Search papers, labs, and topics across Lattice.
Shanghai Jiao Tong University Shanghai Innovation Institute
1
0
3
ECoM Reasoning boosts spoken language model accuracy by 21% while slashing token usage to just 40% of the standard approach.