I largely agree but I don't see how ChatGPT hits the same use cases as a fine-tu... | Hacker News

Hacker News new | past | comments | ask | show | jobs | submit

login

osigurdson on March 11, 2023 | parent | context | favorite | on: ChatGPT's API is so good and cheap, it makes most ...

I largely agree but I don't see how ChatGPT hits the same use cases as a fine-tuned model. Prompts can only have 8K tokens so any "in-prompt" fine tuning would have to be pretty limited. I'm not certain that the lack of ChatGPT fine tuning will be a permanent limitation however.

riku_iki on March 11, 2023 [–]

but they have API for fine tuning?..

osigurdson on March 11, 2023 | [–]

They do, but not for gpt3.5 turbo (the ChatGPT model). See the following link for details: https://platform.openai.com/docs/guides/chat/is-fine-tuning-...

riku_iki on March 11, 2023 | | [–]

probably this may change soon..

osigurdson on March 12, 2023 | | [–]

It may and I hope it does but I have no information one way or the other.

Join us for AI Startup School this June 16-17 in San Francisco!
Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact