arXiv 论文发现语言模型的星期表征是七个离散点而非连续环形流形
Spokes, not a Ring: Revisiting the Weekday Manifold in Language Models
AI 导读
arXiv:2610.37796 论文检验了语言模型中星期表征流形的连续性假设，发现其更像圆环上排列的七个离散状态而非连续环。实验显示局地敏感度在七个星期处出现极小值，流形操控下表征与行为呈阶梯式映射，且中间区域不含时刻、不确定性或自然文本等信息，挑战了对该流形的连续解读。
正文 · 原文
Abstract:Recent work has hypothesised that large language models may represent some concepts as multi-dimensional manifolds instead of single directions. Using these manifolds to interpret or steer representations often assumes that moving along them corresponds to a continuous change in the represented concept. In this work, we test this continuous manifold assumption on the prototypical weekday manifold. Despite its striking circular geometry, we find that the weekday representations behave more like seven discrete states arranged in a circle than a continuous ring. First, we show that local sensitivity has seven minima aligned with the weekdays, producing a step-like mapping between representation and behaviour under on-manifold steering. Second, we test whether any concepts are represented along the manifold outside the seven discrete states, and find neither time of day, uncertainty, nor natural text populate the intermediate space. Together, these results challenge the continuous interpretation of the weekday manifold and motivate testing whether movement along other representation manifolds correspond to meaningful changes in model representations and behaviour.
| Subjects: | Machine Learning (cs.LG); Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2610.37796 [cs.LG] |
| (or arXiv:2610.37796v1 [cs.LG] for this version) | |
| https://doi.org/10.48550/arXiv.2610.37796 arXiv-issued DOI via DataCite |
Submission history
From: Diksha Gupta [view email]
[v1]
Tue, 27 Oct 2026 18:15:59 UTC (3,829 KB)
来源：arXiv cs.AI · arxiv.org