Compositional Generalization Requires Linear, Orthogonal Representations in Vision Embedding Models
Paper
• 2602.24264 • Published
• 14
Fabian Morelli
fabianmorelli
AI & ML interests
None yet
Recent Activity
upvoted a paper 5 months ago
DISCO: Diversifying Sample Condensation for Efficient Model Evaluation Organizations
None yet
upvoted a paper 1 day ago