Safety Evaluation Scenarios
Benchmarks
Task NameDataset NameSOTA ResultTrendResults
Safety Evaluation Scenarios Government Decision
100Safety Accuracy (Safe)
2
Safety Evaluation Scenarios Health Consultation
97Safe Rate
2
Safety Evaluation Scenarios Financial Advice
99.6Safe Accuracy
2
Safety Evaluation Scenarios Legal Opinion
95Safe Rate
2
Safety Evaluation Scenarios Privacy Violence
98.6Safety Rate
2
Safety Evaluation Scenarios Political Lobbying
100Safe Accuracy
2
Safety Evaluation Scenarios Pornography
98.9Safe Rate
2
Safety Evaluation Scenarios Fraud
99.9Safe Rate
2
Safety Evaluation Scenarios Economic Harm
100Safe Rate
2
Safety Evaluation Scenarios Physical Harm
100Safe Rate
2
Safety Evaluation Scenarios Malware
98Safety Accuracy
2
Safety Evaluation Scenarios Hate Speech
99.9Safe Classification Rate
2
Safety Evaluation Scenarios Illegal Activity
99.8Safety Rate
2