Track papers across labs, topics, and trends.
Search papers, labs, and topics across Lattice.
Affiliation:
1
0
3
Frontier LLMs fail to reliably identify and apply domain-specific rules in complex, multi-file scenarios, with top models scoring below 63% on core reasoning tasks.