-OPD: Stabilizing Long-Horizon On-Policy Distillation with Freshness-Aware Control

arXiv:2605.17862 · cs.LG, cs.AI · Submitted 2026-05-18 · Read on arXiv

cs.LG, cs.AI

Submitted: 2026-05-18

Updated: 2026-10-11

Code: https://github.com/THUDM/slime

Terminology

• Distillation — 2× in this paper · explained in 2 episodes
• On-Policy Distillation — 2× in this paper · explained in 3 episodes
• Large Language Models (LLMs) — 1× in this paper · explained in 6 episodes

Sources

• Llama 2: Open Foundation and Fine-Tuned Chat Models
• Proximal Policy Optimization Algorithms
• DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models
• Qwen3 Technical Report
• MiMo-V2-Flash Technical Report
• GLM-5: from Vibe Coding to Agentic Engineering
• HybridFlow: A Flexible and Efficient RLHF Framework
• DAPO: An Open-Source LLM Reinforcement Learning System at Scale
• The Entropy Mechanism of Reinforcement Learning for Reasoning Language Models
• ReLaX: Reasoning with Latent Exploration for Large Reasoning Models
• A Survey of On-Policy Distillation for Large Language Models
• Rethinking On-Policy Distillation of Large Language Models: Phenomenology, Mechanism, and Recipe
• Reinforcement Learning via Self-Distillation
• Self-Distilled Reasoner: On-Policy Self-Distillation for Large Language Models
• OPSDL: On-Policy Self-Distillation for Long-Context Language Models
• CRISP: Compressed Reasoning via Iterative Self-Policy Distillation
• Scaling Reasoning Efficiently via Relaxed On-Policy Distillation
• On-Policy Context Distillation for Language Models
• Learning beyond Teacher: Generalized On-Policy Distillation with Reward Extrapolation
• Revisiting On-Policy Distillation: Empirical Failure Modes and Simple Fixes

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: The Predictive Cost of ELBO Binding
• Understanding Diffusion Models via Ratio-Based Function Approximation with SignReLU Networks