给推理链打时间分，让大模型思考更靠谱。
Chronos: Learning Temporal Dynamics of Reasoning Chains for Test-Time Scaling
- 把推理过程当时间序列建模，动态评估每步质量。
- 在HMMT25上比单次推理提升34.21%，比投票法提升22.70%。
- 轻量插件式设计，适配各类大模型，计算开销极低。
测试时扩展（Test-Time Scaling, TTS）已成为提升大语言模型推理能力的有效范式。然而，现有方法（如多数投票和启发式词元级评分）对推理轨迹或词元一视同仁，易受轨迹质量波动和局部逻辑错误影响。本文提出轻量级、即插即用的时序推理评分器Chronos，将每条推理轨迹建模为时间序列，学习词元概率等轨迹特征，生成质量评分并采用加权投票机制。在领域内与领域外基准上的广泛评估表明，Chronos在多种模型上持续带来显著提升，计算开销可忽略不计。特别地，在使用Qwen3-4B-Thinking-2507模型时，Chronos@128在HMMT25上相较Pass@1提升34.21%，相较Maj@128提升22.70%，充分验证其有效性。
原文摘要 · Abstract (English)
Test-Time Scaling (TTS) has emerged as an effective paradigm for improving the reasoning performance of large language models (LLMs). However, existing methods -- most notably majority voting and heuristic token-level scoring -- treat reasoning traces or tokens equally, thereby being susceptible to substantial variations in trajectory quality and localized logical failures. In this work, we introduce \textbf{Chronos}, a lightweight and plug-and-play chronological reasoning scorer that models each trajectory as a time series. Specifically, Chronos learns to capture trajectory features of token probabilities, assigns quality scores accordingly, and employs a weighted voting mechanism. Extensive evaluations on both in-domain and out-of-domain benchmarks demonstrate that Chronos consistently delivers substantial gains across a variety of models, with negligible computational overhead. Notably, Chronos@128 achieves relative improvements of 34.21\% over Pass@1 and 22.70\% over Maj@128 on HMMT25 using Qwen3-4B-Thinking-2507, highlighting its effectiveness.
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。