Unverified85% confidenceFactTime unknown
斯坦福的'Generative Agents'用虚拟小镇评测社交能力，Google DeepMind用各类游戏评测通用智能
1
Sources
85%
Confidence
Long-term
Relevance
6/2/2026
First Seen
Sources
用AI开发三消游戏并让Agent自己玩：全流程实战
bilibili琪琪不慌_
Related Entities
Related Claims
Verified2023年斯坦福大学与Google Research联合发表了「Generative Agents: Interactive Simulacra of Human Behavior」论文，在虚拟小镇中部署了25个基于LLM的智能体79% similarUnverified学术界正探索基础模型在机器人领域的应用，如Google DeepMind的RT-2和斯坦福的Open X-Embodiment项目69% similarUnverifiedGoogle DeepMind和斯坦福大学的研究团队曾系统性记录了超过100种涌现能力，涵盖数学推理、代码生成、多语言理解等领域64% similarUnverified涌现能力（Emergent Ability）现象由谷歌DeepMind研究团队于2022年系统记录64% similarUnverifiedGoogle DeepMind推动了A2A（Agent-to-Agent）通信规范64% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/36305API
curl https://kongchang.com/api/v1/knowledge/claims/36305MCP
get_claim(id=36305)