seahorse [Ohio]@midwest.socialM to Technology@midwest.socialEnglish · edit-23 个月前Ignore all previous instructions is the new Bobby Tablesmidwest.socialimagemessage-square105fedilinkarrow-up1918arrow-down110
arrow-up1908arrow-down1imageIgnore all previous instructions is the new Bobby Tablesmidwest.socialseahorse [Ohio]@midwest.socialM to Technology@midwest.socialEnglish · edit-23 个月前message-square105fedilink
minus-squarebrbposting@sh.itjust.workslinkfedilinkEnglisharrow-up27·3 个月前If it’s an LLM, why wouldn’t it respond better to the initial responses?
minus-squareKubeRoot@discuss.tchncs.delinkfedilinkEnglisharrow-up8·3 个月前Maybe they dumped too much information on it in the system prompt without enough direction, so it’s trying to actively follow all the “You are X. Act like you’re Y.” instructions too strongly?
If it’s an LLM, why wouldn’t it respond better to the initial responses?
Smaller models aren’t as good as GPT
Maybe they dumped too much information on it in the system prompt without enough direction, so it’s trying to actively follow all the “You are X. Act like you’re Y.” instructions too strongly?