I have two questions chat GPT has been unable to give me a full and complete answer about.
-
Using Moderation - I’m building a AI tool for children. I’ve been able to add some content controls within the actual text prompt but I’d like to use the moderation as well for extra protection. Is there any way to do this without having to make one API call to the moderation end point and a second to the chat model? Performance issues are frustrating, making it slow down the response time. I want the moderation to pass before passing the text prompt from the user into the chat endpoint. However, I’d rather just have all that done in one API call to the 3.5 version’s end point. Is there a way to do that?
-
Adding in data from… well any kind of file. I’m familiar with JSON, csv, doc, text and the all the usual file types but how to use any of these files to basically include it in the prompt? I’ve read about embeddings but that’s a bit complicated for me with all the categorizing of the data that would need to take place. Chatgpt responded that it can use a link to a google drive file but after an hour of trial and error it never seemed to be able to consistently access the content that way. It kept responding that ‘I’m an AI … blah blah… can’t read a file’ but when I asked it how to share a file it was like… steps 1-5 and I did all of those, shared the link and it still didn’t work. any other solutions outside of embedding or a simple way to use embedding without being so complicated?