paper-with-me

Papers

TIP: Token Importance in On-Policy Distillation

2026-04-15 · Yuanda Xu, Hejian Sang, Zhengze Zhou, Ran He, Zhipeng Wang, Alborz Geramifard arxiv

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 question: which tokens carry the most useful learning signal in OPD? Our answer is that informative tokens come from two regions: positions with high student entropy, and positions with low student entropy plus high teacher--student divergence, where the student is overconfident and wrong. Empirically, student entropy is a strong first-order proxy: retaining $50\%$ of tokens with entropy-based sampling matches or exceeds all-token training while reducing peak memory by up to $47\%$. But entropy alone misses a second important region. When we isolate low-entropy, high-divergence tokens, training on fewer than $10\%$ of all tokens nearly matches full-token baselines, showing that overconfident tokens carry dense corrective signal despite being nearly invisible to entropy-only rules. We organize these findings with TIP (Token Importance in on-Policy distillation), a two-axis taxonomy over student entropy and teacher--student divergence, and give a theoretical explanation for why entropy is useful yet structurally incomplete. This view motivates type-aware token selection rules that combine uncertainty and disagreement. We validate this picture across three teacher--student pairs spanning Qwen3, Llama, and Qwen2.5 on MATH-500 and AIME 2024/2025, and on the DeepPlanning benchmark for long-horizon agentic planning, where Q3-only training on $<$$20\%$ of tokens surpasses full-token OPD. Our experiments are implemented by extending the OPD repository https://github.com/HJSang/OPSD_OnPolicyDistillation, which supports memory-efficient distillation of larger models under limited GPU budgets.

📄 PDF Abstract BibTeX arXiv:2604.14084

Code (0)

등록된 구현이 없습니다.

Tasks

Knowledge Distillation

Similar Papers 제목 키워드 기반

On the Position Bias of On-Policy Distillation

2026-06-21 · Yan Xie, Sijie Zhu, Tiansheng Wen, Bo Chen 외 arxiv

On-Policy Distillation (OPD) improves the learning efficiency of standard reinforcement learning through dense, token-level supervision from teachers. In the standard KL objective of OPD, token-level losses are uniformly…

Reinforcement Learning

Who Teaches Which Token? Verifier-Gated Multi-Expert On-Policy Distillation for Scientific Reasoning

2026-09-14 · Xun Xu, Zaixi Zhang arxiv

Multi-teacher on-policy distillation (OPD) is becoming the standard way to integrate specialist capabilities into one model: train experts with RL, then distill them into the student on its own rollouts. Existing recipes…

ReNIO: Reweighting Negative Trajectory Importance for LLM On-Policy Distillation

2026-06-22 · Chen Lin, Kedi Chen, Wei Zhang arxiv

On-policy distillation (OPD) improves LLM reasoning by training a student model on its own generated outputs, but standard OPD treats all student-generated outputs (SGOs) equally regardless of their informativeness. We o…

Reinforcement LearningMathematical ReasoningCode Generation

Breaking the Tokenizer Barrier: On-Policy Distillation across Model Families

2026-06-08 · Yifan Niu, Han Xiao, Dongyi Liu, Zelong Wang 외 arxiv

On-Policy Distillation (OPD) has become a core technique in the post-training of Large Language Models (LLMs) for transferring knowledge from domain experts to student models. However, existing OPD distillation methods r…

Self-Distilled Reasoner: On-Policy Self-Distillation for Large Language Models

2026-01-26 · Siyan Zhao, Zhihui Xie, Mengchen Liu, Jing Huang 외 arxiv

Knowledge distillation improves large language model (LLM) reasoning by compressing the knowledge of a teacher LLM to train smaller LLMs. On-policy distillation advances this approach by having the student sample its own…

Knowledge DistillationMathematical ReasoningReinforcement Learning