# Gains and Collapse in On-Policy Distillation:A Reinforcement Learning Perspective

arXiv: https://arxiv.org/abs/2610.03185 · HTML: https://mlsanity.com/paper/2610.03185/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Placed in Reward hacking and misspecification, one of 10 directions in LLM alignment / RLHF / preference optimisation.
- Builds on 27 of its 31 references that are in the map.
- First posted on arXiv 2 Oct 2026.

## Placed in

- LLM alignment / RLHF / preference optimisation / [Reward hacking and misspecification](https://mlsanity.com/alignment/reward-hacking-and-misspecification/)
- Knowledge distillation / [On-policy and self-distillation for LLMs](https://mlsanity.com/distill/on-policy-and-self-distillation-for-llms/)
- Deep reinforcement learning methods / [RL training of language and agent models](https://mlsanity.com/deeprl/rl-training-of-language-and-agent-models/)

## Builds on (most cited first, 27 of 27)

- [Distilling the Knowledge in a Neural Network](https://mlsanity.com/paper/1503.02531/) (2015)
- [Evaluating Large Language Models Trained on Code](https://mlsanity.com/paper/2107.03374/) (2021)
- [Understanding R1-Zero-Like Training: A Critical Perspective](https://mlsanity.com/paper/2503.20783/) (2025)
- [Does Reinforcement Learning Really Incentivize Reasoning Capacity in LLMs Beyond the Base Model?](https://mlsanity.com/paper/2504.13837/) (2025)
- [On-Policy Distillation of Language Models: Learning from Self-Generated Mistakes](https://mlsanity.com/paper/2306.13649/) (2023)
- [Sequence-Level Knowledge Distillation](https://mlsanity.com/paper/1606.07947/) (2016)
- [Distilling Step-by-Step! Outperforming Larger Language Models with Less Training Data and Smaller Model Sizes](https://mlsanity.com/paper/2305.02301/) (2023)
- [GLM-5: from Vibe Coding to Agentic Engineering](https://mlsanity.com/paper/2602.15763/) (2026)
- [Self-Distilled Reasoner: On-Policy Self-Distillation for Large Language Models](https://mlsanity.com/paper/2601.18734/) (2026)
- [Large Language Models Are Reasoning Teachers](https://mlsanity.com/paper/2212.10071/) (2022)
- [Rethinking On-Policy Distillation of Large Language Models: Phenomenology, Mechanism, and Recipe](https://mlsanity.com/paper/2604.13016/) (2026)
- [MiMo-V2-Flash Technical Report](https://mlsanity.com/paper/2601.02780/) (2026)
- [On-Policy Context Distillation for Language Models](https://mlsanity.com/paper/2602.12275/) (2026)
- [Learning beyond Teacher: Generalized On-Policy Distillation with Reward Extrapolation](https://mlsanity.com/paper/2602.12125/) (2026)
- [Revisiting On-Policy Distillation: Empirical Failure Modes and Simple Fixes](https://mlsanity.com/paper/2603.25562/) (2026)
- [Echo Chamber: RL Post-training Amplifies Behaviors Learned in Pretraining](https://mlsanity.com/paper/2504.07912/) (2025)
- [RRM: Robust Reward Model Training Mitigates Reward Hacking](https://mlsanity.com/paper/2409.13156/) (2024)
- [Self-Distillation Zero: Self-Revision Turns Binary Rewards into Dense Supervision](https://mlsanity.com/paper/2604.12002/) (2026)
- [JustRL: Scaling a 1.5B LLM with a Simple RL Recipe](https://mlsanity.com/paper/2512.16649/) (2025)
- [The Many Faces of On-Policy Distillation: Pitfalls, Mechanisms, and Fixes](https://mlsanity.com/paper/2605.11182/) (2026)
- [Demystifying OPD: Length Inflation and Stabilization Strategies for Large Language Models](https://mlsanity.com/paper/2604.08527/) (2026)
- [Understanding In-Context Learning from Repetitions](https://mlsanity.com/paper/2310.00297/) (2023)
- [Policy Filtration for RLHF to Mitigate Noise in Reward Models](https://mlsanity.com/paper/2409.06957/) (2024)
- [Adversarial Reward Auditing for Active Detection and Mitigation of Reward Hacking](https://mlsanity.com/paper/2602.01750/) (2026)
- [Cross-Tokenizer On-Policy Distillation via Byte-Prefix Marginalization](https://mlsanity.com/paper/2607.22334/) (2026)
- [SimpleOPD: Simple Tokenizer-Agnostic On-Policy Distillation for Long-Context Reasoning](https://mlsanity.com/paper/2608.14277/) (2026)
- [Detecting RLVR Training Data via Structural Convergence of Reasoning](https://mlsanity.com/paper/2602.11792/) (2026)

## Abstract (arXiv)

On-policy distillation (OPD) has become an important approach to language model post-training. However, despite its performance gains, OPD can also collapse into excessively long and repetitive generation, and the mechanism underlying these divergent outcomes remains poorly understood. We explain these outcomes through a reinforcement learning perspective: the teacher implicitly rewards student behaviors, even those it rarely exhibits itself. From this perspective, our experiments show that OPD improves performance without expanding the student's capabilities. When the implicit reward model is reliable, OPD makes correct responses easier to sample. In contrast, when the preference misaligns with quality, reward hacking happens: the implicit reward model amplifies overlong, repetitive student rollouts, even though it rarely generates such text itself. Guided by this diagnosis, we find that masking unhealthy responses during training and using SFT initialization can each effectively mitigate the collapse. Together, these findings show that OPD amplifies student behaviors favored by the teacher's implicit feedback, shifting the focus from how well the teacher generates to how reliably it evaluates student rollouts. Our code is available at https://github.com/HancCui/opd_hacking.