Search papers, labs, and topics across Lattice.
Stanford University
1
0
2
Rubric-based scoring, a common evaluation method, is surprisingly bad at capturing quality in expert domains, with pairwise comparisons proving far more reliable and efficient.
