I am still reading the docs and going through tutorials. I would like to have some questions answered for clearness.
1 - If you fine-tune a model, do you have access to it via another API key so that you still have access to the base models with the other API-keys?
2 - Also, I read the following Help - what-is-chatgpt:
ChatGPT is fine-tuned from GPT-3.5, a language model trained to produce text. ChatGPT was optimized for dialogue by using Reinforcement Learning with Human Feedback (RLHF) – a method that uses human demonstrations and preference comparisons to guide the model toward desired behavior.
Do I need an integration of ChatGPT or one of a GPT-xy model? I would like to have it at expert-level regarding dialogues. I am a bit confused. Or will the GPT-xy models do the job, but htey are not fine-tuned for dialogue, as the site states.
So, would one also have to fine-tune via RLHF?
3 - Is there any information about a date frame when the access to the GPT-4 API will be granted? Docs - Fine-Tuning does not reveal a date.
4 - If I use a plugin that already exists and that has the ability to access and work via existing API, then I could just reference the fine-tuned model, I guess!?