Large Reasoning Models (LRMs) have demonstrated remarkable capabilities by scaling up the length of Chain-of-Thought (CoT). However, excessively long reasoning traces pose substantial challenges for training cost and inference latency. While various CoT compression approaches have emerged to address this challenge, they face inherent trade-offs: token-level methods often disrupt syntactic and logical coherence, while step-level methods based on perplexity fail to reliably capture the logically critical reasoning steps because of the dilution of logical information. In this paper, we propose ASAP (Anchor-guided, SurprisAl-based Pruning), a novel coarse-to-fine framework for CoT compression. ASAP first performs anchor-guided pruning to preserve the core reasoning structure, which efficiently reduces the search space for subsequent processing. Leveraging the insight that logical branching choices are concentrated at the onset of reasoning steps, it then enables logic-aware pruning by selecting logically essential reasoning steps based on a novel first-token surprisal metric. Finally, ASAP distills the models to autonomously generate and leverage these concise CoTs at inference time, enabling efficient reasoning. Experiments show that ASAP achieves state-of-the-art accuracy across multiple benchmarks while substantially reducing training and inference costs.

翻译：大型推理模型（LRMs）通过扩展思维链（CoT）的长度展现了卓越的能力。然而，过长的推理轨迹对训练成本和推理延迟构成了重大挑战。尽管已有多种CoT压缩方法应对这一挑战，但它们面临固有的权衡：词元级方法往往会破坏句法和逻辑连贯性，而基于困惑度的步骤级方法由于逻辑信息被稀释，无法可靠地捕捉逻辑上关键的推理步骤。本文提出ASAP（锚点引导、基于意外度的剪枝），一种新颖的从粗到精的CoT压缩框架。ASAP首先执行锚点引导剪枝以保留核心推理结构，从而高效减少后续处理的搜索空间。利用逻辑分支选择集中在推理步骤起始点的洞见，该框架随后通过基于一种新颖的首词意外度度量来选择逻辑上必需的推理步骤，从而实现逻辑感知剪枝。最后，ASAP通过蒸馏使模型能够在推理时自主生成并利用这些简洁的CoT，从而实现高效推理。实验表明，ASAP在多个基准测试中达到了最先进的准确率，同时显著降低了训练和推理成本。