"text-embedding-ada-002" for Retrieval Plugin

It appears that the “text-embedding-ada-002” model is used when utilizing the Retrieval Plugin. This model is primarily expected to be used for personal or internal document retrieval. However, when I ask a question to ChatGPT and the intent of the question needs to be understood to determine which plugin to invoke, it is possible that this model is also used for that purpose.

Not unless you explicitly do it yourself.

For instance, you could use the retrieval to put text into the context window which could then inform the model as to which function to call.