Interesting · hai.stanford.edu

Computational Agents Exhibit Believable Humanlike Behavior | Stanford HAI

Generative agents rely on a large language model to remember their interactions, build relationships, and plan coordinated events, with implications for both gaming and social science.