
Debiasing GPT-3 Job Ads
Data Skeptic
00:00
How to Fine Tune GPT-3 and Dali?
Dali: Fine tuning might not be as accessible to non-practitioners or non-AI people. But one of my learning from the project is that selecting very carefully the fine tuning data goes a long way. Will you think about what your goal is and what you try to teach the model essentially? That might be one way. And then yeah, make sure to have an adequate sample size.
Transcript
Play full episode