Geometric Limits of Knowledge Distillation: A Minimum-Width Theorem via Superposition Theory
Paper • 2604.04037 • Published • 2
Nilesh Sarkar
sarkarnilesh
·
AI & ML interests
I like to train deep neural networks and large language models! :D
Recent Activity
authored a paper 8 days ago
Geometric Limits of Knowledge Distillation: A Minimum-Width Theorem via Superposition Theory updated a model 15 days ago
sarkarnilesh/katana-cad-experimentOrganizations
authored a
paper 8 days ago