Concept Bottleneck Large Language Models

Interpretable Deep Learning Seminar Series Speaker: Chung-En Sun (UC San Diego) Title: Concept Bottleneck Large Language Models Abstract We introduce Concept Bottleneck Large Language Models (CB-LLMs), a novel framework for building inherently interpretable Large Language Models (LLMs). In contrast to traditional black-box LLMs that rely on limited post-hoc interpretations, CB-LLMs integrate intrinsic interpretability directly into the LLMs -- allowing accurate explanations with scalability and transparency. Interpretable Deep Learning website: https://interpretabledeeplearning.github.io/