4 ms·
Unfortunately, the method OpenAI may be using to reduce bias (by adding words to the prompt unknown to the user) is a naive approach that can affect images unex
by minimaxir 4y ago
Unfortunately, the method OpenAI may be using to reduce bias (by adding words to the prompt unknown to the user) is a naive approach that can affect images unexpectedly and outside of the domain OpenAI intended: https://twitter.com/rzhang88/status/1549472829304741888 https://twitter.com/rzhang88/status/1549472829304741888
I have also seeing some cases where the bias correction may not be working at all, so who knows. And it's why transparancy is important.
- kmeisthax 4y agoThis sounds like something that could backfire very badly on certain prompts. "person eating a watermelon" for example.
- CobrastanJorji 4y agoWhat a fascinating hack. I mean, yeah, naive and simplistic and doesn't really do anything interesting with the model itself, but props to the person who was given the "make this more diverse" instruction and said "okay, what's the simplest thing that could possibly work? What if I just append some races and genders onto the end of the query string, would that mostly work?" and then it did! Was it a GOOD idea? Maybe not. But I appreciate the optimization.