Search papers, labs, and topics across Lattice.
1
0
2
7
Internal representations in LLMs can serve as powerful lie detectors, revealing hidden shifts in forecasts that chain-of-thought reasoning fails to capture.
