CAC-CoT: Connector-Aware Compact Chain-of-Thought for Efficient Reasoning Data Synthesis Across Dual-System Cognitive Tasks
On the map: LLM reasoning → Prompting and chain-of-thought reasoning
also in 1 more
Abstract (arXiv)
Long chain-of-thought (CoT) prompting helps Large Language Models (LLMs) solve difficult problems, but very long traces often slow or even degrade performance on fast, intuitive "System-1" tasks. We introduce Connector-Aware Compact CoT (CAC-CoT) -- a method that deliberately restricts reasoning to a small, fixed set of connector phrases, steering the model toward concise and well -- structured explanations. Despite its simplicity, our synthetic method with general-purpose LLMs yields a high-quality training quality. CAC-CoT achieves approximately 85% on GSM8K and approximately 40% on GPQA (System-2) while also achieving approximately 85% on S1-Bench (System-1), surpassing the baseline by over 20%. Its reasoning traces average approximately 300 tokens(ART), about one-third the length of baseline traces, delivering higher efficiency without loss of accuracy.
Key facts. Cited by 1 paper 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 20 of its 39 references that are in the map. First posted on arXiv 26 Aug 2025.
Builds on
20 of 39 references.
Closest ancestors · 7
Same direction, newest first
- 2025Towards Reasoning Era: A Survey of Long Chain-of-Thought for Reasoning Large Language Models12 Mar
- 2025Small Models Struggle to Learn from Strong Reasoners17 Feb
- 2025LLMs Can Easily Learn to Reason from Demonstrations Structure, not content, is what matters!11 Feb
- 2025When More is Less: Understanding Chain-of-Thought Length in LLMs11 Feb
- 2025RedStar: Does Scaling Long-CoT Data Unlock Better Slow-Reasoning Systems?20 Jan
- 2024SIKeD: Self-guided Iterative Knowledge Distillation for mathematical reasoning24 Oct
- 2023Beyond Chain-of-Thought, Effective Graph-of-Thought Reasoning in Language Models26 May
Other references · 13
- 2025DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learningcited by 6,330
- 2024GPT-4o System Cardcited by 4,588
- 2024Scaling LLM Test-Time Compute Optimally can be More Effective than Scaling Model Parameterscited by 1,986
- 2025s1: Simple test-time scalingcited by 1,341
- 2024OpenAI o1 System Cardcited by 1,244
- 2024Do NOT Think That Much for 2+3=? On the Overthinking of o1-Like LLMscited by 570
- 2025LIMO: Less is More for Reasoningcited by 498
- 2025Stop Overthinking: A Survey on Efficient Reasoning for Large Language Modelscited by 464
- 2020Dataset Cartography: Mapping and Diagnosing Datasets with Training Dynamicscited by 425
- 2025LIMR: Less is More for RL Scalingcited by 139
- 2025Phi-4-Mini-Reasoning: Exploring the Limits of Small Reasoning Language Models in Mathcited by 44
- 2025Long Is More Important Than Difficult for Training Reasoning Modelscited by 11
1 more
Built on by
Citing papers per year: 2026: 1.
1 paper.
Similar, not cited
Close in wording, no citation either way
- 2025Scaling Code-Assisted Chain-of-Thoughts and Instructions for Model Reasoning
- 2025SCOUT: Teaching Pre-trained Language Models to Enhance Reasoning via Flow Chain-of-Thought
- 2025The Challenge of Teaching Reasoning to LLMs Without RL or Distillation
- 2025CoT Vectors: Transferring and Probing the Reasoning Mechanisms of LLMs
- 2026D-COT: Disciplined Chain-of-Thought Learning for Efficient Reasoning in Small Language Models
- 2026CtrlCoT: Dual-Granularity Chain-of-Thought Compression for Controllable Reasoning
- 2026S3-CoT: Self-Sampled Succinct Reasoning Enables Efficient Chain-of-Thought LLMs
- 2024Break the Chain: Large Language Models Can be Shortcut Reasoners