Simulating Agents
Raphaël discusses the concept of models simulating specific agents based on prompts, delving into the idea of in-context learning and linguistic emulation. He questions whether models truly simulate mental states like beliefs and desires or simply mimic patterns from training data, using examples like Jane Austen's style and Bing chat's behavior.In this clip
From this podcast

Machine Learning Street Talk (MLST)
#107 - Dr. RAPHAËL MILLIÈRE - Linguistics, Theory of Mind, Grounding
Related Questions