逻辑相变：理解大语言模型逻辑推理中的崩溃
Logical Phase Transitions: Understanding Collapse in LLM Logical Reasoning
AI总结：
研究通过控制逻辑复杂性分析大语言模型的逻辑推理性能，发现逻辑相变现象，提出神经符号课程调优框架以提升高复杂度下的推理能力。
摘要：
符号逻辑推理是大语言模型的关键但未被充分探索的能力，在数学推理和法律判断等高风险领域提供可靠且可验证的决策。本文系统分析了逻辑推理在受控增加的逻辑复杂性下的表现，揭示了一种未被识别的现象，称为逻辑相变：而非平滑退化，推理性能在某一范围内保持稳定，但在临界逻辑深度后突然崩溃，类似于水在临界温度阈值后冻结的物理相变。基于这一见解，我们提出了神经符号课程调优框架，该框架能够自适应地对齐自然语言与逻辑符号，建立共享表示，并重塑训练动态以围绕相变边界重新塑造训练动态，逐步增强推理能力。在五个基准测试中，我们的方法有效缓解了高复杂度下的逻辑推理崩溃，产生平均准确率提升+1.26在简单提示和+3.95在CoT，同时提升对未见逻辑组合的泛化能力。代码和数据可在https://github.com/AI4SS/Logical-Phase-Transitions获取。
英文摘要：
Symbolic logical reasoning is a critical yet underexplored capability of large language models (LLMs), providing reliable and verifiable decision-making in high-stakes domains such as mathematical reasoning and legal judgment. In this study, we present a systematic analysis of logical reasoning under controlled increases in logical complexity, and reveal a previously unrecognized phenomenon, which we term Logical Phase Transitions: rather than degrading smoothly, logical reasoning performance remains stable within a regime but collapses abruptly beyond a critical logical depth, mirroring physical phase transitions such as water freezing beyond a critical temperature threshold. Building on this insight, we propose Neuro-Symbolic Curriculum Tuning, a principled framework that adaptively aligns natural language with logical symbols to establish a shared representation, and reshapes training dynamics around phase-transition boundaries to progressively strengthen reasoning at increasing logical depths. Experiments on five benchmarks show that our approach effectively mitigates logical reasoning collapse at high complexity, yielding average accuracy gains of +1.26 in naive prompting and +3.95 in CoT, while improving generalization to unseen logical compositions. Code and data are available at https://github.com/AI4SS/Logical-Phase-Transitions.
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。
