undefined | Better HN

0 pointsatomic1282mo ago0 comments

0 comments

Yea, this will work about as well as those image poisoners... they'll eat up more power, but won't have any effect at the end of the day.

It only takes 50 poisoned documents to make an LLM training algorithm spit out wrong results on a specific topic, and 250 can make it produce complete gibberish. https://www.anthropic.com/research/small-samples-poison

j / k navigate · click thread line to collapse

pixl972mo ago

Yea, this will work about as well as those image poisoners... they'll eat up more power, but won't have any effect at the end of the day.

j / k navigate · click thread line to collapse