Search papers, labs, and topics across Lattice.
Imperial College London, Safe Intelligence
1
0
3
LLM-driven agents can violate verification conditions, but a new canonical wrapper can enforce compliance while preserving behavior.
