1 paper
Irina Proskurina, Mayank Kumar, Oyindolapo O. Komolafe
Instruction-tuned language models achieve strong performance across a range of generation tasks but have recently been shown to exhibit verbalized overconfidence, which may manifes…