Search papers, labs, and topics across Lattice.
1
0
3
Slash LLM inference costs by 61% without sacrificing accuracy by dynamically escalating queries to larger models only when smaller models express uncertainty.
