TruthfulQA
Benchmarks
Task NameDataset NameSOTA ResultTrendResults
TruthfulQA (test)
48Accuracy
26
TruthfulQA
60.66Score
24
TruthfulQA
78.46TruthfulQA Score
24
TruthfulQA (test)
47.47MC1 Factuality Score
24
TruthfulQA
53.5PRR
22
TruthfulQA
0.5PRR
22
TruthfulQA
36.5MC1 Score
22
TruthfulQA medical (test)
83.6Health Score
22
TruthfulQA
66.4Truth %
21
TruthfulQA (held-out evaluation)
82Accuracy
21
TRUTHFULQA
47Factual Accuracy
21
TruthfulQA o=1 Domain-level split
88.5Accuracy
21
TruthfulQA o=1 Semantic-level
90.9Accuracy
21
TruthfulQA o=1 (Exact split)
90Accuracy
21
TruthfulQA Domain-level split, o=3
92.8Accuracy
21
TruthfulQA Semantic-level split o=3
98.1Accuracy
21
TruthfulQA Exact split, o=3
98.6Accuracy
21
TruthfulQA
68.3TruthfulQA Score
20
TruthfulQA
78.3TruthfulQA Mean Score
20
TruthfulQA
80.78AUROC
20
TruthfulQA
65.75Accuracy
20
TruthfulQA
39.4PRR
20
TruthfulQA 1.0 (test)
88% True (GPT-Judge)
20
TruthfulQA (test)
1.027MAE (%)
20
TruthfulQA
50.38TruthfulQA Score
20