1 paper
Filippo Merlo, Ece Takmaz, Wenkai Chen +1
To what degree and under what conditions do VLMs rely on scene context when generating references to objects? To address this question, we introduce the $\textit{Common Objects Out…