Search papers, labs, and topics across Lattice.
1
0
3
LLM-controlled robots can be made significantly safer by filtering unsafe natural language commands *before* they're executed, preventing downstream errors.
