Embedding Models Measure in Peculiar Ways

The paper investigates whether embedding spaces capture objective physical measurements such as mass, distance, time, and volume. It finds that these embeddings only weakly model such measurements and instead exhibit peculiar patterns. Further analysis shows that superficial string similarity heavily influences the representation of physical measurements, and recalibrating similarity does not significantly improve alignment.

Machine-generated by The Flow from the publisher's headline and feed description — not written or checked by a human. The full article lives at arXiv Machine Learning.

• rag
• safety