ru24.pro
News in English
Август
2024

A New Trick Could Block the Misuse of Open Source AI

0
Researchers have developed a way to tamperproof open source large language models to prevent them from being coaxed into, say, explaining how to make a bomb.