Top AI Repos — open-source AI, indexed and scored
Top AI Repos tracks AI repositories on GitHub and answers two different questions about each one: is it moving right now, and would you bet a product on it.
Top AI Repos tracks AI repositories on GitHub and answers two different questions about each one: is it moving right now, and would you bet a product on it.
This repository contains a regularly updated paper list for LLMs-reasoning-in-latent-space.
| Date | Stars |
|---|---|
| 2026-07-31 | 367 |
| 2026-08-06 | 367 |
Today
— stars today
This week
— stars this week
This month
— stars this month
Momentum
0.0
growth rate 0.00%/day
<div align="center">
<h1><img src="assets/logo.png" height="32px"/> Reasoning Beyond Language: A Comprehensive Survey on Latent Chain-of-Thought Reasoning</h1>
</div>
<div align="center">
<b>Xinghao Chen</b><sup>1,2</sup>,
<b>Anhao Zhao</b><sup>2</sup>,
<b>Heming Xia</b><sup>1</sup>,
<b>Xuan Lu</b><sup>2</sup>,
<b>Hanlin Wang</b><sup>1</sup>,
</div>
<div align="center">
<b>Yanjun Chen</b><sup>1,2</sup>,
<b>Wei Zhang</b><sup>2</sup>,
<b>Jian Wang</b><sup>1</sup>,
<b>Wenjie Li</b><sup>1</sup>,
<b>Xiaoyu Shen</b><sup>2</sup>
</div>
<div align="center">
<sup>1</sup>Department of Computing, The Hong Kong Polytechnic University
</div>
<div align="center">
<sup>2</sup>Ningbo Digital Twin Institute, Eastern Institute of Technology, Ningbo, China
</div>
This repository contains a regularly updated paper list for **Latent CoT Reasoning**.
<p align="center">
<a href="https://arxiv.org/abs/2505.16782">
<img src="https://img.shields.io/badge/Arxiv-2505.16782-orange.svg"></a>
<a href="https://awesome.re">
<img src="https://awesome.re/badge.svg" alt="Awesome"></a>
<a href="https://opensource.org/licenses/Apache-2.0">
<img src="https://img.shields.io/badge/License-Apache_2.0-green.svg"></a>
<a href="https://github.com/EIT-NLP/Awesome-Latent-CoT/pulls">
<img src="https://img.shields.io/badge/Contributions-welcome-blue.svg?style=flat"></a>
<img src="https://img.shields.io/github/last-commit/EIT-NLP/Awesome-Latent-CoT/main?logo=github&color=blue" alt="Last Commit">
</p>
<p align="center"><em>
Whereof one cannot speak, thereof one must be silent.
-- Ludwig Wittgenstein
</em></p>
Reasoning in latent space shifts the way AI models think, moving beyond language tokens to represent thought processes in a more abstract, non-language space. Just as humans often think without words, latent space allows for more flexible and efficient reasoning.
### Key advantages include:
1. **Richer Thought Representation**: Latent space captures complex, non-verbal thoughts that language alone can't express.
2. **Lower Latency**: It allows for higher information density, reducing the need for token-based decoding and speeding up reasoning.
This approach brings AI closer to human-like cognition, enabling faster, more flexible, and powerful models for real-world tasks.
## Citation
If you find our survey useful for your research, please consider citing the following paper:
```bibtex
@article{eit2025latentcot,
title={Reasoning Beyond Language: A Comprehensive Survey on Latent Chain-of-Thought Reasoning},
author={Xinghao Chen and Anhao Zhao and Heming Xia and Xuan Lu and Hanlin Wang and Yanjun Chen and Wei Zhang and Jian Wang and Wenjie Li and Xiaoyu Shen},
year={2025},
eprint={2505.16782},
archivePrefix={arXiv},
primaryClass={cs.CL},
url={https://arxiv.org/abs/2505.16782},
}
```
## Updates
- 2025-05-22: 📝 The survey is now available on [arXiv](https://arxiv.org/abs/2505.16782)!
- 2025-02-16: 🚀 Latent CoT Repo launched!
## Content
- [Keywords Convention](#keywords-convention)
- [Papers](#papers)
- [Token-wise Strategies](#token-wise-strategies)
- [Discrete Tokens](#discrete-tokens)
- [Continuous Tokens](#continuous-tokens)
- [Internal Mechanisms](#internal-mechanisms)
- [Structural CoT](#structural-cot)
- [Representational CoT](#representational-cot)
- [Analysis and Interpretability](#analysis-and-interpretability)
- [Applications and Future Directions](#applications-and-future-directions)
- [Resources](#resources)
- [Acknowledgements](#acknowledgements)
## Keywords Convention
 Abbreviation
 Conference
 Main Features
## Papers
### Token-wise Strategies
#### Discrete Tokens
- **Think before you speak: Training language models with pause tokens**
*Sachin Goyal,ZiwExcerpt of 94,514 characters
Read on GitHubWould you bet a product on this? Bounded 0–100 and slow moving.
matched fp:59bff8e3cca089a8, llm:Repository description: 'contains a regularly updated paper list for LLMs-reasoning-in-latent-space' (awesome list of papers on latent chain-of-thought / LLM reasoning in latent space).
matched fp:59bff8e3cca089a8, llm:Repository description: 'contains a regularly updated paper list for LLMs-reasoning-in-latent-space' (awesome list of papers on latent chain-of-thought / LLM reasoning in latent space).
matched fp:59bff8e3cca089a8, llm:Repository description: 'contains a regularly updated paper list for LLMs-reasoning-in-latent-space' (awesome list of papers on latent chain-of-thought / LLM reasoning in latent space).