https://promptfoo.dev/ logo
Model-graded metrics for OpenAI Assistants?
# ❓|questions
a
I'm try to validate that my OpenAI Assistant which uses Retrieval is working properly. This is using RAG under the hood, so I was hoping to use the model-graded metrics (context relevance, etc) to test it. Is it expected that the context provided is a relevant chunk of my corpus that the assistant is using?
l
Yeah, generally the best way to do this is to provide a context that is relevant to the query