让大模型按难易程度智能剪枝，既省计算又不丢逻辑。
SAT: Balancing Reasoning Accuracy and Efficiency with Stepwise Adaptive Thinking
- 用状态机动态管理推理步骤，难的深思，易的快过。
- 在7个评测上减少40%推理词数，准确率基本不变或提升。
- 适合追求效率又怕出错的AI系统开发者使用。
大型推理模型虽大幅提升复杂问题求解能力，但普遍存在“过度思考”现象，生成冗长推理链。现有方法虽提升了词元效率，却常牺牲细粒度控制或破坏推理逻辑完整性。为此，本文提出分步自适应思维（SAT）框架，实现基于难度感知的步骤级剪枝，同时保持核心推理结构。SAT将推理建模为有限状态机（FSM），包含慢、常、快、跳过四种思维模式，通过轻量级过程奖励模型（PRM）动态切换状态，对简单步骤压缩，对困难步骤保留深度。在9个大型推理模型和7个基准测试上的实验表明，SAT可实现最高40%的推理词元减少，且准确率普遍维持或提升。
原文摘要 · Abstract (English)
Large Reasoning Models (LRMs) have revolutionized complex problem-solving, yet they exhibit a pervasive "overthinking", generating unnecessarily long reasoning chains. While current solutions improve token efficiency, they often sacrifice fine-grained control or risk disrupting the logical integrity of the reasoning process. To address this, we introduce Stepwise Adaptive Thinking (SAT), a framework that performs step-level, difficulty-aware pruning while preserving the core reasoning structure. SAT formulates reasoning as a Finite-State Machine (FSM) with distinct thinking modes (Slow, Normal, Fast, Skip). It navigates these states dynamically using a lightweight Process Reward Model (PRM), compressing easy steps while preserving depth for hard ones. Experiments across 9 LRMs and 7 benchmarks show that SAT achieves up to 40% reduction in reasoning tokens while generally maintaining or improving 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。