SOURCE-LINKED INTELLIGENCE

Tracing distinguishability through transformer processing with stochastic LayerNorm

arXiv · AI, language, vision and robotics · article · Aug 31, 2026 · UTC

Representational similarity is foundational to analyses of deep networks, yet distances between point-valued representations are not intrinsically tied to downstream function: nearby states may produce different behaviors, while distant states may behave similarly. We instead give representations volume, turning similarity into statistical distinguishability. Overlapping stochastic representations necessarily induce overlapping downstream distributions, grounding latent comparison in model function and bringing it under information-theoretic tools such as the data-processing inequality. We rea

Read original source ↗ Open in workspace

recordType

paper

region

Global