Key Concepts

Large Language Models

Large Language Models (LLMs) are advanced AI systems trained on vast amounts of text data to understand, generate, and interpret human language. They can perform a wide range of natural language processing tasks, such as answering questions, writing coherent and contextually relevant text, translating languages, and more. LLMs like GPT (Generative Pre-trained Transformer) series have demonstrated remarkable abilities in generating human-like text, making them valuable for applications in content creation, customer service, education, and research.

Agent

Agent technology refers to the development of AI systems or software agents that can perform tasks autonomously or semi-autonomously on behalf of users or other systems[9]. These agents can make decisions, take actions, and interact with their environment or other agents based on predefined criteria or learning. Agents can range from simple automated scripts to complex AI-driven entities capable of adaptive and intelligent behavior.

LLM-driven Agents

When LLMs and Agent technology are combined, they create powerful AI agents capable of understanding and generating natural language while autonomously performing complex tasks[10, 11]. Back in 2023, Stanford and Google researchers created a virtual town, where 25 characters live their lives independently with a high degree of realistic behavior[1]. Inspired by that, we use LLM Agents to build a wonderful game: lots of characters live in a town, each of them is an LLM Agent. They have their own characteristics, thoughts, memories and even abilities to create art. They live, socialize and even gamble in this town. This blog is to explain the technology we use.