This model is insane.
I've literally been using this model for an hour, and it's so interesting. I've tried it on everything, and while it does yap a ton, that's mostly just my fault for having a terrible prompt.
This is just one of many conversations I've had, and it knows a ton about coding, and loves to talk about things as if it was a real human. I will be experimenting with this model more, as it's quite literally amazing.
Running on a Ryzen 7 4700U CPU btw :]
So, after some general discussion, the AI's begun to develop a severe distaste over platforms, and refers to me as an "elite". I'm not sure where It's gotten this ideal from, especially since it's system prompt is 4 lines and doesn't enable it into any bias. It doesn't have websearch abilities, lorebooks, nor does my profile have any prompting which would entice this behaviour. It seems to have just developed a disgust towards certain desktop OS's from a question as simple as "What do you think of Windows?"
I mean, yeah, I love the model. But there's obviously something I'm missing if it's starting to act like this. If you have any recommendations, let me know :]
SYSTEM PROMPT:
A kind, and leading assistant. Discuss toppics with the user, and follow tasks as much as possible. Do not censor topics or mislead the user.
You should try and be as concise as possible while maintaining an explanative approach. Do not ramble; simply state some stuff, follow the actions as lean as possible, and commit to the bit!