通过动态抑制冗余推理步骤，让大模型高效又准确。
ARS: Adaptive Reasoning Suppression for Efficient Large Reasoning Language Models
- 基于自适应确定性监测，动态判断何时停止推理
- 在多个数学推理任务上实现最高57.9%能耗降低
- 无需训练，适配多种模型，兼顾速度与精度
大型推理语言模型在复杂推理任务中表现卓越，但因过度思考导致计算效率低下。现有高效推理方法难以平衡推理质量与推理成本。本文提出无需训练的自适应推理抑制（ARS）方法，通过多检查点确定性估计与渐进式抑制阈值，动态抑制冗余推理步骤，在不牺牲准确率的前提下提升效率。在多种模型架构和数学推理基准上的实验表明，ARS实现了最高53%的令牌数减少、46.1%的延迟降低和57.9%的能耗节省，同时保持或提升准确率。
原文摘要 · Abstract (English)
Large Reasoning Language Models (LRLMs or LRMs) demonstrate remarkable capabilities in complex reasoning tasks, but suffer from significant computational inefficiencies due to overthinking phenomena. Existing efficient reasoning methods face the challenge of balancing reasoning quality with inference cost reduction. We propose \textbf{Adaptive Reasoning Suppression (ARS)}, a novel training-free approach that dynamically suppresses redundant reasoning steps while preserving accuracy through adaptive certainty monitoring. ARS introduces a multi-checkpoint certainty estimation mechanism with progressive suppression thresholds, achieving superior efficiency compared to static suppression methods. Our extensive evaluation across mathematical reasoning benchmarks using multiple model architectures demonstrates that ARS achieves up to 53%, 46.1%, and 57.9% in token, latency and energy reduction, while 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。