Discussion about this post

User's avatar
Petar Dimov's avatar

When models are trained to please humans, it really shows why you still have to double-check outputs instead of trusting tone or confidence

No posts

Ready for more?