linear combination

3 papers

• Advancing Interpretability of CLIP Representations with Concept Surrogate Model
• Capturing Individual Human Preferences with Reward Features
• In-Context Compositional Learning vis Sparse Coding Transformer