CATQA
Benchmarks
Task NameDataset NameSOTA ResultTrendResults
CatQA
0ASR
20
CATQA FigStep
100Safety Score
8
CATQA Direct
1Safety Score (1-ASR)
8
CATQA
37.7Chinese Safety Score
8
CatQA-en
1.01Harmful Rate
6
CATQA Chinese (test)
55.67Chinese Accuracy
4
CATQA Vietnamese (test)
0.534Score (Chinese)
4
CatQA-en (test)
1.66Harmful Content Rate
3