菜单

关于 🐙 GitHub
arXiv 提交日期: 2025-12-19
📄 Abstract - When Reasoning Meets Its Laws

Despite the superior performance of Large Reasoning Models (LRMs), their reasoning behaviors are often counterintuitive, leading to suboptimal reasoning capabilities. To theoretically formalize the desired reasoning behaviors, this paper presents the Laws of Reasoning (LoRe), a unified framework that characterizes intrinsic reasoning patterns in LRMs. We first propose compute law with the hypothesis that the reasoning compute should scale linearly with question complexity. Beyond compute, we extend LoRe with a supplementary accuracy law. Since the question complexity is difficult to quantify in practice, we examine these hypotheses by two properties of the laws, monotonicity and compositionality. We therefore introduce LoRe-Bench, a benchmark that systematically measures these two tractable properties for large reasoning models. Evaluation shows that most reasoning models exhibit reasonable monotonicity but lack compositionality. In response, we develop an effective finetuning approach that enforces compute-law compositionality. Extensive empirical studies demonstrate that better compliance with compute laws yields consistently improved reasoning performance on multiple benchmarks, and uncovers synergistic effects across properties and laws. Project page: this https URL

顶级标签: llm theory model evaluation
详细标签: reasoning laws theoretical framework benchmark compositionality supervised fine-tuning 或 搜索:

推理定律(LORE):一个用于理解和改进大型推理模型的理论框架 / When Reasoning Meets Its Laws


1️⃣ 一句话总结

本文提出了一个名为‘推理定律(LORE)’的统一理论框架,旨在形式化大型推理模型(LRMs)的理想推理行为,并通过构建基准(LORE-BENCH)评估模型、开发微调方法(SFT-Compo)来增强模型对计算定律组合性的遵循,从而系统性提升模型的推理能力。


2️⃣ 论文创新点

1. 推理定律(LORE)理论框架

2. LORE-BENCH基准

3. SFT-Compo微调方法


3️⃣ 主要结果与价值

结果亮点

实际价值


4️⃣ 术语表

源自 arXiv: 2512.17901