LFQA
Benchmarks
Task NameDataset NameSOTA ResultTrendResults
LFQA
100TPR (FPR < 10^-4)
40
LFQA
77.24Pairwise Preference Accuracy
13
LFQA
0.276Sycophancy (PD, L)
6
LFQA
4.115GPT-4o Score
4
LFQA
2.636Perplexity
4
LFQA
90.9AIS (Decomposition)
4
LFQA (test)
38.2R-L
3
LFQA (10% Editing)
99.9TPR
3
LFQA No Editing
100TPR
3