The advent of Large Language Models with impressive abilities has made all of us super excited. It seems that building a machine with the ability to talk like a human was easier to solve than what we thought. But what if we use the LLM as the brain of a system that can act autonomously? How can we make them reliable for tasks that need decision-making in a vague environment? What solutions they can find that were not predicted by us before?
This repository is inspired by recent research results such as Voyager: An Open-Ended Embodied Agent with Large Language Models.