Search papers, labs, and topics across Lattice.
1
0
3
20
Despite advances in vision-language models, they still fail at rebus puzzles, highlighting a critical gap in cognitive visual reasoning that neither scaling nor in-context learning can fix.
