There is also no way to over-write content moderation settings, and half of the responses you generate via open ai endpoint end up being moderated.