Generative Agents is a computational platform for simulating autonomous agents that exhibit human-like social behaviors and decision-making processes. The system functions as a multi-agent simulator where individual participants operate within a virtual environment, driven by large language models to process observations and generate natural language actions.
The framework distinguishes itself through a hierarchical memory system that allows agents to store, retrieve, and synthesize past experiences into higher-level insights. This architecture supports the development of complex social dynamics by enabling agents to maintain personal histories and evolve their behavior based on long-term memory and periodic reflection.
Users can design custom virtual environments using spatial-graph mapping and define specific agent narratives to study social interactions in controlled settings. The platform includes tools for state persistence and simulation replay, allowing for the systematic analysis of behavioral trends and the reconstruction of past events through a browser-based interface.