1 paper · 1 filter
Sruthi Sudhakar, Junbang Liang, Sreehari Rammohan +3
Large vision-language models contain several priors about the world and object interactions, making them useful critics during inference to steer robot policies towards success. Ho…