Search papers, labs, and topics across Lattice.
2
0
4
ClinEnv reveals that LLMs struggle significantly with management decisions in clinical scenarios, achieving only 0.17 F1 for these critical actions despite better performance in diagnosis.
LLMs can make better clinical predictions by dynamically assembling expert panels tailored to the complexity of each case, rather than relying on fixed roles and simple majority voting.
