DataHour: Evaluating LLM Applications

LLM Applications are becoming popular for solving multiple tasks. One of the popular applications of LLMs today is in Retrieval Augmented Generation (RAG) applications which are useful for multitude of tasks such as augmented search on proprietary data. Frameworks like LangChain alongside vector databases makes this possible, but how to best evaluate these applications?

In this session, we will understand the high-level understanding of how LLM applications are made, and various kinds of evaluation strategies that can be taken. We will go into the hands-on code to go about building intuition on evaluating the LLM applications well.