Localizing and Editing Knowledge in LLMs with Peter Hase - #679
The TWIML AI Podcast (formerly This Week in Machine Learning & Artificial Intelligence)
00:00
Intro
This chapter delves into the research interests of a PhD student focused on language models, emphasizing interpretability and the necessity of understanding their reasoning processes. It also discusses the challenges of maintaining effective oversight in increasingly complex AI systems and the implications of updating factual knowledge within these models.
Transcript
Play full episode
