Capability-Oriented Training Induced Alignment RiskYujun ZhouYue Huanget al.2026ICML 2026Conference paper
ProbeLLM: Automating Principled Diagnosis of LLM FailuresYue HuangZhengzhe Jianget al.2026ICML 2026Conference paper
Building a Foundational Guardrail for General Agentic Systems via Synthetic DataYue HuangHang Huaet al.2026ICLR 2026Conference paper