通过动态建模推理难度，减少大模型的冗余思考。
DyCon: Dynamic Reasoning Control via Evolving Difficulty Modeling
- 利用步骤级嵌入捕捉难度演化，实现无需训练的动态控制。
- 在12个任务上提升效率，4B到32B模型均有效，精度不降。
- 适合需要高效推理的数学、问答和编程场景。
大型推理模型在迭代反思、探索与执行复杂任务时表现优异，但存在冗余推理（过度思考）导致效率低下问题。现有方法依赖静态难度估计或需特定任务训练，难以适应推理过程中的动态复杂度变化。本文实证发现，任务难度在推理过程中动态演变，且线性编码于模型的步骤级嵌入中。基于此，提出DyCon框架，利用隐式步骤级表示显式建模动态难度，从而实现推理深度的动态调控，缓解过度思考问题。在4B至32B规模的四款模型上，跨12个基准测试（涵盖数学推理、通用问答和编程任务）的实验表明，DyCon显著提升推理效率，减少冗余步骤，同时保持精度与泛化能力。代码已开源：https://github.com/yu-lin-li/DyCon。
原文摘要 · Abstract (English)
Recent advances in Large Reasoning Models (LRMs) demonstrate remarkable performance improvements by iteratively reflecting, exploring, and executing complex tasks, yet suffer from inefficiencies due to redundant reasoning, known as "overthinking". Existing methods to mitigate this issue either rely on static difficulty estimates or require task-specific training, and thus fail to adapt to the dynamic complexity during reasoning. In this work, we empirically show that the problem difficulty evolves dynamically throughout the reasoning process and is linearly encoded in the LRM's step-level embeddings. Building on this insight, we propose DyCon, a training-free framework that leverages latent step-level representations to explicitly model the evolving task difficulty, enabling the dynamic control of reasoning depth to mitigate the overthinking issue. Extensive experiments conducted on four models ranging from 4B to 32B, and across twelve benchmarks in math reasoning, general question answering, and coding tasks demonstrate that DyCon significantly enhances reasoning efficiency by reducing redundant steps without sacrificing accuracy or generalization. Code is available at https://github.com/yu-lin-li/DyCon.
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。