I apologize for the delayed response. We fit into the 2nd camp, so leveraging huggingface llms works nicely for us. From a security stand point, I was just trying to understand if any of our data would be sent to arize when using AutoEmbeddings.
When using AutoEmbeddings my understanding is that the models and tokenizers are still being downloaded from huggingface and all the computations to generate the embeddings are ran locally. Are there ever any situations where the data would be sent to Arize AI hosted llms when using generator.generate_embeddings(). Does Arize AI host their own llms?
