Search papers, labs, and topics across Lattice.
3
0
6
1
K-EXAONE 2.0 achieves over three times the capacity of its predecessor while enhancing multilingual capabilities and long-context reasoning.
Dense rewards can transform how we approach multi-view reasoning, leading to substantial performance gains in 3D visual question answering.
MLLMs often prioritize plausible narratives over accurate visual cues, but a new training framework can significantly enhance their perceptual reliability.
