allowing a model to compress its own reasoning without an external teacher.
By Jiacheng Ruan, Jun Tang, Wenzhen Yuan, Ting Liu, Shuai Bai, Dayiheng Liu, Zhibo Yang, Yuzhuo Fu