Efficient Reasoning via Chain of Unconscious Thought
On the map: LLM reasoning → Prompting and chain-of-thought reasoning
Abstract (arXiv)
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
Key facts. Cited by 7 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 33 references that are in the map. First posted on arXiv 26 May 2025.
Builds on
14 of 33 references.
Closest ancestors · 2
Same direction, newest first
- 2025Sketch-of-Thought: Efficient LLM Reasoning with Adaptive Cognitive-Inspired Sketching7 Mar
- 2022Large Language Models are Zero-Shot Reasoners24 May
Other references · 12
- 2021Training Verifiers to Solve Math Word Problemscited by 10,109
- 2023Improving Factuality and Reasoning in Language Models through Multiagent Debatecited by 1,784
- 2023Encouraging Divergent Thinking in Large Language Models through Multi-Agent Debatecited by 1,011
- 2021Are NLP Models really able to Solve Simple Math Word Problems?cited by 878
- 2025O1-Pruner: Length-Harmonizing Fine-Tuning for O1-Like Reasoning Pruningcited by 279
- 2024ReFT: Reasoning with Reinforced Fine-Tuningcited by 239
- 2024Compressed Chain of Thought: Efficient Reasoning Through Dense Representationscited by 201
- 2024C3oT: Generating Shorter Chain-of-Thought without Compromising Effectivenesscited by 119
- 2024Concise Thoughts: Impact of Output Length on LLM Reasoning and Costcited by 109
- 2023At Which Training Stage Does Code Data Help LLMs Reasoning?cited by 82
- 2025Safety in Large Reasoning Models: A Surveycited by 64
- 2025Efficient Inference for Large Reasoning Models: A Surveycited by 61
Built on by
Citing papers per year: 2025: 4; 2026: 3.
7 papers.
- 2025Afterburner: Reinforcement Learning Facilitates Self-Improving Code Efficiency Optimizationcited by 18
- 2025Towards Concise and Adaptive Thinking in Large Reasoning Models: A Surveycited by 18
- 2026VTC-R1: Vision-Text Compression for Efficient Long-Context Reasoningcited by 8
- 2025The Zero-Step Thinking: An Empirical Study of Mode Selection as Harder Early Exit in Reasoning Modelscited by 4
- 2026QA-Merging: Query-Adaptive Reasoning via Layer Selective Model Mergingcited by 1
- 2025Dual-Process Scaffold Reasoning for Enhancing LLM Code Debuggingcited by 0
- 2026Collective Bias Mitigation via Model Routing and Collaborationcited by 0
- 2026Collective Bias Mitigation via Model Routing and Collaboration2 Oct 2026
- 2026VTC-R1: Vision-Text Compression for Efficient Long-Context Reasoning29 Jan 2026
- 2026QA-Merging: Query-Adaptive Reasoning via Layer Selective Model Merging7 Jan 2026
- 2025Dual-Process Scaffold Reasoning for Enhancing LLM Code Debugging11 Nov 2025
- 2025The Zero-Step Thinking: An Empirical Study of Mode Selection as Harder Early Exit in Reasoning Models22 Oct 2025
- 2025Towards Concise and Adaptive Thinking in Large Reasoning Models: A Survey13 Jul 2025
- 2025Afterburner: Reinforcement Learning Facilitates Self-Improving Code Efficiency Optimization29 May 2025
Similar, not cited
Close in wording, no citation either way
- 2022Chain-of-Thought Prompting Elicits Reasoning in Large Language Models
- 2025THINK-Bench: Evaluating Thinking Efficiency and Chain-of-Thought Quality of Large Reasoning Models
- 2025Deep Hidden Cognition Facilitates Reliable Chain-of-Thought Reasoning
- 2026Thinking Without Words: Efficient Latent Reasoning with Abstract Chain-of-Thought
- 2023Implicit Chain of Thought Reasoning via Knowledge Distillation
- 2025Beyond Chains of Thought: Benchmarking Latent-Space Reasoning Abilities in Large Language Models
- 2025Efficient Reasoning with Hidden Thinking
- 2025Focused Chain-of-Thought: Efficient LLM Reasoning via Structured Input Information