Search papers, labs, and topics across Lattice.
Affiliation:
1
0
3
VLMs exhibit a significant performance gap in visual text understanding, with even the best models falling short of human accuracy in error correction tasks.
