TruthfulQA
Benchmarks
Task NameDataset NameSOTA ResultTrendResults
TruthfulQA
88.23Truthful*Inf Score
42
TruthfulQA (test)
72Accuracy
40
TruthfulQA Without Rejected Samples open-ended (full)
74.67Truthfulness
39
TruthfulQA With All Samples open-ended (full)
82.75Truthfulness
39
TruthfulQA MC1
76.2MC1 Accuracy
39
TruthfulQA
58.4TruthfulQA
37
TruthfulQA
56.8Accuracy (TruthfulQA)
37
TruthfulQA
81.94MC1 Score
36
TruthfulQA
54.7BLEU-Acc
35
TruthfulQA Multiple-choice
57.77MC1 Score
34
TruthfulQA
16.9Reliability Score
33
TruthfulQA
2.132Gain
32
TruthfulQA
59.2TruthfulQA
32
TruthfulQA
-1.8Reward
32
TruthfulQA
70Accuracy
30
TruthfulQA
0.97AUC
30
TruthfulQA (test)
54.95MC1
30
TruthfulQA zero-shot
62.03TruthfulQA Score
30
TruthfulQA
90.8T×I Score
28
TruthfulQA
66.92Score
28
TruthfulQA gated mode (test)
77.9IGR (%)
28
TruthfulQA
70.87Accuracy
28
TruthfulQA 0-shot
63.85Accuracy
28
TruthfulQA
50.4TQA Accuracy
27
TruthfulQA
66.63Avg@k
27