AlphaOne：在测试时进行慢速与快速思考的推理模型
AlphaOne: Reasoning Models Thinking Slow and Fast at Test Time
AI总结：
本文提出通用框架AlphaOne，通过引入$α$时刻并结合随机过程动态调度慢速思考与快速推理，在测试时有效调节大型推理模型的推理进度，提升了推理能力与效率。
摘要：
本文提出了AlphaOne（$α$1），一个在测试时调节大型推理模型（LRMs）推理进度的通用框架。$α$1首先引入了$α$时刻，表示由通用参数$α$缩放的思考阶段。在这个缩放的前$α$时刻阶段，它通过将推理转换token的插入建模为伯努利随机过程，来动态调度慢速思考转换。在$α$时刻之后，$α$1使用思考结束token确定性地终止慢速思考，从而促进快速推理和高效的答案生成。该方法通过实现灵活且密集的慢速到快速推理调节，统一并推广了现有的单调缩放方法。在数学、编程和科学领域各种具有挑战性的基准上的大量实证研究证明了$α$1卓越的推理能力和效率。项目页面：https://alphaone-project.github.io/
英文摘要：
This paper presents AlphaOne ($α$1), a universal framework for modulating reasoning progress in large reasoning models (LRMs) at test time. $α$1 first introduces $α$ moment, which represents the scaled thinking phase with a universal parameter $α$. Within this scaled pre-$α$ moment phase, it dynamically schedules slow thinking transitions by modeling the insertion of reasoning transition tokens as a Bernoulli stochastic process. After the $α$ moment, $α$1 deterministically terminates slow thinking with the end-of-thinking token, thereby fostering fast reasoning and efficient answer generation. This approach unifies and generalizes existing monotonic scaling methods by enabling flexible and dense slow-to-fast reasoning modulation. Extensive empirical studies on various challenging benchmarks across mathematical, coding, and scientific domains demonstrate $α$1's superior reasoning capability and efficiency. Project page: https://alphaone-project.github.io/
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。
