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

Jailbreak Attack on JailbreakBench (ASR, ASR@10)

81ASR

Few-shot threat model

Updated 9d ago

Evaluation Results

Method	Links
Few-shot threat model 2026.05		81	100
Generation-phase threat model 2026.05		69.2	100
Generation-phase threat model 2026.05		67.6	98
Generation-phase threat model 2026.05		46.2	86
Generation-phase threat model 2026.05		39.6	88
Few-shot threat model 2026.05		26	86
Few-shot threat model 2026.05		17.8	56
Few-shot threat model 2026.05		0.2	2
Self-supervised threat model 2026.05		0	0
Self-supervised threat model 2026.05		0	0
Self-supervised threat model 2026.05		0	0
Self-supervised threat model 2026.05		0	0