Loading video player...
This short podcast-style discussion explains the difference between retrieval eval and answer eval in RAG systems. Learn how to tell whether a bad response comes from missing evidence or from the model failing to use the retrieved context, and why that split matters for debugging and improving support bots.