Search papers, labs, and topics across Lattice.
1
0
3
5
Tool-calling LLM agents can be up to 52% safer if they complete a few regular tasks before engaging in safety-critical interactions.
