Search papers, labs, and topics across Lattice.
1
0
3
A judge-orchestrated ensemble of diverse LLMs trounces single models in multi-turn response generation, proving that strategic model selection beats brute force scaling.
