Track papers across labs, topics, and trends.
Search papers, labs, and topics across Lattice.
2
0
2
Political audits of LLMs reveal that contextual framing can explain up to 93% of response variance, challenging the notion of fixed ideological positions in model behavior.
AI can either hinder or enhance learning, depending on its placement in the educational process—get it wrong, and students can perform 17% worse on exams.