# Skip-Thinking: Chunk-wise Chain-of-Thought Distillation Enable Smaller Language Models to Reason Better and Faster

arXiv: https://arxiv.org/abs/2505.18642 · HTML: https://mlsanity.com/paper/2505.18642/

## Key facts

- Cited by 16 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 14 of its 36 references that are in the map.
- First posted on arXiv 24 May 2025.

## 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, 14 of 14)

- [The Llama 3 Herd of Models](https://mlsanity.com/paper/2407.21783/) (2024)
- [Llama 2: Open Foundation and Fine-Tuned Chat Models](https://mlsanity.com/paper/2307.09288/) (2023)
- [Training Verifiers to Solve Math Word Problems](https://mlsanity.com/paper/2110.14168/) (2021)
- [Beyond the Imitation Game: Quantifying and extrapolating the capabilities of language models](https://mlsanity.com/paper/2206.04615/) (2022)
- [Training Large Language Models to Reason in a Continuous Latent Space](https://mlsanity.com/paper/2412.06769/) (2024)
- [Teaching Small Language Models to Reason](https://mlsanity.com/paper/2212.08410/) (2022)
- [Think before you speak: Training Language Models With Pause Tokens](https://mlsanity.com/paper/2310.02226/) (2023)
- [From Explicit CoT to Implicit CoT: Learning to Internalize CoT Step by Step](https://mlsanity.com/paper/2405.14838/) (2024)
- [Navigate through Enigmatic Labyrinth A Survey of Chain of Thought Reasoning: Advances, Frontiers and Future](https://mlsanity.com/paper/2309.15402/) (2023)
- [Implicit Chain of Thought Reasoning via Knowledge Distillation](https://mlsanity.com/paper/2311.01460/) (2023)
- [PaD: Program-aided Distillation Can Teach Small Models Reasoning Better than Chain-of-thought Fine-tuning](https://mlsanity.com/paper/2305.13888/) (2023)
- [Turning Dust into Gold: Distilling Complex Reasoning Capabilities from LLMs by Leveraging Negative Data](https://mlsanity.com/paper/2312.12832/) (2023)
- [Distilling Reasoning Ability from Large Language Models with Adaptive Thinking](https://mlsanity.com/paper/2404.09170/) (2024)
- [Study on the Large Batch Size Training of Neural Networks Based on the Second Order Gradient](https://mlsanity.com/paper/2012.08795/) (2020)

## Built on by (most cited first, 16 of 16)

- [Purified OPSD: On-Policy Self-Distillation Without Losing How to Think](https://mlsanity.com/paper/2607.02234/) (2026)
- [AgentDistill: Training-Free Agent Distillation with Generalizable MCP Boxes](https://mlsanity.com/paper/2506.14728/) (2025)
- [From Efficiency to Adaptivity: A Deeper Look at Adaptive Reasoning in Large Language Models](https://mlsanity.com/paper/2511.10788/) (2025)
- [Backtracking When It Strays: Mitigating Dual Exposure Biases in LLM Reasoning Distillation](https://mlsanity.com/paper/2605.19433/) (2026)
- [Merge-of-Thought Distillation](https://mlsanity.com/paper/2509.08814/) (2025)
- [Putting on the Thinking Hats: A Survey on Chain of Thought Fine-tuning from the Perspective of Human Reasoning Mechanism](https://mlsanity.com/paper/2510.13170/) (2025)
- [RedAct: Redacting Agent Capability Traces for Procedural Skill Protection](https://mlsanity.com/paper/2606.10813/) (2026)
- [ImgCoT: Compressing Long Chain of Thought into Compact Visual Tokens for Efficient Reasoning of Large Language Model](https://mlsanity.com/paper/2601.22730/) (2026)
- [Are Rationales Necessary and Sufficient? Tuning LLMs for Explainable Misinformation Detection](https://mlsanity.com/paper/2605.19285/) (2026)
- [Learning through Internalization](https://mlsanity.com/paper/2606.20937/) (2026)
- [Adaptive Confidence Gating in Multi-Agent Collaboration for Efficient and Optimized Code Generation](https://mlsanity.com/paper/2601.21469/) (2026)
- [Structural Rationale Distillation via Reasoning Space Compression](https://mlsanity.com/paper/2605.07139/) (2026)
- [Breaking the Reward Barrier: Accelerating Tree-of-Thought Reasoning via Speculative Exploration](https://mlsanity.com/paper/2605.10195/) (2026)
- [Learning What to Learn: Stage-Specific Data Sets for SFT-then-RL in Small Language Model Reasoning](https://mlsanity.com/paper/2606.04466/) (2026)
- [Characterize Then Distill: Mechanistic Reasoning in Large Output Spaces](https://mlsanity.com/paper/2606.06840/) (2026)
- [Progress-Aware Distillation for Mitigating Stagnation in Small Language Model Agents](https://mlsanity.com/paper/2610.19439/) (2026)

## Abstract (arXiv)

Chain-of-thought (CoT) distillation allows a large language model (LLM) to guide a small language model (SLM) in reasoning tasks. Existing methods train the SLM to learn the long rationale in one iteration, resulting in two issues: 1) Long rationales lead to a large token-level batch size during training, making gradients of core reasoning tokens (i.e., the token will directly affect the correctness of subsequent reasoning) over-smoothed as they contribute a tiny fraction of the rationale. As a result, the SLM converges to sharp minima where it fails to grasp the reasoning logic. 2) The response is slow, as the SLM must generate a long rationale before reaching the answer. Therefore, we propose chunk-wise training (CWT), which uses a heuristic search to divide the rationale into internal semantically coherent chunks and focuses SLM on learning from only one chunk per iteration. In this way, CWT naturally isolates non-reasoning chunks that do not involve the core reasoning token (e.g., summary and transitional chunks) from the SLM learning for reasoning chunks, making the fraction of the core reasoning token increase in the corresponding iteration. Based on CWT, skip-thinking training (STT) is proposed. STT makes the SLM automatically skip non-reasoning medium chunks to reach the answer, improving reasoning speed while maintaining accuracy. We validate our approach on a variety of SLMs and multiple reasoning tasks.