Research map / Large language models / Agents, code and math
Decoupling Knowledge and Reasoning in LLMs: An Exploration Using Cognitive Dual-System Theory
Mutian Yang, Jiandong Gao, Ji Wu
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| slow thinking (CoT) instead of fast thinking | Qwen 1.5B · MMLU | accuracy: 53.9 -> 50.0 | Hurt |
| slow thinking (CoT) instead of fast thinking | Qwen 7B · MMLU | accuracy: 67.8 -> 71.8 | Improved |
| slow thinking (CoT) instead of fast thinking | LLaMA 1B · MMLU | accuracy: 35.1 -> 35.1 | No clear effect |
| slow thinking in mathematics domain | 15 LLMs · MMLU Mathematics | accuracy: average delta 19.0 | Improved |
| slow thinking in political science domain | 15 LLMs · MMLU Political Science | accuracy: average delta -3.0 | Hurt |
| scale LLaMA from 1B to 70B | overthinking rate · MMLU | ro (%): 52.6 -> 7.2 | Improved |
| scale LLaMA from 1B to 70B | correction rate · MMLU | rc (%): 29.4 -> 38.1 | Improved |
Closest papers by meaning
- Cognitive Mirrors: Exploring the Diverse Functional Roles of Attention Heads in LLM Reasoning (NeurIPS 2025 · similarity 0.90)
- Procedural Knowledge in Pretraining Drives Reasoning in Large Language Models (ICLR 2025 · similarity 0.90)
- Unveiling the Magic of Code Reasoning through Hypothesis Decomposition and Amendment (ICLR 2025 · similarity 0.90)
- RATT: A Thought Structure for Coherent and Correct LLM Reasoning (AAAI 2025 · similarity 0.89)
- The Illusion of Thinking: Understanding the Strengths and Limitations of Reasoning Models via the Lens of Problem Complexity (NeurIPS 2025 · similarity 0.89)
- Eliciting Reasoning in Language Models with Cognitive Tools (NeurIPS 2025 · similarity 0.89)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
