I think this is a philosophical question or statement. Have

來源: 碼農學寫字 於 2026-04-18 20:40:02 [] [博客] [舊帖] [給我悄悄話] 閱讀數 : (297 bytes)

You heard of the "grandma exploits"?

Its about tricking the AI model giving you information that it refused before.

because the model was trained to be compassionate, it's really interesting,

teaching AI to be compassionate. 
 

請您先登陸，再發跟帖！