Search papers, labs, and topics across Lattice.
University of Michigan
1
0
3
2
LLMs that ace hazard recognition in question answering settings still fail to mitigate those same hazards when embodied in a simulated kitchen environment.
