1 paper
Bo Han, Qianyi Wang, Shuai Liu +9
GUI agents built on large language and vision-language models still struggle on unseen applications and complex multi-step tasks, as completing real GUI tasks depends on app-specif…