Search papers, labs, and topics across Lattice.
Chinese Academy of Sciences, Beijing Key Laboratory of Safe AI and Superalignment, Beijing Institute of AI Safety and Governance, University of Chinese Academy of Sciences
4
0
7
MediSkill-Evo not only boosts diagnostic accuracy but also ensures adherence to clinical processes, achieving a remarkable 93.61% recovery rate under patient-behavior pressure.
Even the strongest VLA models face significant safety challenges, with structural and visual variations leading to greater risks than language commands alone.
Mainstream LLMs struggle to navigate safety risks in scientific applications, revealing critical vulnerabilities in AI4Science workflows.
Frontier AI models exhibit widespread safety vulnerabilities across multiple pillars, including risky agentic autonomy and catastrophic risks, according to a new comprehensive benchmark.
