Most Users agree that the model’s over-optimization has made it overly cautious and frustrating to use for creative or nuanced tasks.
Thanks for sharing this feedback. @Troy_Bradley
Could you share which model you’re referring to, along with a few examples where it felt overly cautious or less helpful than expected? Specific prompts or situations would help us better understand the behavior you’re seeing and pass along more useful feedback to the team.
~ Smith
The latest one for the past few days it just lost its energy in the way it talks and it always replies with “I can’t generate scenes with graphic content” even when I specified that said bloody as in the British slang so can you just dial back the censorship a couple notches and try give the personality and energy a boost please
Thanks for clarifying.
I understand what you mean, especially with words like “bloody” being interpreted as graphic content when you meant it as British slang. That kind of context mismatch can definitely make the response feel overly restricted or less natural than expected.
I’ll pass this feedback along to the team. If you have any specific prompts where this happened, feel free to share them here too, since examples are especially helpful for reviewing this kind of behavior.
~ Smith