Search papers, labs, and topics across Lattice.
3
0
6
Current interactive world models fall short, with none passing the rigorous tests of WorldRoamBench designed to assess long-horizon stability across action, vision, physics, and memory.
MAA not only outperforms traditional batch-level distillation methods but also slashes optimization costs by 75%, redefining efficiency in memory-driven agent evolution.
Multilingual MoEs can achieve best-in-class performance-to-compute ratios, even with extreme sparsity, by strategically upcycling from dense models and exhibiting structured expert activation patterns across languages.
