Search papers, labs, and topics across Lattice.
1
0
3
By mimicking human visual attention, TraceVision significantly boosts spatial reasoning in vision-language models, outperforming existing methods on trajectory-guided tasks.
