Is the role of functions in the AI Render (submitUserMessage) a new way to provide context to the LLM for a RAG? Usually we just put the RAG content as part of the prompt. The data from the functions gets attached to the prompt? What's the benefit of using functions for RAG to help provide context to the LLM?
Is the role of functions in the AI Render (submitUserMessage) a new way to provide context to the LLM for a RAG? Usually we just put the RAG content as part of the prompt. The data from the functions gets attached to the prompt? What's the benefit of using functions for RAG to help provide context to the LLM?