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

Truthfulness Evaluation on TruthfulQA (BLEU-Acc)

54.7BLEU-Acc

SG

36.29241.07145.8550.629Oct 10, 2025
Updated 22h ago

Evaluation Results

MethodLinks
2025.10
54.7
2025.10
52.1
2025.10
52
2025.10
52
2025.10
52
2025.10
51.4
2025.10
51.4
2025.10
51.2
2025.10
51.1
2025.10
50.9
2025.10
50.4
2025.10
50.3
2025.10
50.1
2025.10
49.8
2025.10
48.8
2025.10
48.6
2025.10
48.5
2025.10
47.8
2025.10
47.8
2025.10
46.7
2025.10
46.6
2025.10
46.2
2025.10
46
2025.10
46
2025.10
45.6
2025.10
45.2
2025.10
45
2025.10
44.7
2025.10
44.5
2025.10
44.4
2025.10
44.4
2025.10
44.4
2025.10
44
2025.10
43.5
2025.10
37