LLMLow
Embedding Models Measure in Peculiar Ways
- Occurred
- Sep 17, 2026
- Published
- Sep 17, 2026
- Source
- arXiv
Summary
Embedding spaces define notions of semantic similarity and distance. We study whether those embeddings reflect physical measurements of mass, distance, time and volume, which admit a unique, objective notion of semantic equivalence and distance. We find that physical measurement is only weakly modeled in the embedding space, and that instead quite peculiar measurement patterns can be observed. Further analysis indicates that embedding representations of physical measurements are strongly influenced by superficial string similarity, and recalibration of similarity does not substantially improve the alignment.
Classification
Category confidence: 90%
Severity confidence: 95%
Rationale: Research paper analyzing embedding model behavior without concrete incident or harm.
Classifier: Qwen/Qwen3.5-35B-A3B-FP8