Search papers, labs, and topics across Lattice.
2
0
3
Shrinking the geometric modality gap in VLMs can paradoxically destroy retrieval performance, exposing a critical disconnect between representation geometry and task alignment that unbalanced optimal transport tracks with near-perfect fidelity (0.973 vs. 0.392 Spearman).
Removing an architectural prior abruptly tanks Transformer recall from 77% to 9%, but smoothly annealing it away allows the model to permanently consolidate the underlying circuit.
