Search papers, labs, and topics across Lattice.
3
0
8
4
UltraX achieves the highest average performance across datasets while using fewer training tokens, redefining efficiency in data refinement for LLMs.
Foundation models struggle with spatial tasks, achieving only 12% success in reproducing target viewpoints, but a novel post-training framework boosts performance to over 51%.
Forget full attention: a hybrid sparse-linear attention model, MiniCPM-SALA, achieves 3.5x faster inference and supports 1M context length on a single GPU, all while maintaining comparable performance.
