English
Related papers

Related papers: SD-Search: On-Policy Hindsight Self-Distillation f…

200 papers

On-policy self-distillation trains a reasoning model on its own rollouts while a teacher, often the same model conditioned on privileged context, provides dense token-level supervision. Existing objectives typically weight the teacher's…

Artificial Intelligence · Computer Science 2026-05-14 Junlong Ke , Zichen Wen , Weijia Li , Conghui He , Linfeng Zhang

On-policy distillation (OPD), which aligns the student with the teacher's logit distribution on student-generated trajectories, has demonstrated strong empirical gains in improving student performance and often outperforms off-policy…

Machine Learning · Computer Science 2026-02-27 Wenkai Yang , Weijie Liu , Ruobing Xie , Kai Yang , Saiyong Yang , Yankai Lin

On-policy distillation (OPD), which supervises a student on its own sampled trajectories, has emerged as a data-efficient post-training method for improving reasoning while avoiding the reward dependence of reinforcement learning and the…

Computer Vision and Pattern Recognition · Computer Science 2026-05-14 Ke Zhang , Yunjie Tian , Dongdi Zhao , Yijiang Li , Yuanye Liu , Vishal M Patel , Di Fu

On-policy distillation offers dense, per-token supervision for training reasoning models; however, it remains unclear under which conditions this signal is beneficial and under which it is detrimental. Which teacher model should be used,…

On-policy self-distillation has become a strong recipe for LLM reasoning, where a privileged teacher supervises the student's own rollouts while conditioning on the reference solution. A design choice shared by nearly all such methods,…

Artificial Intelligence · Computer Science 2026-05-28 Zihao Han , Tiangang Zhang , Huaibin Wang , Yilun Sun

Adapting large language models (LLMs) to long-context tasks requires post-training methods that remain accurate and coherent over thousands of tokens. Existing approaches are limited in several ways: 1) off-policy methods such as supervised…

Computation and Language · Computer Science 2026-05-13 Miguel Moura Ramos , Duarte M. Alves , André F. T. Martins

On-policy self-distillation has emerged as a promising paradigm for post-training language models, in which the model conditions on environment feedback to serve as its own teacher, providing dense token-level rewards without external…

Machine Learning · Computer Science 2026-05-13 Guobin Shen , Lei Huang , Xiang Cheng , Chenxiao Zhao , Jindong Li , Dongcheng Zhao , Xing Yu

Large language models trained with reinforcement learning (RL) for mathematical reasoning face a fundamental challenge: on problems the model cannot solve at all - "cliff" prompts - the RL gradient vanishes entirely, preventing any learning…

Machine Learning · Computer Science 2026-03-26 Ken Ding

Direct Preference Optimization (DPO) is a powerful paradigm to align language models with human preferences using pairwise comparisons. However, its binary win-or-loss supervision often proves insufficient for training small models with…

Computation and Language · Computer Science 2025-09-23 Minchan Kwon , Junwon Ko , Kangil Kim , Junmo Kim

Large language models are often post-trained with sparse verifier rewards, which indicate whether a sampled trajectory succeeds but provide limited guidance about where reasoning succeeds or fails. On-policy distillation (OPD) offers denser…

Machine Learning · Computer Science 2026-05-14 Weichen Yu , Xiaomin Li , Yizhou Zhao , Xiaoze Liu , Ruowang Zhang , Haixin Wang , Yinyi Luo , Chen Henry Wu , Gaurav Mittal , Matt Fredrikson , Yu Hu

Distilling reasoning traces from strong large language models into smaller ones is a promising route to improve intelligence in resource-constrained settings. Existing approaches face a fundamental trade-off: offline distillation from…

Computation and Language · Computer Science 2026-05-15 Yumeng Zhang , Zhengbang Yang , Yevin Nikhel Goonatilake , Zhuangdi Zhu

Self-supervised learning solves pretext prediction tasks that do not require annotations to learn feature representations. For vision tasks, pretext tasks such as predicting rotation, solving jigsaw are solely created from the input data.…

Computer Vision and Pattern Recognition · Computer Science 2021-07-01 Prashant Bhat , Elahe Arani , Bahram Zonooz

Solving complex long-horizon tasks requires strong planning and reasoning capabilities. Although datasets with explicit chain-of-thought (CoT) rationales can substantially benefit learning, they are costly to obtain. To address this…

Artificial Intelligence · Computer Science 2026-05-13 Shengjie Wang , Guanghe Li , Zonghan Yang , Yang Gao

Current post-training methods in verifiable settings fall into two categories. Reinforcement learning (RLVR) relies on binary rewards, which are broadly applicable and powerful, but provide only sparse supervision during training.…

Computation and Language · Computer Science 2026-04-15 Yinghui He , Simran Kaur , Adithya Bhaskar , Yongjin Yang , Jiarui Liu , Narutatsu Ri , Liam Fowl , Abhishek Panigrahi , Danqi Chen , Sanjeev Arora

Reinforcement learning can solve decision-making problems and train an agent to behave in an environment according to a predesigned reward function. However, such an approach becomes very problematic if the reward is too sparse and so the…

Artificial Intelligence · Computer Science 2024-06-12 Matej Pecháč , Michal Chovanec , Igor Farkaš

Policy distillation, which transfers a teacher policy to a student policy has achieved great success in challenging tasks of deep reinforcement learning. This teacher-student framework requires a well-trained teacher model which is…

Machine Learning · Computer Science 2020-06-09 Kwei-Herng Lai , Daochen Zha , Yuening Li , Xia Hu

Deep Research agents tackle knowledge-intensive tasks through multi-round retrieval and decision-oriented generation. While reinforcement learning (RL) has been shown to improve performance in this paradigm, its contributions remain…

Computation and Language · Computer Science 2026-02-24 Yinuo Xu , Shuo Lu , Jianjie Cheng , Meng Wang , Qianlong Xie , Xingxing Wang , Ran He , Jian Liang

Large reasoning models achieve strong performance on complex tasks through long chain-of-thought (CoT) trajectories, but directly transferring such reasoning processes to smaller models remains challenging. A key difficulty is that not all…

Computation and Language · Computer Science 2026-04-06 Chaoqun He , Yingfa Chen , Chaojun Xiao , Xu Han , Lijie Wen

On-policy distillation (OPD), which samples trajectories from the student model and supervises them with a teacher at the token level, avoids relying solely on verifiable terminal rewards and can yield better generalization than off-policy…

On-policy reinforcement learning has become the dominant paradigm for reasoning alignment in large language models, yet its sparse, outcome-level rewards make token-level credit assignment notoriously difficult. On-Policy Distillation (OPD)…

Machine Learning · Computer Science 2026-04-14 Binbin Zheng , Xing Ma , Yiheng Liang , Jingqing Ruan , Xiaoliang Fu , Kepeng Lin , Benchang Zhu , Ke Zeng , Xunliang Cai