Hey folks, how are y’all passing in context retrieved from documents while creating a openai.createChatCompletion object? Do you just append the context text as the first user message or is there a better way to do it?

it shouldn’t really matter as long as you label the prompt correctly. E.g. something like:

You are a helpful agent and will receive a user query along with context information to use when answering the user query

User query: ${userQuery}

Context: ${context}

Answer: