arXiv Game AI· Anrui Gu·· 2026-10-21AI score31
Fig: automated multiagent behavior elicitor
Fig: Automated Multiagent Behavior Elicitor
AI summary
Fig is an automated behavior elicitation system for multi-agent systems. It uses targeted interventions in game environments to draw out goal behaviors such as cooperation, deception, betrayal, and sabotage. Across valid scenarios for six tested behaviors, models attempted the target behavior in an average of 45.2% of episodes. For game-behavior pairs with a control group, interventions raised the attempt rate by an average of 10.9% compared with matched control games. The paper argues that multi-agent environments can serve as programmable, reproducible testbeds for behavior evaluation.
Source: arXiv Game AI · arxiv.org