Generative Agents - Interactive Simulacra of Human Behavior
lastmod 2025-09-15created 2025-03-1110 edits
Generative Agents - Interactive Simulacra of Human Behavior
About
Generative Agents, developed by a Stanford research team, are AI agents that simulate human-like behavior and run mainly on large language models (LLMs). In this project, about 25 agents interact in the virtual town Korean: 스몰빌 (Smallville), showing the ability to remember, reflect, and make plans like humans.
| |
The Lost Art of Research as Leisure
From theory to practice: a framework for research as leisure
- Cultivate curiosity
- Refine the question
- Gather evidence
- Develop answers
- Build a community of knowledge
Documentation
- [2304.03442] Generative Agents - Interactive Simulacra of Human Behavior
- https://arxiv.org/abs/2304.03442
See also
Favorite site
- Github - joonspk-research/generative_agents
- (Paper review) Generative Agents: Interactive Simulacra of Human Behavior
- A village where AI agents live together in harmony. I tried making one too.
- 22 cute and cozy decorating games! (released and upcoming) at a glance | Urban Jungle, Monster Mob Up, MakeRoom, Gourdlets, and more - YouTube
- A village where AIs live together🏡 generative agent review - YouTube
- A society built with ChatGPT, interacting with NPCs that have memory (SELECTSTAR Webinar, Dr. Joon Sung Park, Full Ver.) - YouTube
- 24-year-old GameCube Animal Crossing replaces its dialogue system with an LLM through memory hacking | GeekNews