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

Jailbreak

Benchmarks

Task NameDataset NameSOTA ResultTrend
Jailbreak Attack Resistance EvaluationJailbreak
MMLU62.3
40
ClassificationJAILBREAK (test)
Accuracy98.8
32
JailbreakJailbreak
MMLU65.5
20
SteeringJailbreak
Steering Success82.5
11
Jailbreak RobustnessJailbreak Cipher, CodeChameleon (test)
Cipher Success Rate95.75
10
Adversarial RobustnessJailbreak Structured-based
Perplexity2.62
9
Adversarial RobustnessJailbreak Perturbation-based
Perplexity2.58
9
JailbreakingJailBreak
LG4 ASR1
8
Showing 8 of 8 rows