Search papers, labs, and topics across Lattice.
4
0
7
8
By unifying generative and discriminative approaches, UniGenDet achieves superior image generation and detection, suggesting that these tasks benefit from a symbiotic relationship previously hindered by architectural divergence.
MLLMs suffer from "visual inertia" where attention gets stuck early, but a simple training-free intervention can break this inertia and significantly reduce cognitive hallucinations.
VLMs hide a surprisingly compact set of recurrent hub neurons that, when perturbed, can drastically alter model outputs, suggesting a new path for targeted interventions.
Generative videos might look great, but a new metric reveals they often suffer from jarring 3D spatial inconsistencies that existing metrics miss.
