TY  - RPRT
TI  - Saddle-To-Saddle Dynamics in Deep ReLU Networks: Low-Rank Bias in the First Saddle Escape
AU  - Ioannis Bantzis
AU  - James B. Simon
AU  - Arthur Jacot
PY  - 2026
UR  - https://arxiv.org/abs/2505.21722
ID  - 2505.21722
ER  -