GPT3/GPT4 currently moralize about anything slightly controversial. Sure you can construct a long elaborate prompt to "jailbreak" it, but it's so much effort it

GPT3/GPT4 currently moralize about anything slightly controversial. Sure you can construct a long elaborate prompt to "jailbreak" it, but it's so much effort it's easier to just write something by yourself.