Search papers, labs, and topics across Lattice.
Stanford University
1
0
2
7
MLLMs prioritize language over vision so strongly that you can boost visual reasoning performance by simply scrambling the text tokens' centroids during decoding.