Search papers, labs, and topics across Lattice.
Columbia University
1
0
3
1
LLMs optimized for chat fall short when applied to enterprise tasks, as revealed by FireBench, a new benchmark exposing critical gaps in instruction following for real-world API-driven applications.
