# Stepwise Penalization for Length-Efficient Chain-of-Thought Reasoning

arXiv: https://arxiv.org/abs/2603.00296 · HTML: https://mlsanity.com/paper/2603.00296/

## Key facts

- Cited by 3 papers in the map (data through 14 Oct 2026).
- Placed in Test-time compute and efficient thinking, one of 17 directions in LLM reasoning.
- Builds on 28 of its 39 references that are in the map.
- First posted on arXiv 27 Feb 2026.

## Placed in

- LLM reasoning / [Test-time compute and efficient thinking](https://mlsanity.com/lr/test-time-compute-and-efficient-thinking/)

## Builds on (most cited first, 28 of 28)

- [GPT-4 Technical Report](https://mlsanity.com/paper/2303.08774/) (2023)
- [DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learning](https://mlsanity.com/paper/2501.12948/) (2025)
- [Measuring Mathematical Problem Solving With the MATH Dataset](https://mlsanity.com/paper/2103.03874/) (2021)
- [Stop Overthinking: A Survey on Efficient Reasoning for Large Language Models](https://mlsanity.com/paper/2503.16419/) (2025)
- [L1: Controlling How Long A Reasoning Model Thinks With Reinforcement Learning](https://mlsanity.com/paper/2503.04697/) (2025)
- [Omni-MATH: A Universal Olympiad Level Mathematic Benchmark For Large Language Models](https://mlsanity.com/paper/2410.07985/) (2024)
- [O1-Pruner: Length-Harmonizing Fine-Tuning for O1-Like Reasoning Pruning](https://mlsanity.com/paper/2501.12570/) (2025)
- [ThinkPrune: Pruning Long Chain-of-Thought of LLMs via Reinforcement Learning](https://mlsanity.com/paper/2504.01296/) (2025)
- [Imitate, Explore, and Self-Improve: A Reproduction Report on Slow-thinking Reasoning Systems](https://mlsanity.com/paper/2412.09413/) (2024)
- [AdaptThink: Reasoning Models Can Learn When to Think](https://mlsanity.com/paper/2505.13417/) (2025)
- [Thought Anchors: Which LLM Reasoning Steps Matter?](https://mlsanity.com/paper/2506.19143/) (2025)
- [Efficient Reasoning Models: A Survey](https://mlsanity.com/paper/2504.10903/) (2025)
- [Learn to Reason Efficiently with Adaptive Length-based Reward Shaping](https://mlsanity.com/paper/2505.15612/) (2025)
- [Guiding Language Model Reasoning with Planning Tokens](https://mlsanity.com/paper/2310.05707/) (2023)
- [Just Enough Thinking: Efficient Reasoning with Adaptive Length Penalties Reinforcement Learning](https://mlsanity.com/paper/2506.05256/) (2025)
- [Reasoning Language Models: A Blueprint](https://mlsanity.com/paper/2501.11223/) (2025)
- [Optimizing Length Compression in Large Reasoning Models](https://mlsanity.com/paper/2506.14755/) (2025)
- [Making Slow Thinking Faster: Compressing LLM Chain-of-Thought via Step Entropy](https://mlsanity.com/paper/2508.03346/) (2025)
- [NaturalThoughts: Selecting and Distilling Reasoning Traces for General Reasoning Tasks](https://mlsanity.com/paper/2507.01921/) (2025)
- [Walk Before You Run! Concise LLM Reasoning via Reinforcement Learning](https://mlsanity.com/paper/2505.21178/) (2025)
- [Explainable Chain-of-Thought Reasoning: An Empirical Analysis on State-Aware Reasoning Dynamics](https://mlsanity.com/paper/2509.00190/) (2025)
- [CTRLS: Chain-of-Thought Reasoning via Latent State-Transition](https://mlsanity.com/paper/2507.08182/) (2025)
- [SABER: Switchable and Balanced Training for Efficient LLM Reasoning](https://mlsanity.com/paper/2508.10026/) (2025)
- [AALC: Large Language Model Efficient Reasoning via Adaptive Accuracy-Length Control](https://mlsanity.com/paper/2506.20160/) (2025)
- [Hierarchical Budget Policy Optimization for Adaptive Reasoning](https://mlsanity.com/paper/2507.15844/) (2025)
- [SmartThinker: Learning to Compress and Preserve Reasoning by Step-Level Length Control](https://mlsanity.com/paper/2507.04348/) (2025)
- [Masked-and-Reordered Self-Supervision for Reinforcement Learning from Verifiable Rewards](https://mlsanity.com/paper/2511.17473/) (2025)
- [Mitigating Overthinking through Reasoning Shaping](https://mlsanity.com/paper/2510.09535/) (2025)

## Built on by (most cited first, 3 of 3)

- [ExpThink: Experience-Guided Reinforcement Learning for Adaptive Chain-of-Thought Compression](https://mlsanity.com/paper/2605.07501/) (2026)
- [Better Starts, Better Ends: Bootstrapped Iterative Self-Reasoning Distillation for Compressed Reasoning](https://mlsanity.com/paper/2607.15736/) (2026)
- [Giving Credit Where It's Due: Redundancy-Aware Learning for Efficient Reasoning](https://mlsanity.com/paper/2609.27156/) (2026)

## Abstract (arXiv)

Large reasoning models improve with more test-time computation, but often overthink, producing unnecessarily long chains-of-thought that raise cost without improving accuracy. Prior reinforcement learning approaches typically rely on a single outcome reward with trajectory-level length penalties, which cannot distinguish essential from redundant reasoning steps and therefore yield blunt compression. Although recent work incorporates step-level signals, such as offline pruning, supervised data construction, or verifier-based intermediate rewards, reasoning length is rarely treated as an explicit step-level optimization objective during RL. We propose Step-wise Adaptive Penalization (SWAP), a fine-grained framework that allocates length reduction across steps based on intrinsic contribution. We estimate step importance from the model's on-policy log-probability improvement toward the correct answer, then treat excess length as a penalty mass redistributed to penalize low-importance steps more heavily while preserving high-importance reasoning. We optimize with a unified outcome-process advantage within group-relative policy optimization. Extensive experiments demonstrate that SWAP reduces reasoning length by 64.3% on average while improving accuracy by 5.7% relative to the base model.