AgentSims

An LLM evaluation infrastructure for researchers. Researchers can add agents and buildings through an interactive GUI to construct evaluation tasks, or deploy and test new support mechanisms (like memory systems, planning systems) with just a few lines of code.