🚀 Awesome Long2Short Papers

A curated list of papers about making LLM reasoning more efficient (shorter/faster/better).

📰 News

🎉 Latest Update: Our paper collection is continuously updated. Feel free to contribute!

📝 Citation

If you find this resource helpful, please cite the relevant papers:

@misc{awesome-long2short-papers,
  title={Awesome Long2Short Papers},
  author={Community Contributors},
  year={2024},
  publisher={GitHub},
  howpublished={\url{https://github.com/yzhangchuck/awesome-llm-reasoning-long2short-papers}}
}

📚 Table of Contents

• 🔍 Analysis and Understanding
• 🤔 Reasoning Scaling
• ⚡ Inference Intervention
• 🧠 Latent Reasoning
• 📚 Supervised Fine-tuning
• 🎛️ Steering Vector
• 🎯 Reinforcement Learning
• 🌟 General Papers

---

🔍 Analysis and Understanding

Deep dives into LLM reasoning

Title	Year	Venue	Paper	Code
How Well do LLMs Compress Their Own Chain-of-Thought? A Token Complexity Approach	2025	arXiv preprint	[Paper]	[Code]
Towards Thinking-Optimal Scaling of Test-Time Compute for LLM Reasoning	2025	arXiv preprint	[Paper]	-
Do NOT Think That Much for 2+3=? On the Overthinking of o1-Like LLMs	2024	arXiv preprint	[Paper]	-

🤔 Reasoning Scaling

Scaling up reasoning

Title	Year	Venue	Paper	Code
Test Time Preference Optimization On the Fly Alignment via Iterative Textual Feedback	2025	arXiv preprint	[Paper]	[Code]
Beyond Human Data Scaling Self Training for Problem Solving with Language Models	2024	TMLR	[Paper]	-
REFT Reasoning with REinforced Fine Tuning	2024	ACL	[Paper]	[Code]
ReST MCTS∗ LLM Self Training via Process Reward Guided Tree Search	2024	NeurIPS	[Paper]	[Code]
Recursive Introspection Teaching Language Model Agents How to Self Improve	2024	NeurIPS	[Paper]	[Code]
Scaling LLM Test Time Compute Optimally can be More Effective than Scaling Model Parameters	2024	arXiv preprint	[Paper]	-
Let’s Verify Step by Step	2023	ICLR	[Paper]	-

⚡ Inference Intervention

Interventions during model inference process

Title	Year	Venue	Paper	Code
When More is Less- Understanding Chain-of-Thought Length in LLMs	2025	arXiv preprint	[Paper]	-

🧠 Latent Reasoning

Reasoning in latent spaces

Title	Year	Venue	Paper	Code
TwT: Thinking without Tokens by Habitual Reasoning Distillation with Multi-Teachers’ Guidance	2025	arXiv preprint	[Paper]	-
LightThinker- Thinking Step-by-Step Compression	2025	arXiv preprint	[Paper]	-
Scaling up Test-Time Compute with Latent Reasoning- A Recurrent Depth Approach	2025	arXiv preprint	[Paper]	[Code]
Token Assorted- Mixing Latent and Text Tokens for Improved Language Model Reasoning	2025	arXiv preprint	[Paper]	-
Compressed Chain of Thought- Efficient Reasoning through Dense Representations	2024	arXiv preprint	[Paper]	-
Quiet-STaR- Language Models Can Teach Themselves to Think Before Speaking	2024	arXiv preprint	[Paper]	-
Training Large Language Models to Reason in a Continuous Latent Space	2024	arXiv preprint	[Paper]	-

📚 Supervised Fine-tuning

Direct optimization for efficiency

Title	Year	Venue	Paper	Code
Self-Training Elicits Concise Reasoning in Large Language Models	2025	arXiv preprint	[Paper]	[Code]
TokenSkip- Controllable Chain-of-Thought Compression in LLMs	2025	arXiv preprint	[Paper]	[Code]
s1- Simple test-time scaling	2025	arXiv preprint	[Paper]	[Code]

🎛️ Steering Vector

Steering model behavior through vector manipulation

Title	Year	Venue	Paper	Code
CoT-Valve- Length-Compressible Chain-of-Thought Tuning	2025	arXiv preprint	[Paper]	-

🎯 Reinforcement Learning

Training models to reason efficiently

Title	Year	Venue	Paper	Code
Kimi k1.5: Scaling Reinforcement Learning with LLMs	2025	arXiv preprint	[Paper]	-
L1- Controlling How Long A Reasoning Model Thinks With Reinforcement Learning	2025	arXiv preprint	[Paper]	[Code]
O1-Pruner- Length-Harmonizing Fine-Tuning for O1-Like Reasoning Pruning	2024	arXiv preprint	[Paper]	[Code]
Training Language Models to Reason Efficiently	2024	arXiv preprint	[Paper]	[Code]

🌟 General Papers

Latest advances in efficient reasoning

Title	Year	Venue	Paper	Code
Inner Thinking Transformer- Leveraging Dynamic Depth Scaling to Foster Adaptive Internal Thinking	2025	arXiv preprint	[Paper]	-
Learning to Reason from Feedback at Test-Time	2025	arXiv preprint	[Paper]	[Code]
S*: Test Time Scaling for Code Generation	2025	arXiv preprint	[Paper]	[Code]
Efficiently Serving LLM Reasoning Programs with Certaindex	2024	arXiv preprint	[Paper]	-

🤝 Contributing

We welcome contributions! Please feel free to submit a pull request to add more papers or improve the existing content.

Contribution Guidelines

1. Please ensure the paper is related to efficient LLM reasoning
2. Follow the existing format
3. Add a brief description if possible
4. Include links to paper and code (if available)

📄 License

This project is licensed under the MIT License - see the LICENSE file for details.

⭐ Star History