English
Related papers

Related papers: Tailoring Teaching to Aptitude: Direction-Adaptive…

200 papers

Reasoning-centric large language models (LLMs) achieve strong performance by generating intermediate reasoning trajectories, but often incur excessive token usage and high inference-time decoding cost. We observe that, when solving the same…

Artificial Intelligence · Computer Science 2026-05-12 Han Yang , Mingyan Wu , Bailan He , Zeyu Cao , Sikuan Yan , Kevin Qinghong Lin , Zifeng Ding

On-policy distillation (OPD) leverages dense teacher rewards to enhance reasoning models. However, scaling OPD to long-horizon tasks exposes a critical flaw: as the student's generated prefix inevitably diverges from the teacher's thought…

Machine Learning · Computer Science 2026-05-29 Zhicheng Yang , Zhijiang Guo , Yifan Song , Minrui Xu , Yongxin Wang , Yiwei Wang , Xiaodan Liang , Jing Tang

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,…

Knowledge distillation is central to LLM post-training, yet its design space remains poorly understood, especially alongside reinforcement learning (RL). We show that the prevailing paradigms, off-policy distillation and on-policy…

Machine Learning · Computer Science 2026-05-19 Anhao Zhao , Haoran Xin , Yingqi Fan , Junlong Tong , Wenjie Li , Xiaoyu Shen

On-policy distillation (OPD) has shown strong potential for transferring reasoning ability from frontier or domain-specific models to smaller students. While effective on static single-turn tasks, its behavior in multi-turn agent settings…

Machine Learning · Computer Science 2026-04-30 Jiaqi Wang , Wenhao Zhang , Weijie Shi , Yaliang Li , James Cheng

Reinforcement learning for multi-turn agents suffers from a credit-assignment mismatch: rewards are sparse and trajectory-level, while success often hinges on a few local decisions. Existing online policy distillation (OPD) provides denser…

Artificial Intelligence · Computer Science 2026-05-27 Yanfei Zhang , Xu Lin , Chenglin Wu

On-policy knowledge distillation (OPD) trains a student on its own rollouts under token-level supervision from a teacher. Not all token positions matter equally, but existing views of token importance are incomplete. We ask a direct…

Machine Learning · Computer Science 2026-05-22 Yuanda Xu , Hejian Sang , Zhengze Zhou , Ran He , Zhipeng Wang , Alborz Geramifard

On-policy self-distillation (SD) improves LLM reasoning by using teacher-side privileged information (PI) to turn sparse verifier outcomes into dense token-level supervision. Existing methods usually assume trusted PI, such as reference…

Computation and Language · Computer Science 2026-05-28 Jiazhen Huang , Xiao Chen , Xiao Luo , Yong Dai , Senkang Hu , Yuzhi Zhao

Multimodal Large Language Models (MLLMs) still struggle with fine-grained visual understanding, where answers often depend on small but decisive evidence in the full image. We observe a regional-to-global perception gap: the same MLLM…

Computer Vision and Pattern Recognition · Computer Science 2026-05-28 Qianhao Yuan , Jie Lou , Xing Yu , Hongyu Lin , Le Sun , Xianpei Han , Yaojie Lu

The push to compress and impart the proficiency of Large Language Models (LLMs) into more deployable and efficient Small Language Models (SLMs) has benefited from improvements in knowledge distillation (KD) techniques. These techniques…

Artificial Intelligence · Computer Science 2025-07-02 Shreyansh Padarha

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

GRPO-style RLVR trains reasoning models from multiple on-policy attempts per prompt, but typically uses these attempts only through terminal rewards. We show that a mixed group contains a richer process signal: a correct completion is a…

Machine Learning · Computer Science 2026-05-19 Zhiquan Tan , Yinrong Hong

On-policy distillation (OPD) trains student models under their own induced distribution while leveraging supervision from stronger teachers. We identify a failure mode of OPD: as training progresses, on-policy rollouts can undergo abrupt…

Computation and Language · Computer Science 2026-04-10 Feng Luo , Yu-Neng Chuang , Guanchu Wang , Zicheng Xu , Xiaotian Han , Tianyi Zhang , Vladimir Braverman

The landscape of high-performance image generation models is currently shifting from the inefficient multi-step ones to the efficient few-step counterparts (e.g, Z-Image-Turbo and FLUX.2-klein). However, these models present significant…

Computer Vision and Pattern Recognition · Computer Science 2026-05-27 Dengyang Jiang , Xin Jin , Dongyang Liu , Zanyi Wang , Mingzhe Zheng , Ruoyi Du , Xiangpeng Yang , Qilong Wu , Zhen Li , Peng Gao , Harry Yang , Steven Hoi

On-Policy Distillation (OPD) has emerged as a dominant post-training paradigm for large language models, especially for reasoning domains. However, OPD remains unstable in practice due to the high gradient variance of its single-sample…

Machine Learning · Computer Science 2026-05-11 Minjae Oh , Sangjun Song , Gyubin Choi , Yunho Choi , Yohan Jo

Large language models (LLMs) are increasingly adapted to proprietary and domain-specific corpora that contain sensitive information, creating a tension between formal privacy guarantees and efficient deployment through model compression.…

Machine Learning · Computer Science 2026-04-07 Fatemeh Khadem , Sajad Mousavi , Yi Fang , Yuhong Liu

Knowledge distillation (KD) is a standard route to compress Large Language Models (LLMs) into compact students, yet most pipelines uniformly apply token-wise loss regardless of teacher confidence. This indiscriminate supervision amplifies…

Computation and Language · Computer Science 2025-11-18 Haiduo Huang , Jiangcheng Song , Yadong Zhang , Pengju Ren

Self-Distillation Policy Optimization (SDPO) provides dense token-level credit assignment for reinforcement learning with large language models by leveraging the model's own feedback-conditioned predictions as a self-teacher. Unlike GRPO,…

Machine Learning · Computer Science 2026-05-28 Zehao Liu , Yuanpu Cao , Jinghui Chen , Vasant G. Honavar

Large language models (LLMs) have achieved remarkable performance across diverse domains, yet their enormous computational and memory requirements hinder deployment in resource-constrained environments. Knowledge distillation offers a…

Computation and Language · Computer Science 2026-05-05 Hao Zhang , Zhibin Zhang , Guangxin Wu , Wanyi Ning , Jiafeng Guo , Xueqi Cheng

Knowledge distillation (KD) is a key technique for compressing large-scale language models (LLMs), yet prevailing logit-based methods typically employ static strategies that are misaligned with the dynamic learning process of student…

Computation and Language · Computer Science 2025-10-14 Xurong Xie , Zhucun Xue , Jiafu Wu , Jian Li , Yabiao Wang , Xiaobin Hu , Yong Liu , Jiangning Zhang