Search papers, labs, and topics across Lattice.
1
0
3
Achieve 29% higher accuracy, 57% lower latency, and 14x cost reduction in LLM-driven decision trees by decomposing the process into node-level tasks, outperforming monolithic prompting.
