A Token-Level Analysis of Sampled-Token Reverse-KL On-Policy Distillation
cs.LG, cs.CL
Submitted: 2026-08-26
Updated: 2026-08-27
Comments: 16 pages, 7 figures; v2 adds Boyang Liu and Junlin Shang to the author list; scientific content unchanged
License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Terminology
Sources
- On the Direction of RLVR Updates for LLM Reasoning: Identification and Exploitation
- Reinforcement Learning via Self-Distillation
- Entropy-Aware On-Policy Distillation of Language Models
- Scaling Reasoning Efficiently via Relaxed On-Policy Distillation
- GLM-5: from Vibe Coding to Agentic Engineering
- Rethinking On-Policy Distillation of Large Language Models: Phenomenology, Mechanism, and Recipe
- DeepMath-103K: A Large-Scale, Challenging, Decontaminated, and Verifiable Mathematical Dataset for Advancing Reasoning
- Distilling the Knowledge in a Neural Network
- DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models
- Kimi K3: Open Frontier Intelligence
- Qwen3 Technical Report
- Beyond the 80/20 Rule: High-Entropy Minority Tokens Drive Effective Reinforcement Learning for LLM Reasoning
- MiMo-V2-Flash Technical Report
- Learning beyond Teacher: Generalized On-Policy Distillation with Reward Extrapolation
- DAPO: An Open-Source LLM Reinforcement Learning System at Scale
- Instruction-Following Evaluation for Large Language Models
Related papers
- Polynomial-Augmented Neural Networks (PANNs) with Weak Orthogonality Constraints for Enhanced Function and PDE Approximation
- AIRL-S: Unifying Reinforcement Learning and Search-Based Test-Time Scaling via Adversarial Inverse Reinforcement Learning
- Transformers as Bayesian In-Context Experimenters: Smoothness-Adaptive Efficient ATE Estimation
- Convergence issues in Relational Concept Analysis based on AOC-posets
- Beliefs Beyond Posteriors: Local-Consistency Optimisation for Bayesian Neural Networks
- Understanding Diffusion Models via Ratio-Based Function Approximation with SignReLU Networks