Salvato in:
| Autori principali: | , , , , , , , , , , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2025
|
| Soggetti: | |
| Accesso online: | https://arxiv.org/abs/2505.19756 |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866912395176181760 |
|---|---|
| author | Gong, Ruihan Liu, Yue Qu, Wenjie Du, Mingzhe He, Yufei Ma, Yingwei Chen, Yulin Liu, Xiang Wen, Yi Li, Xinfeng Wang, Ruidong Zhu, Xinzhong Hooi, Bryan Zhang, Jiaheng |
| author_facet | Gong, Ruihan Liu, Yue Qu, Wenjie Du, Mingzhe He, Yufei Ma, Yingwei Chen, Yulin Liu, Xiang Wen, Yi Li, Xinfeng Wang, Ruidong Zhu, Xinzhong Hooi, Bryan Zhang, Jiaheng |
| contents | Large Reasoning Models (LRMs) achieve promising performance but compromise token efficiency due to verbose reasoning processes. Unconscious Thought Theory (UTT) posits that complex problems can be solved more efficiently through internalized cognitive processes. Inspired by UTT, we propose a new reasoning paradigm, termed Chain of Unconscious Thought (CoUT), to improve the token efficiency of LRMs by guiding them to mimic human unconscious thought and internalize reasoning processes. Concretely, we first prompt the model to internalize the reasoning by thinking in the hidden layer. Then, we design a bag of token-efficient strategies to further help models reduce unnecessary tokens yet preserve the performance. Our work reveals that models may possess beneficial unconscious thought, enabling improved efficiency without sacrificing performance. Extensive experiments demonstrate the effectiveness of CoUT. Remarkably, it surpasses CoT by reducing token usage by 47.62% while maintaining comparable accuracy, as shown in Figure 1. The code of CoUT is available at this link: https://github.com/Rohan-GRH/CoUT |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2505_19756 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Efficient Reasoning via Chain of Unconscious Thought Gong, Ruihan Liu, Yue Qu, Wenjie Du, Mingzhe He, Yufei Ma, Yingwei Chen, Yulin Liu, Xiang Wen, Yi Li, Xinfeng Wang, Ruidong Zhu, Xinzhong Hooi, Bryan Zhang, Jiaheng Computation and Language Large Reasoning Models (LRMs) achieve promising performance but compromise token efficiency due to verbose reasoning processes. Unconscious Thought Theory (UTT) posits that complex problems can be solved more efficiently through internalized cognitive processes. Inspired by UTT, we propose a new reasoning paradigm, termed Chain of Unconscious Thought (CoUT), to improve the token efficiency of LRMs by guiding them to mimic human unconscious thought and internalize reasoning processes. Concretely, we first prompt the model to internalize the reasoning by thinking in the hidden layer. Then, we design a bag of token-efficient strategies to further help models reduce unnecessary tokens yet preserve the performance. Our work reveals that models may possess beneficial unconscious thought, enabling improved efficiency without sacrificing performance. Extensive experiments demonstrate the effectiveness of CoUT. Remarkably, it surpasses CoT by reducing token usage by 47.62% while maintaining comparable accuracy, as shown in Figure 1. The code of CoUT is available at this link: https://github.com/Rohan-GRH/CoUT |
| title | Efficient Reasoning via Chain of Unconscious Thought |
| topic | Computation and Language |
| url | https://arxiv.org/abs/2505.19756 |