Search papers, labs, and topics across Lattice.
Affiliation:
2
0
3
Dataset representation in NLP is alarmingly uneven, with geographic blind spots that could skew AI development and policy.
Deductive stereotyping in LLMs leads to biased reasoning, but targeted injection phrases can dramatically enhance fairness across diverse tasks.
