human: je pense
llm: je ponce
Comment on Chatbots Make Terrible Doctors, New Study Finds
plyth@feddit.org 16 hours agoThey simply string words together according to statistical likelihood, without having any notion of what the words mean
What gives you the confidence that you don’t do the same?
Digit@lemmy.wtf 13 hours ago
pkjqpg1h@lemmy.zip 6 hours ago
because I don’t just look the words I know and feel their meaning, and I’m aware of process (this is why ALL LLMs is vulnurable to Promt Injections) also talking about Promt Injection think about when ChatGPT confidently give you an advice that would kill you because of prompt injection