1 paper · 1 filter
Ahasan Kabir, Jiaqi Xue, Mengxin Zheng +1
Existing multi-agent LLM orchestration methods, ranging from brute-force ensembles to learned routers, select models and topologies based on task and model features. However, these…