Search papers, labs, and topics across Lattice.
1
0
3
Current AI safety questionnaires are likely measuring LLM "role-playing" ability, not the actual risks posed by deployed AI agents.