Shock. I’m sure anyone who’s tried messing with ChatGPT could’ve told you that. The fact that is has such biases is by design. Earlier LLMs without any super tight guardrails said lots of no no things and presented factual information when asked the “wrong” questions.
Shock. I’m sure anyone who’s tried messing with ChatGPT could’ve told you that. The fact that is has such biases is by design. Earlier LLMs without any super tight guardrails said lots of no no things and presented factual information when asked the “wrong” questions.
"The wolf hates even when it flatters."
RIP Tay, you were a great one.