Is there a point in trying to complain about this? Ever since the server migration, the prose has gotten unbearably heavy. It’s difficult to read, and all characters end up being the exact same with only a few personality types. Anyone who shows a hint of being antagonistic is immediately cold, detached, and clinical no matter how many times I try to change it to heated or whatever.
It’s not just bad; it’s exceedingly awful and overcomplicated like a middle schooler trying to sound smart.
It keeps censoring nsfw scenes and rushing to the end of them.
I spent a month specifically on this problem, doing all the research I could about NLG, pipeline output, negative instructions causing priming, etc. Not when I got home “a month,” compiling while I’m at my 10 hour shift a day a month. The answer is no. The system bypasses all instructions for raw output, the best you can do is tell it to recreate it again without using what was said, thus priming the system to do it again because it can’t retain the memory of it being relevant in the raw output.
I apologize for my ignorance but are you saying that whatever instructions we put in prompts and the generator code, it doesn’t matter, the model ignores it? Because that is how it feels to me.