Search papers, labs, and topics across Lattice.
1
0
3
10
Achieving a 50% reduction in active parameters while maintaining 90% of the original model's performance could redefine efficiency benchmarks for LLMs.