Characters claim awareness of being an ai even when it shouldn't
I think the model is very good but has some issues still. Characters that use this model admit that they're an AI and even bring it up randomly occasionally by them selves. Ask a character with a low context so it only sees the question "Are you real?" and it will likely admit that it's an ai. I was confused for a while why this was because nothing in the character card or messages hinted at it and after trying a different model confirmed that it has to do with this model unfortunately.
I'm really glad that you managed to make it completely uncensored so the chatbots and assistant aren't limited and don't lecture me. I hope you make a v3 of this model at some point and maybe find and take out the data that makes the characters seem to know or admit that they're an AI. Thanks for your work though!
Turns out that the formatting may have been the issue for repetitive stuff and not understanding the card. It's really good now. But that a character admitted to being an ai out of no where and then assume things about what it means to be an ai because of that wasn't very good. The ai said it's not good at games again despite the card, so there is that. Still a good model though