AI.info

Yousung Lee

Explore Yousung Lee on AI.info.

• Why Linear Interpretability Works: Invariant Subspaces as a Result of Architectural Constraints