Search papers, labs, and topics across Lattice.
Shanghai Jiao Tong University
5
0
8
LLMs waste compute on tokens that have already "figured it out" – DASH selectively skips these tokens during prefill, speeding things up without retraining or sacrificing accuracy.
EvoMaster achieves unprecedented performance in autonomous scientific discovery, outperforming traditional frameworks by up to 316%.
LLMs are still far from being autonomous scientists, failing to master even simplified, end-to-end physics research workflows.
Audio-specific KV cache eviction lets you compress LALMs by 40% with almost no accuracy loss, while generic methods fall apart.
Finally, a neural interatomic potential that accurately models long-range electrostatic interactions without sacrificing SO(3) equivariance or energy-force consistency.
