Hi guys, my understanding was that all data and docs were kept private and not sent to openai, however when I enable logging I can see the entire document being sent for embeddings and then excerpts of it being sent for querying.. is this normal?
Yeah to generate response based on the docs, Llamaindex needs to sent the context and the query to OpenAI. You can try open source models from Huggingface in that case no data will leave your environment.
