How deterministic are these processes? Is there a way to invoke stochasticity in the embedding and generative process with the OpenAI endpoints?
No way to enable stochasticity for embedding models, but the generation models from OpenAI can all add stochasticity by specifying the temperature