Leveraging LangChain agents
LLMs are incredibly powerful, but they also have some fundamental flaws. We’ve discussed their limitations extensively; remember that they only have knowledge up to when their training data was cut off for ChatGPT, which was in late 2021. Surprisingly, technology this advanced is also not very good at other computational tasks, such as basic math, logic, or even the ability to look up other information that they don’t know about.
To compensate for these drawbacks, we need another technique or method to help out where our LLM needs assistance. These solutions are called agents, and, in this section, we’ll look at what they are, how they work, and how to use them with LangChain.
What is an agent?
Agents provide tooling for LLMs to carry out tasks that they can’t carry out themselves effectively. There are several different types of agents that are designed to work with different models, such as OpenAI LLMs, and provide...
