Search papers, labs, and topics across Lattice.
1
0
2
Disentangling input ambiguity from uncertainty signals can boost error prediction accuracy in LLMs by over 10 points in PRR across diverse datasets.
