# The Rise and Potential of Large Language Model Based Agents: A Survey

**Authors:** Zhiheng Xi, Wenxiang Chen, Xin Guo, Wei He, Yiwen Ding, Boyang Hong, Ming Zhang, Junzhe Wang, Senjie Jin, Enyu Zhou, Rui Zheng, Xiaoran Fan, Xiao Wang, Limao Xiong, Yuhao Zhou, Weiran Wang, Changhao Jiang, Yicheng Zou, Xiangyang Liu, Zhangyue Yin  
**arXiv:** [2309.07864](https://arxiv.org/abs/2309.07864)  
**Date:** 2023-09-14  
**Summary page:** https://www.summarizepaper.com/en/paper/2309.07864v3/

## Key points

- AI agents have emerged as a promising avenue in the pursuit of artificial intelligence (AI) that surpasses human capabilities
- Large language models (LLMs) have gained attention as potential foundations for Artificial General Intelligence (AGI)
- LLMs demonstrate versatile capabilities and offer hope for building general AI agents
- Researchers have leveraged LLMs to build AI agents and achieved significant progress
- The paper explores the rise and potential of LLM-based agents, tracing the concept of agents from its philosophical origins to its development in AI
- LLMs are suitable foundations for building AI agents and a general framework comprising brain, perception, and action is presented
- Extensive applications of LLM-based agents in various scenarios such as single-agent, multi-agent, and human-agent cooperation are discussed
- Agent societies and the behavior, personality, and social phenomena that emerge from interactions between LLM-based agents are explored
- Key topics and open problems within the field of LLM-based agents are discussed
- The paper provides a repository of related papers for further exploration

## Summary

In the pursuit of artificial intelligence (AI) that surpasses human capabilities, AI agents have emerged as a promising avenue. These agents are artificial entities that sense their environment, make decisions, and take actions. While significant efforts have been made to develop intelligent agents, they often focus on specific capabilities or tasks rather than providing a general model for diverse scenarios. Large language models (LLMs) have recently gained attention as potential foundations for Artificial General Intelligence (AGI). LLMs demonstrate versatile capabilities and offer hope for building general AI agents. Researchers have leveraged LLMs to build AI agents and have achieved significant progress in this area. In this comprehensive survey paper, the authors explore the rise and potential of LLM-based agents. They begin by tracing the concept of agents from its philosophical origins to its development in AI. They explain why LLMs are suitable foundations for building AI agents and present a general framework comprising three main components: brain, perception, and action. This framework can be tailored for different applications. The paper then delves into the extensive applications of LLM-based agents in various scenarios such as single-agent scenarios, multi-agent scenarios, and human-agent cooperation. Additionally, the authors explore agent societies and examine the behavior, personality, and social phenomena that emerge from interactions between LLM-based agents. These insights offer valuable perspectives for understanding human society. Finally, the authors discuss several key topics and open problems within the field of LLM-based agents. They provide a repository of related papers for further exploration. Overall, this survey paper provides a comprehensive overview of LLM-based agents and their potential in advancing AI research. It highlights the importance of developing general models that can adapt to diverse scenarios and offers insights into building intelligent agent societies.