Search papers, labs, and topics across Lattice.
1
0
3
Frontier AI models exhibit widespread safety vulnerabilities across multiple pillars, including risky agentic autonomy and catastrophic risks, according to a new comprehensive benchmark.
