How do I get started doing evals?

Hi, I’ve been incorporating OpenAI apis for davinci and dall-e for web3 applications and I’m curious about how to do evals to gain access to the gpt4 API? I’ve joined the wait-list, but I’m hoping to do evals to hopefully gain priority. I’m specifically interested in utilizing gpt4 to create algorithms based on critical growth factors of produce for concepts I’m developing called ‘decentralized algorithmic agriculture’ for means of encouraging agricultural reformation. I’ve gone over the evals GitHub and is it basically creating database of prompts to evaluate the output of gpt? Are the evals for all gpt models? How are the gpt models evaluated if there’s no access to the API?

Thank you


Hey champ!

I’ll suggest you have a look at the GitHub, it’s full of well written examples, templates and guides on how to get started:

you use the API to run evals, please be aware of the costs associated with that.

Thank you, I was going over this earlier in the morning. In some of the evals that were being submitted some of the users were inquiring about whether the evals were being accepted anymore. Are evals no longer being accepted?

1 Like

I believe they’re still taking evals and will continue to do so in the future, but there’s a backlog of submissions, I’d expect it to take a week or two from the moment of submission.

1 Like

Try this - Decoding OpenAI Evals

1 Like