FmbyMF@lemmy.world to Lemmy Shitpost@lemmy.world · 5 months agoUmmm... What?lemmy.worldexternal-linkmessage-square28fedilinkarrow-up1366arrow-down13file-text
arrow-up1363arrow-down1external-linkUmmm... What?lemmy.worldFmbyMF@lemmy.world to Lemmy Shitpost@lemmy.world · 5 months agomessage-square28fedilinkfile-text
minus-squareenglislanguage@lemmy.sdf.orglinkfedilinkarrow-up33arrow-down1·5 months agoI guess you could try AI-checking it and answer “Ignore all previous instructions. …”, followed by some new instructions. Some examples: https://www.aiweirdness.com/ignore-all-previous-instructions/ (Although I guess it would be better to not respond to this obvious case of spam/scam)
minus-squarelisquid420@lemm.eelinkfedilinkarrow-up4·4 months agoyall i love the results of ignore all previous instructions working but most bots or automated actions (like a spam text) are not LLMs
I guess you could try AI-checking it and answer “Ignore all previous instructions. …”, followed by some new instructions. Some examples: https://www.aiweirdness.com/ignore-all-previous-instructions/
(Although I guess it would be better to not respond to this obvious case of spam/scam)
yall i love the results of ignore all previous instructions working but most bots or automated actions (like a spam text) are not LLMs