@PopulistRight well, the interesting thing about LLMs is that they in general have enough input ( hence the large ) and the black box model, they've decided to do reinforcement training rather than removing "problematic" inputs.