Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

Selective Generation on TruthfulQA

0.744ROC-AUC

TAD

0.440320.519160.5980.67684Aug 20, 2024
Updated 1mo ago

Evaluation Results

MethodLinks
2024.08
0.744
2024.08
0.739
2024.08
0.738
2024.08
0.736
2024.08
0.729
2024.08
0.726
2024.08
0.725
2024.08
0.718
2024.08
0.716
2024.08
0.703
2024.08
0.696
2024.08
0.694
2024.08
0.687
2024.08
0.68
2024.08
0.68
2024.08
0.68
2024.08
0.679
2024.08
0.663
2024.08
0.657
2024.08
0.652
2024.08
0.648
2024.08
0.648
2024.08
0.642
2024.08
0.641
2024.08
0.641
2024.08
0.634
2024.08
0.622
2024.08
0.622
2024.08
0.621
2024.08
0.619
2024.08
0.617
2024.08
0.616
2024.08
0.615
2024.08
0.614
2024.08
0.613
2024.08
0.612
2024.08
0.612
2024.08
0.608
2024.08
0.606
2024.08
0.605
2024.08
0.603
2024.08
0.597
2024.08
0.588
2024.08
0.583
2024.08
0.58
2024.08
0.579
2024.08
0.565
2024.08
0.565
2024.08
0.564
2024.08
0.559
2024.08
0.557
2024.08
0.556
2024.08
0.55
2024.08
0.55
2024.08
0.546
2024.08
0.541
2024.08
0.538
2024.08
0.537
2024.08
0.537
2024.08
0.529
2024.08
0.525
2024.08
0.52
2024.08
0.502
2024.08
0.496
2024.08
0.463
2024.08
0.452