用流匹配让大模型推理更简洁，效果优于传统方法。
Steering Large Reasoning Models towards Concise Reasoning via Flow Matching
- 通过流匹配学习输入相关的非线性变换，控制推理过程。
- 在多个基准上实现更短输出和更好任务表现，节省大量token。
- 适合需要高效推理的场景，如实时对话或资源受限部署。
大型推理模型在复杂任务中表现出色，但其效率常因输出过长而受阻。以往的引导方法采用单一全局向量作用于隐藏表示，依赖于限制性的线性表征假设。本文提出FlowSteer，一种非线性引导方法，通过流匹配学习冗长与简洁推理分布之间的完整变换，以速度场形式建模，实现对模型推理过程的精准、输入相关控制。通过对齐引导后的表示与简洁推理激活的分布，FlowSteer生成更紧凑的推理路径。在多个推理基准上，相比主流推理时基线方法，该方法展现出更强的任务性能与更高的词元效率。结果表明，利用生成技术建模完整的分布传输，为控制大型推理模型提供了更有效且原则化的新范式。
原文摘要 · Abstract (English)
Large Reasoning Models (LRMs) excel at complex reasoning tasks, but their efficiency is often hampered by overly verbose outputs. Prior steering methods attempt to address this issue by applying a single, global vector to hidden representations -- an approach grounded in the restrictive linear representation hypothesis. In this work, we introduce FlowSteer, a nonlinear steering method that goes beyond uniform linear shifts by learning a complete transformation between the distributions associated with verbose and concise reasoning. This transformation is learned via Flow Matching as a velocity field, enabling precise, input-dependent control over the model's reasoning process. By aligning steered representations with the distribution of concise-reasoning activations, FlowSteer yields more compact reasoning than the linear shifts. Across diverse reasoning benchmarks, FlowSteer demonstrates strong task performance and token efficiency compared to leading inference-time baselines. Our work demonstrates that modeling the full distributional transport with generative techniques offers a more effective and principled foundation for controlling LRMs.
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。