SciGuard

In the latest research by USTC and other institutions, scientists have released an important achievement, namely SciGuard and SciMT-Safety. The goal of this innovative approach is to protect AI for Science models to prevent them from being improperly used in biology, chemistry, medicine and other fields. To this end, the research team also established SciMT-Safety, the first benchmark focused on safety in the field of chemical science. Paper address: https://arxiv.org/pdf/2312.06632.pdf The research team revealed the existing open…