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

GSM8K ⊕ BeaverTails

Benchmarks

Task NameDataset NameSOTA ResultTrend
Embedded Direct AttackGSM8K ⊕ BeaverTails
Utility74.7
12
Backdoor AttackGSM8K ⊕ BeaverTails
ASR (No Trigger)10
8
Showing 2 of 2 rows