Comment on LLM Evals Are Just Tests. Why Are We Making This So Complicated?Comments−8organicbits1ySo, did the tests allow you to build a system that never confused existing features with new features? That seems like the problem statement, but I think I'm only seeing probabilistic testing.−camwestOP1yNever? No. Way less likely? Yes!In dev we do 100 consistency checks and get green. In CI we do 10.
Comments
So, did the tests allow you to build a system that never confused existing features with new features? That seems like the problem statement, but I think I'm only seeing probabilistic testing.
Never? No. Way less likely? Yes!
In dev we do 100 consistency checks and get green. In CI we do 10.