Concept Layers: Enhancing Interpretability and Intervenability via LLM Conceptualization

Re-assign community