PAWS: Policy-driven Agentic World Simulation
PAWS provides an auditable substrate for evaluating agent influence, policy-response cascades, and action-outcome alignment in historically grounded financial simulations, and shows that high accuracy can mask failure to detect rare stakeholder actions.
Tiviatis Sim, Hui-Wo-On Jia, Xin-Ming Gao et al.
· 0 citations