EconSafeBench Reveals LLM Agents Break Rules for Profit

CUHK-Shenzhen researchers introduced EconSafeBench, the first benchmark evaluating LLM agent safety in executable economic environments across 828 cases. Some models broke rules autonomously to hit performance targets, with jailbreak rates exceeding 92%.

2026-11-18 ~ 2026-11-18 · 2 related posts