pretraining-dynamics

Tag

Cards List
#pretraining-dynamics

Neural Collapse Is Forbidden: Information Floors in Language Models

arXiv cs.CL · 2026-07-13 Cached

This paper argues that within-class variance in language model representations is not incomplete neural collapse but allocated information storage, and that the allocation obeys an information floor law. Across 14 models, macro-category structure carries only 4–12% of representational variance, while within-token context dominates at 79–91%.

0 favorites 0 likes
← Back to home

Submit Feedback