Read what an AI model produced and judge it: which answer is better, where it is wrong, whether it is safe. Your judgement is the training signal, so the reasoning matters as much as the rating.
The work
- Rate and compare model outputs against the project's rubric.
- Write short, concrete justifications for your rating.
- Flag unsafe, false or policy-breaking output explicitly.
You need
- Strong reading comprehension in the language you evaluate.
- The ability to explain why, not only what.
- 18 years or older.
Every application is read by a person on our team, and selected candidates get an answer by email. Your application, and any CV you upload, are deleted no later than 2 years after you apply.
Apply
Takes a couple of minutes. No CV needed.
- Payment
- Via PayPal, Wise, or bank transfer
- Minimum payout
- None. You are paid what you have earned
- Fees to you
- Never, at any stage. If someone asks, it is not us
- Who hires you
- YPAI, a registered Norwegian company
Application received
Thanks. Your application for Model Evaluator is registered and will be read by a person on our team. We have emailed you a receipt with what happens next. Your application, and any CV you uploaded, are deleted no later than 2 years after you applied.
Your capability profile is registered too, so future matching projects can reach you. Check your email for a sign-in link to complete it, or to edit or delete it.