This is interesting. I'm curious how you calibrate it. I've found that sometimes AI is very confident about being able to do something (or not do something), and I've had to pick its response apart a bit to see what the real answer is. Are you getting accurate results with one pass?
1 comments