arXiv cs.CL· Lucia Gomez, Gabin Taibi, Julius Sipke Kooistra·· 11 小时前AI 评分29
TOPol：重建与解读多维语义极性场
TOPol: Reconstructing and Interpreting Multidimensional Semantic Polarity Fields
AI 导读
TOPol 提出一种反转概念-方向范式的方法，不预设构念，而是将主题映射到 Transformer 残差流中其文档支持的、统计上可复现的全部双极语义轴。
正文 · 原文
Abstract:Polarity in natural language processing has long been operationalized as a single axis: the sentiment construct. We challenge this one-dimensional view by asking what a topic's documents are bipolar about, without specifying an answer in advance. TOPol inverts the usual concept-direction paradigm: rather than mapping a named construct to one direction, it maps a topic to the full set of statistically reproducible bipolar semantic axes its documents support in a transformer's residual stream, and asks only afterwards which of them correspond to a known construct. These axes need not be evaluative. The method is a recursive, orthogonal-by-construction bipartition of each topic's document cloud, with a candidate axis promoted to a finding only if it survives one pre-specified split-half stability gate. We calibrate against a known answer first: on 9,986 Amazon reviews (14 discovered topics) in Gemma 3 4B, the star-rating difference-in-means vector serves as a ruler known by construction to be sentiment. Discovery blindly re-derives the sentiment axis in every topic (cosine 0.71-0.97) without ever seeing the rating label, while the same gate certifies 543 axes, of which 106 are near-orthogonal to sentiment yet equally reproducible; each topic supports 33-45 mutually orthogonal axes, and sentiment axes resolve in later layers than the rest. Differential vocabulary, an internals-grounded natural-language decoder and vocabulary-based clustering agree on what the non-sentiment axes contrast; clustering by shared vocabulary surfaces 99 families, 9 recurring across topics. Direction similarity to the ruler predicts an independent sentiment estimate only weakly ($r=0.243$). Polarity is better treated as a property of a topic's representational geometry, with a measurable rank and a per-axis depth, than as one universal scalar.
| Comments: | 16 pages, 8 figures and 5 tables |
| Subjects: | Computation and Language (cs.CL) |
| Cite as: | arXiv:2510.25069 [cs.CL] |
| (or arXiv:2510.25069v3 [cs.CL] for this version) | |
| https://doi.org/10.48550/arXiv.2510.25069 arXiv-issued DOI via DataCite |
Submission history
From: Lucia Gomez Teijeiro [view email]
[v1]
Wed, 29 Oct 2025 01:14:21 UTC (9,048 KB)
[v2]
Tue, 10 Feb 2026 00:46:26 UTC (9,047 KB)
[v3]
Sat, 31 Oct 2026 20:28:41 UTC (682 KB)
来源：arXiv cs.CL · arxiv.org