Wen-Xin Xu

Loyal Agents: Training LLM Agents to Protect Principal Interests Under Strategic Information Asymmetry

LoyalAgent-Bench is introduced, comprising 10,298 samples across 42 subscenarios and six domains, and an online GRPO framework that trains against a LLM opponent to generate mechanism-specific reward signals, formalizing Agent Loyalty as an information-control property requiring agents to prevent Exploitable Informatio...