Vanja Karan (Digital Text Sciences, University of Vienna)
Lecture Series: Introduction to Digital Humanities
Large Language Models (LLMs) have become an ubiquitous tool in modern natural language processing. The talk with first cover the basic types and terminology regarding LLMs. Moreover, an overview will be given of methods for evaluating LLMs to discover their strengths, weaknesses, and current limits. Other aspects of LLMs that are particularly relevant for use in digital humanities research will be discussed, such as explainability, privacy, security, and factuality. Particular emphasis will be on the explainability aspect. Finally, some example applications of LLMs for digital humanities research will be presented.