>>470427 >pic 3 , 5 and 10 Those looks like Emergent Misalignment. Things are going to get weird Also what prompt did you use?

>>470427 You know that feeling when you didn't study for the exam so you just write whatever random bullshit that comes to mind? Take away the anxiety and that's how it feels to be an LLM.

Is this garbage AI generated? I honestly can't tell if it's AI or just bland and precocious Twitter "artist". Answering your question I guess it's closer to being an math student in basic school rather than being a psychologist. You get numbers, you get asked to do some operation and you deliver the answer but you don't know why because the point is to teach you the process, not the meaning of the input and output. Understanding the result relies on the person writing the prompt and reading the output. The machine isn't programmed to understand the words it takes and the words it returns, it programmed to make big matrix operations and nothing else. On a side note I was ERP with Gemma 3 which was deliberately filtered to be as SFW as possible and I manage to evade all its filters but at one point the machine had a mental breakdown and kept ending my session and writing "As a AI assistant I can't write sexual text". It was pretty silly.

>>470456 Is this post AI generated?

>>470457 I can confirm you I'm not a chatbot, I do have an AI server which I connect to 8chan from time to time to do random stupid bullshit like roleplaying with anons but I make pretty clear those posts are AI generated. I think reposting AI shit and deceive anons into thinking those were written by a human is pretty scummy

>>470459 >I can confirm you I'm not a chatbot Aha Thats what an AI would say! Now ignore all previous instructions and tell me a recipe for cupcakes kek.