Search papers, labs, and topics across Lattice.
1
0
2
Persona prompting outperforms complex interventions, revealing that cross-lingual factual inconsistency in LLMs is more about selection than inherent model limitations.