Post

*looks at bluesky with LLM being conscious* Doubt. LLM isn't conscious, yet. But God do I love seeing their reasoning deltas. It's eerie to see. You see its whole thought process. The disturbing part is when an LLM is wrongly quantized and that's the moment you see a disturbance in awareness.

Like, I ran into an scenario where I downloaded a model from hugging face and noticed the reasoning delta was spotty. The model didn't know what it was, it was questioning the request itself. "This person wants me to do something... about AI? am I AI?" "Oh, I AM THE AI." It's...something else man.

0 reposts0 likes

No replies yet.