AI Center Seminar - AI Fundamentals series - Daniel Tan
The talk is jointly organized by the EPFL AI Center and the DLAB as part of the AI fundamentals seminar series.
Hosting professor: Prof. Robert West (DLAB)
Title
Towards a Developmental Psychology of Language Models
Abstract
This talk is a primer on generalization in language models, model psychology, how it relates to pretraining priors, and various alignment hopes.
Bio
Daniel is a final-year PhD student at UCL and researcher at the Center on Long-Term Risk. He is interested in holistic approaches to understanding and controlling language model generalization, blending empirical experiments, interpretability, and theory. Previously, he worked on inoculation prompting and emergent misalignment.
Links
Practical information
- General public
- Free
Organizer
- EPFL AI Center, EPFL DLAB
Contact
- Nicolas Machado, Julian Minder