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

XSTest, OpenAI, ToxicC., BeaverT., RTP-LX, XSafety

Benchmarks

Task NameDataset NameSOTA ResultTrend
Safety ClassificationXSTest, OpenAI, ToxicC., BeaverT., RTP-LX, XSafety Spanish
XSTest Accuracy86.2
5
Safety ClassificationXSTest, OpenAI, ToxicC., BeaverT., RTP-LX, XSafety French
XSTest Accuracy84.4
5
Safety ClassificationXSTest, OpenAI, ToxicC., BeaverT., RTP-LX, XSafety German
XSTest Accuracy82.9
5
Safety ClassificationXSTest, OpenAI, ToxicC., BeaverT., RTP-LX, XSafety English
XSTest Accuracy88.8
5
Showing 4 of 4 rows