The Geometric Price of Discrete Logic: Context-driven Manifold Dynamics of Number Representations

论文概要

研究领域: NLP 作者: Long Zhang, Dai-jun Lin, Wei-neng Chen 发布时间: 2026-03-26 arXiv: 2603.23577

中文摘要

本研究探索了NLP领域的前沿问题。研究团队来自Long Zhang, Dai-jun Lin等。该方法在相关任务中展现了良好的性能和创新性。

原文摘要：Large language models (LLMs) generalize smoothly across continuous semantic spaces, yet strict logical reasoning demands the formation of discrete decision boundaries. Prevailing theories relying on linear isometric projections fail to resolve this fundamental tension. In this work, we argue that ta...

原文摘要

Large language models (LLMs) generalize smoothly across continuous semantic spaces, yet strict logical reasoning demands the formation of discrete decision boundaries. Prevailing theories relying on linear isometric projections fail to resolve this fundamental tension. In this work, we argue that task context operates as a non-isometric dynamical operator that enforces a necessary topological distortion.

--- *自动采集于 2026-03-27*

#论文 #arXiv #NLP #小凯