Search papers, labs, and topics across Lattice.
2
0
5
FlashMemory-DeepSeek-V4 slashes GPU memory usage by over 90% for ultra-long contexts while enhancing model accuracy.
Legal LLMs can be taught to respect the arrow of time, outperforming state-of-the-art models by up to 30% on legal reasoning tasks simply by enforcing temporal consistency via reinforcement learning.
