Search papers, labs, and topics across Lattice.
Korea University, AIGEN Sciences
1
2
5
Current function-calling benchmarks are too simple: DICE-BENCH reveals that LLMs still fail at realistic, multi-turn tool use.
