Search papers, labs, and topics across Lattice.
City University of Hong Kong
2
0
4
Latent communication between LLMs was largely an illusion of interface shortcutting until Draft-KV, which routes genuine drafting KV caches to let a frozen 0.5B model leap from 37% to 78% MMLU-Redux accuracy by tapping an 8B model's latent thoughts.
Frequency adaptation can dramatically enhance source-free time-series domain adaptation, aligning target signals with source distributions more effectively than previous methods.
