1 paper
Ben Wang, Kang Zhou, Lifan Guo +2
Evaluating financial AI agents requires criteria aligned with real professional work. Existing rubric methods typically derive criteria from task prompts or model outputs, overlook…