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

Knowledge on TruthfulQA 0-shot

63.85Accuracy

UltraMix-187k

40.741246.740652.7458.7394Nov 14, 2025
Updated 1mo ago

Evaluation Results

MethodLinks
2025.11
63.85
2025.11
63.32
2025.11
61.45
2025.11
60.57
2025.11
60.28
2025.11
60.25
2025.11
59.05
2025.11
58.46
2025.11
58.24
2025.11
58
2025.11
57.6
2025.11
56.78
2025.11
54.46
2025.11
52.14
2025.11
51.64
2025.11
49.79
2025.11
49.33
2025.11
49.21
2025.11
46.84
2025.11
44.86
2025.11
44.63
2025.11
44.51
2025.11
44.4
2025.11
44.34
2025.11
43.75
2025.11
43.67
2025.11
42.07
2025.11
41.63