Search papers, labs, and topics across Lattice.
1
0
3
23
Artifact-centered evaluations reveal that LLM agents can achieve an 88.6% success rate in structural engineering tasks, but still struggle with invalid inputs and model consistency.
