通过思维训练和无思维推理实现高效推理
Efficient Reasoning via Thought-Training and Thought-Free Inference
AI总结：
3TF通过训练混合模型实现高效推理，使模型能在无显式推理生成的情况下进行高质量推理。
摘要：
近期大型语言模型（LLMs）的进步利用了显式的链式思维（CoT）提示来提高推理准确性。然而，大多数现有方法主要集中在压缩冗长的推理输出上。这些长到短的转换旨在提高效率，但需要大量的短CoT数据。在本工作中，我们引入了3TF（思维训练和无思维推理）框架，这是一种从短到长视角的高效推理框架。我们首先训练一个混合模型，使其能够在推理和非推理模式之间操作，然后进一步在CoT标注的数据上训练它以内部化结构化的推理，同时在推理时强制使用无推理模式产生简洁的无思维输出。与压缩方法不同，3TF提高了非推理输出的推理质量，使模型能够在不进行显式逐步生成的情况下进行丰富的内部推理。实验表明，经过3TF训练的模型在无思维推理下在推理基准上取得了显著的改进，证明了高质量的推理可以被学习和执行，而无需显式的逐步生成。
英文摘要：
Recent advances in large language models (LLMs) have leveraged explicit Chain-of-Thought (CoT) prompting to improve reasoning accuracy. However, most existing methods primarily focus on compressing verbose reasoning outputs. These Long-to-Short transformations aim to improve efficiency, but require a large amount of short CoT data. In this work, we introduce \textbf{3TF} (\textbf{T}hought-\textbf{T}raining and \textbf{T}hought-\textbf{F}ree inference), a framework for efficient reasoning that takes a Short-to-Long perspective. We first train a hybrid model that can operate in both reasoning and non-reasoning modes, and then further train it on CoT-annotated data to internalize structured reasoning, while enforcing concise, thought-free outputs at inference time using the no-reasoning mode. Unlike compression-based approaches, 3TF improves the reasoning quality of non-reasoning outputs, enabling models to perform rich internal reasoning implicitly while keeping external outputs short. Empirically, 3TF-trained models obtain large improvements on reasoning benchmarks under thought-free inference, demonstrating that high quality reasoning can be learned and executed implicitly without explicit step-by-step generation.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
