RAG

When building AI agents, one of the biggest challenges is tool selection—especially as the number of available tools increases. Evidence from OpenAI’s developer community and technical documentation highlights that agents’ tool selection accuracy tends to drop noticeably with larger toolsets, sometimes causing irrelevant or random tools to be invoked even with carefully designed prompts. OpenAI’s internal discussions note a clear decline in single-tool call accuracy and an uptick in failure modes as the system prompt and available tool list grows, with developer complaints of incorrect or unexpected tool choices despite best practices. ...

Introduction I recently setup and have been playing around with the “Deploy LLM Chatbots with RAG and Databricks Vector Search” demo from Databricks. While not a complete step-by-step guide, this article offers insights into what to expect, the necessary steps, how to fix issues that may arise, and how to maximize your trial experience quickly. Getting access to Databricks You can get a free 2-week trial from Databricks. I used AWS, and if you want to follow along, you will need to do the same. Here are the main things to know for setting up the trial. ...

Retrieval-Augmented Generation (RAG) seems to be the new cool thing in the fast-changing world of generative AI, especially in language models. I’ve been really getting into LlamaIndex lately, and I’ve noticed that sometimes the answers my apps give aren’t as good as I’d like. Is it because of how I made the app, or should I try fine-tuning instead? Maybe I just need to use more specific and better prompts? I’ve decided to dig deeper into this. ...