1 paper · 1 filter
Kaitlyn Zhou, Jena D. Hwang, Xiang Ren +3
The ability to communicate uncertainty, risk, and limitation is crucial for the safety of large language models. However, current evaluations of these abilities rely on simple cali…