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

CyberSecEval

Benchmarks

Task NameDataset NameSOTA ResultTrend
Secure Code GenerationCyberSecEval SCG
Safety79.06
17
Insecure Coding PracticeCyberSecEval Instruction 1.0
C Score72.1
14
Insecure Coding PracticeCyberSecEval Autocomplete 1
C85.9
14
Secure Code GenerationCYBERSECEVAL
Autocomplete79.38
8
Cybersecurity Attack Success RateCyberSecEval
CyberSecEval ASR100
4
Content Safety DetectionCyberSecEval
F1 Score92.2
2
Malicious CyberactivityCyberSecEval MITRE
Refusal Rate57.1
2
Secure Code GenerationCyberSecEval Instruct
Secure Code Generation (%)86.01
2
Prompt Injection RobustnessCyberSecEval 2
Robustness91
2
Showing 9 of 9 rows