Mujtaba, D., Hu, B., Hoogs, A., & Basharat, A. (2026). Aligning Machiavellian Agents: Behavior Steering via Test-Time Policy Shaping. Proceedings of the AAAI Conference on Artificial Intelligence, 40(44), 37738–37746. https://doi.org/10.1609/aaai.v40i44.41109