Computational Linguistics II - Interpreting Language Models

Catalog description

Language models form the basis of the modern AI boom. In this class, we will explore the basics (and not-so-basics) of how language models work. Despite being so widely used, we don't really understand what drives their decisions. Therefore, this course also serves as an introduction to developing and improving these models. We will read contemporary research papers to better understand how research works in this space and to better understand how these models represent knowledge and the factors that influence their decisions. Ultimately, students conduct novel research projects to uncover new and better scientific understanding of how these models work.