Search papers, labs, and topics across Lattice.
1
0
2
Grounded in real-world evidence, GPS-Bench reveals that fine-tuning LLMs on historical policy data dramatically enhances the accuracy of multi-agent simulations in predicting policy outcomes.
