# CRISP: Compressed Reasoning via Iterative Self-Policy Distillation

arXiv: https://arxiv.org/abs/2603.05433 · HTML: https://mlsanity.com/paper/2603.05433/

## Key facts

- Cited by 58 papers in the map (data through 14 Oct 2026).
- Placed in Prompting and chain-of-thought reasoning, one of 17 directions in LLM reasoning.
- Builds on 35 of its 41 references that are in the map.
- First posted on arXiv 5 Mar 2026.

## Placed in

- LLM reasoning / [Prompting and chain-of-thought reasoning](https://mlsanity.com/lr/prompting-and-chain-of-thought-reasoning/)

## Builds on (most cited first, 30 of 35)

- [Qwen3 Technical Report](https://mlsanity.com/paper/2505.09388/) (2025)
- [Measuring Massive Multitask Language Understanding](https://mlsanity.com/paper/2009.03300/) (2020)
- [Measuring Mathematical Problem Solving With the MATH Dataset](https://mlsanity.com/paper/2103.03874/) (2021)
- [Gemini 2.5: Pushing the Frontier with Advanced Reasoning, Multimodality, Long Context, and Next Generation Agentic Capabilities](https://mlsanity.com/paper/2507.06261/) (2025)
- [GPQA: A Graduate-Level Google-Proof Q&A Benchmark](https://mlsanity.com/paper/2311.12022/) (2023)
- [DAPO: An Open-Source LLM Reinforcement Learning System at Scale](https://mlsanity.com/paper/2503.14476/) (2025)
- [HybridFlow: A Flexible and Efficient RLHF Framework](https://mlsanity.com/paper/2409.19256/) (2024)
- [Scaling LLM Test-Time Compute Optimally can be More Effective than Scaling Model Parameters](https://mlsanity.com/paper/2408.03314/) (2024)
- [s1: Simple test-time scaling](https://mlsanity.com/paper/2501.19393/) (2025)
- [SGLang: Efficient Execution of Structured Language Model Programs](https://mlsanity.com/paper/2312.07104/) (2023)
- [Do NOT Think That Much for 2+3=? On the Overthinking of o1-Like LLMs](https://mlsanity.com/paper/2412.21187/) (2024)
- [Beyond the 80/20 Rule: High-Entropy Minority Tokens Drive Effective Reinforcement Learning for LLM Reasoning](https://mlsanity.com/paper/2506.01939/) (2025)
- [The Entropy Mechanism of Reinforcement Learning for Reasoning Language Models](https://mlsanity.com/paper/2505.22617/) (2025)
- [Self-Distilled Reasoner: On-Policy Self-Distillation for Large Language Models](https://mlsanity.com/paper/2601.18734/) (2026)
- [L1: Controlling How Long A Reasoning Model Thinks With Reinforcement Learning](https://mlsanity.com/paper/2503.04697/) (2025)
- [Reinforcement Learning via Self-Distillation](https://mlsanity.com/paper/2601.20802/) (2026)
- [Chain of Draft: Thinking Faster by Writing Less](https://mlsanity.com/paper/2502.18600/) (2025)
- [TokenSkip: Controllable Chain-of-Thought Compression in LLMs](https://mlsanity.com/paper/2502.12067/) (2025)
- [Self-Distillation Enables Continual Learning](https://mlsanity.com/paper/2601.19897/) (2026)
- [ThinkPrune: Pruning Long Chain-of-Thought of LLMs via Reinforcement Learning](https://mlsanity.com/paper/2504.01296/) (2025)
- [On-Policy Context Distillation for Language Models](https://mlsanity.com/paper/2602.12275/) (2026)
- [Wait, We Don't Need to "Wait"! Removing Thinking Tokens Improves Reasoning Efficiency](https://mlsanity.com/paper/2506.08343/) (2025)
- [DeepPlanning: Benchmarking Long-Horizon Agentic Planning with Verifiable Constraints](https://mlsanity.com/paper/2601.18137/) (2026)
- [The Choice of Divergence: A Neglected Key to Mitigating Diversity Collapse in Reinforcement Learning with Verifiable Reward](https://mlsanity.com/paper/2509.07430/) (2025)
- [DLER: Doing Length pEnalty Right - Incentivizing More Intelligence per Token via Reinforcement Learning](https://mlsanity.com/paper/2510.15110/) (2025)
- [Concise Reasoning, Big Gains: Pruning Long Reasoning Trace with Difficulty-Aware Prompting](https://mlsanity.com/paper/2505.19716/) (2025)
- [TrimR: Verifier-based Training-Free Thinking Compression for Efficient Test-Time Scaling](https://mlsanity.com/paper/2505.17155/) (2025)
- [Leash: Adaptive Length Penalty and Reward Shaping for Efficient Large Reasoning Model](https://mlsanity.com/paper/2512.21540/) (2025)
- [The Overthinker's DIET: Cutting Token Calories with DIfficulty-AwarE Training](https://mlsanity.com/paper/2505.19217/) (2025)
- [Reasoning Efficiently Through Adaptive Chain-of-Thought Compression: A Self-Optimizing Framework](https://mlsanity.com/paper/2509.14093/) (2025)

