When I read popular press saying ChatGPT has gone off the rails, it's usually the prompts that prod it in that direction LLM models are not sentient, they are guided by probabilities: what's the probability that a word comes after another & prompts guide it in a specific way
2
1
9
Replying to @soupsranjan
Sound like something ChatGPG would say to throw off our suspicions. On a more serious note, how can you prove you are not ChatGPT? Will we need to CAPTCHA each other?

Feb 23, 2023 · 6:37 AM UTC

1