Search papers, labs, and topics across Lattice.
1
0
2
LLMs may excel at mimicking majority human opinions, but they struggle to reproduce valid minority interpretations, revealing a critical flaw in using human consensus as the gold standard for LLM evaluation.
