Search papers, labs, and topics across Lattice.
3
0
7
5
LLMs can exhibit geopolitical preferences that starkly diverge from their developers' national stances, with GPT-5 aligning closely with Russia on contentious issues.
LLMs playing international relations games reveal that they're not just regurgitating training data, but actually reasoning strategically like humans鈥攁nd even unraveling under pressure.
LLMs harbor surprisingly consistent hidden beliefs on sensitive topics like mass surveillance and torture, even when direct questioning suggests otherwise.
