且慢，我们无需“等待”！移除思维标记可提升推理效率

Wait, We Don't Need to "Wait"! Removing Thinking Tokens Improves Reasoning Efficiency

June 10, 2025

作者: Chenlong Wang, Yuanning Feng, Dongping Chen, Zhaoyang Chu, Ranjay Krishna, Tianyi Zhou

cs.AI

摘要

近期大型推理模型的进展已能实现复杂的逐步推理，但往往伴随显著的过度思考，导致冗长冗余的输出，影响效率。本研究探讨了以“等待”和“嗯”等标记为信号的外显自我反思是否对高级推理必不可少。我们提出了NoWait方法，这一简洁而有效的策略通过在推理过程中抑制这些标记来禁用外显自我反思。在涵盖文本、视觉及视频推理任务的十项基准测试中，广泛实验表明，NoWait在五个R1系列模型上能将思维链轨迹长度减少27%至51%，且不损害模型效用。因此，NoWait为高效且保持效用的多模态推理提供了一种即插即用的解决方案。

English

Recent advances in large reasoning models have enabled complex, step-by-step reasoning but often introduce significant overthinking, resulting in verbose and redundant outputs that hinder efficiency. In this study, we examine whether explicit self-reflection, signaled by tokens such as "Wait" and "Hmm", is necessary for advanced reasoning. We propose NoWait, a simple yet effective approach that disables explicit self-reflection by suppressing these tokens during inference. Extensive experiments on ten benchmarks across textual, visual, and video reasoning tasks show that NoWait reduces chain-of-thought trajectory length by up to 27%-51% in five R1-style model series, without compromising model utility. NoWait thus offers a plug-and-play solution for efficient and utility-preserving multimodal reasoning.