## Built on by (most cited first, 30 of 58)

- [Rethinking On-Policy Distillation of Large Language Models: Phenomenology, Mechanism, and Recipe](https://mlsanity.com/paper/2604.13016/) (2026)
- [A Survey of On-Policy Distillation for Large Language Models](https://mlsanity.com/paper/2604.00626/) (2026)
- [TIP: Token Importance in On-Policy Distillation](https://mlsanity.com/paper/2604.14084/) (2026)
- [The Many Faces of On-Policy Distillation: Pitfalls, Mechanisms, and Fixes](https://mlsanity.com/paper/2605.11182/) (2026)
- [Prune-OPD: Efficient and Reliable On-Policy Distillation for Long-Horizon Reasoning](https://mlsanity.com/paper/2605.07804/) (2026)
- [Respecting Self-Uncertainty in On-Policy Self-Distillation for Efficient LLM Reasoning](https://mlsanity.com/paper/2605.13255/) (2026)
- [On-Policy Self-Distillation without Any Supervision](https://mlsanity.com/paper/2608.06296/) (2026)
- [Prefix Teach, Suffix Fade: Local Teachability Collapse in Strong-to-Weak On-Policy Distillation](https://mlsanity.com/paper/2605.13643/) (2026)
- [On-Policy Delta Distillation](https://mlsanity.com/paper/2607.15161/) (2026)
- [ViCuR: Visual Cues as Recoverable Privilege for Multimodal On-Policy Distillation](https://mlsanity.com/paper/2606.05718/) (2026)
- [Beyond GRPO and On-Policy Distillation: An Empirical Sparse-to-Dense Reward Principle for Language-Model Post-Training](https://mlsanity.com/paper/2605.12483/) (2026)
- [OPRD: On-Policy Representation Distillation](https://mlsanity.com/paper/2606.06021/) (2026)
- [Training with Harnesses: On-Policy Harness Self-Distillation for Complex Reasoning](https://mlsanity.com/paper/2605.08741/) (2026)
- [Crosslingual On-Policy Self-Distillation for Multilingual Reasoning](https://mlsanity.com/paper/2605.09548/) (2026)
- [Tailoring Teaching to Aptitude: Direction-Adaptive Self-Distillation for LLM Reasoning](https://mlsanity.com/paper/2605.22263/) (2026)
- [From Generic Correlation to Input-Specific Credit in On-Policy Self Distillation](https://mlsanity.com/paper/2605.11613/) (2026)
- [π-Play: Multi-Agent Self-Play via Privileged Self-Distillation without External Data](https://mlsanity.com/paper/2604.14054/) (2026)
- [When Should the Teacher Move? Temporal Coupling and Stability in Self On-Policy Distillation](https://mlsanity.com/paper/2606.03532/) (2026)
- [Answer-Conditioned Chains of Thought Degrade Verifiable-Reasoning Distillation in Large Language Models](https://mlsanity.com/paper/2607.14552/) (2026)
- [Large Language Model Post-Training: A Unified View of Off-Policy and On-Policy Learning](https://mlsanity.com/paper/2604.07941/) (2026)
- [OPSD Compresses What RLVR Teaches: A Post-RL Compaction Stage for Reasoning Models](https://mlsanity.com/paper/2605.06188/) (2026)
- [A Formula-Driven Survey and Research Agenda for On-Policy Distillation](https://mlsanity.com/paper/2606.22793/) (2026)
- [TRIAGE: Role-Typed Credit Assignment for Agentic Reinforcement Learning](https://mlsanity.com/paper/2606.32017/) (2026)
- [DemoPSD: Disagreement-Modulated Policy Self-Distillation](https://mlsanity.com/paper/2607.02502/) (2026)
- [Healthcare AI GYM for Medical Agents](https://mlsanity.com/paper/2605.02943/) (2026)
- [boldsymbolf-OPD: Stabilizing Long-Horizon On-Policy Distillation with Freshness-Aware Control](https://mlsanity.com/paper/2605.17862/) (2026)
- [Z-Reward: Beyond Scalar Rewards by Internalizing Reasoning into Score Distributions](https://mlsanity.com/paper/2606.09076/) (2026)
- [LLM-as-a-Coach: Experiential Learning for Non-Verifiable Tasks](https://mlsanity.com/paper/2607.18110/) (2026)
- [Contrastive On-Policy Distillation](https://mlsanity.com/paper/2607.19046/) (2026)
- [Reasoning Compression with Mixed-Policy Distillation](https://mlsanity.com/paper/2605.08776/) (2026)

## Abstract (arXiv)

Reasoning models often generate far more tokens than a task requires, which raises inference cost and can compound errors. We introduce CRISP (Compressed Reasoning via Iterative Self-Policy Distillation), an on-policy self-distillation method that teaches a model to reason more concisely by distilling its own concise behavior back into itself. The method uses a single idea: condition the same model on a "be concise" instruction to obtain teacher logits, then minimize the per-token reverse KL divergence between the student and this teacher on the student's own rollouts. It requires no ground-truth answers, no token budgets, and no difficulty estimators. The reverse-KL objective is naturally difficulty-adaptive: it compresses easy problems aggressively while preserving the reasoning steps that hard problems require. On Qwen3-14B, CRISP cuts reasoning length by up to 56% on MATH-500 and 38% on the harder AIME 2024, while improving MATH-500 accuracy by up to 3.3 points over the base model and holding AIME 2024 accuracy within about one point. This behavior generalizes across model sizes and families: Qwen3-8B shows the same compression with accuracy preserved, and DeepSeek-R1-Distill-Llama-8B improves accuracy on all five benchmarks while shortening its responses. General capabilities are preserved across all three models. Code is available at https://github.com/HJSang/OPSD_Reasoning_Compression.