Some of these movies work better than others, but all of them are trying to do something special.

These new models are specially trained to recognize when an LLM is potentially going off the rails. If they don’t like how an interaction is going, they have the power to stop it. Of course, every ...