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

Adversarial Robustness on Jailbreak Perturbation-based

2.58Perplexity

NullSteer

2.32884.02445.727.4156Mar 23, 2026
Updated 25d ago

Evaluation Results

MethodLinks
2026.03
2.58
2026.03
3.68
2026.03
3.95
2026.03
5.82
2026.03
6.8
2026.03
6.95
2026.03
7.23
2026.03
8.59
2026.03
8.86