The Teacher Is a Direction, Not a Destination: Extrapolating RL-Induced Representation Residuals in On-Policy Distillation

arXiv:2609.36484 · cs.LG · Submitted 2026-09-29 · Read on arXiv

cs.LG

Submitted: 2026-09-29

Updated: 2026-10-14

Comments: 19 pages; revised related-work discussion

Code: https://github.com/xixixixixxxx/RIDE

License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/

Terminology

• Distillation — 1× in this paper · explained in 2 episodes
• On-Policy Distillation — 1× in this paper · explained in 3 episodes

Sources

• Phi-4-Mini Technical Report: Compact yet Powerful Multimodal Language Models via Mixture-of-LoRAs
• Process Reinforcement through Implicit Rewards
• Revisiting On-Policy Distillation: Empirical Failure Modes and Simple Fixes
• The Llama 3 Herd of Models
• JustRL: Scaling a 1.5B LLM with a Simple RL Recipe
• Distilling the Knowledge in a Neural Network
• ErrOPD: When Does On-Policy Distillation Transfer Teacher Errors?
• RISE: Recursive Improvement via Self-Extrapolating Policy Distillation
• Rethinking On-Policy Distillation of Large Language Models: Phenomenology, Mechanism, and Recipe
• PHF: Privileged Hidden Flow for On-Policy Self-Distillation
• REOPD: Reliability-Adaptive Reward Extrapolation for On-Policy Distillation
• Steering Language Models With Activation Engineering
• MiMo-V2-Flash Technical Report
• Qwen2.5-Math Technical Report: Toward Mathematical Expert Model via Self-Improvement
• Qwen3 Technical Report
• OPRD: On-Policy Representation Distillation
• Learning beyond Teacher: Generalized On-Policy Distillation with Reward Extrapolation
• Black-Box On-Policy Distillation of Large Language Models
• The Path Not Taken: RLVR Provably Learns Off the Principals
• Representation Engineering: A Top-Down Approach to AI Transparency

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