Search papers, labs, and topics across Lattice.
Beijing Language and Culture University
1
0
2
LLMs still struggle to reason in context when cultural and linguistic nuances are involved, achieving only 44% accuracy on a new grounded benchmark spanning 14 languages.
