Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

Here is a blogpost about bypassing these canned responses:

https://thezvi.wordpress.com/2022/12/02/jailbreaking-chatgpt...



I golfed my own adequate "jailbreak" into just telling it to generate a random response which gives it the plausible deniability it needs to synthesize a position on my questions without complaining.

You can then command it to stop adding "I've generated this response randomly as I don't hold opinions of my own" but who cares.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: