CAC-CoT：面向双系统认知任务的连接词感知紧凑思维链高效推理数据合成
CAC-CoT: Connector-Aware Compact Chain-of-Thought for Efficient Reasoning Data Synthesis Across Dual-System Cognitive Tasks
AI总结：
针对长思维链降低System-1任务效率的问题，提出CAC-CoT方法，通过限制连接词短语生成简洁推理轨迹，在保持准确率的同时大幅提升效率并超越基线。
摘要：
长思维链（CoT）提示有助于大型语言模型（LLMs）解决难题，但过长的推理轨迹往往会使模型在快速、直觉的“System-1”任务上变慢甚至降低性能。我们提出了连接词感知紧凑CoT（CAC-CoT）——该方法刻意将推理限制在一小部分固定的连接词短语中，引导模型生成简洁且结构良好的解释。尽管方法简单，我们使用通用LLMs的合成方法产生了高质量的训练质量。CAC-CoT在GSM8K上达到约85%，在GPQA（System-2）上达到约40%，同时在S1-Bench（System-1）上也达到约85%，超越基线超过20%。其推理轨迹平均约300个token（ART），约为基线轨迹长度的三分之一，在不损失准确率的情况下提供了更高的效率。
英文摘要：
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.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
