LLM Toolbox
So far, we’ve explored some of the basics of LLMs – transformers, prompt engineering, and some of the popular open source models. In this chapter, we’ll dive into some of the tools that allow you to build full-fledged systems with these models – this will allow us to move beyond simple chat interactions with models to interconnected systems that can retrieve information from external sources, execute various applications, remember the history of your personal interactions with the model, and customize results based on user-specific sets of documents that provide context to requests. To do so, we’ll need to store documents in vector databases, retrieve relevant documents from those stores to enhance the context of our prompts, link models that have been specialized for specific tasks as “agents,” and log the results of our experiments. In the process of building these “agentic” systems, we’ll also touch...
