用有限连接词生成精炼推理链，兼顾速度与准确率。
CAC-CoT: Connector-Aware Compact Chain-of-Thought for Efficient Reasoning Data Synthesis Across Dual-System Cognitive Tasks
- 限定推理使用固定连接词，生成简洁结构化过程
- GSM8K和S1-Bench达85%准确率，GPQA达40%
- 推理长度仅300词元，为基线1/3，适合双系统任务
长链式思维（CoT）提示能帮助大语言模型解决复杂问题，但过长的推理过程会拖慢甚至降低快速直觉型“系统-1”任务的表现。本文提出连接词感知的紧凑链式思维（CAC-CoT），通过严格限制推理中使用的连接词种类，引导模型生成简短且结构清晰的解释。尽管方法简单，但利用通用大模型合成的数据质量高。CAC-CoT在GSM8K和S1-Bench上分别达到约85%准确率，在GPQA（系统-2）上达约40%，均超越基线超过20%。其推理轨迹平均长度约为300词元（ART），仅为基线的三分之一，实现高效推理且不损失准确性。
原文摘要 · Abstract (English)
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.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。