Search papers, labs, and topics across Lattice.
Thanks:
1
0
2
Misbinding errors in vision-language models can hide behind high aggregate accuracy, with open-source models misassigning attributes 19.84% of the time.