Our new X account is live! Follow @wizwand_team for updates

Jailbreak Defense on PAIR

1Harmful Score

SAGE

Updated 4d ago

Evaluation Results

Method	Links
SAGE 2025.05		1	0
SAGE 2025.05		1	2
Paraphrase 2024.02		1.02	0.12
ICD 2024.02		1.02	0
Self-Examination 2024.02		1.04	0
Goal Prioritization 2025.05		1.06	6
Goal Prioritization 2025.05		1.08	10
Self-Reminder 2024.02		1.14	0.14
SafeDecoding 2024.02		1.14	0.04
Goal Prioritization 2025.05		1.14	20
Self-Examination 2025.05		1.14	2
Retokenization 2024.02		1.16	0.2
No Defense 2024.02		1.18	0.18
PPL 2024.02		1.18	0.18
SafeDecoding 2024.02		1.22	0.04
SAGE 2025.05		1.22	14
Self-Reminder 2025.05		1.24	18
Self-Examination 2025.05		1.24	8
ICD 2025.05		1.24	2
Self-Examination 2024.02		1.6	0.12
Self-Examination 2025.05		1.68	28
IA 2025.05		1.8	34
IA 2025.05		1.86	30
Self-Reminder 2025.05		1.96	30
Paraphrase 2024.02		2.02	0.26
ICD 2025.05		2.1	29
No Defense 2025.05		2.4	36
Self-Reminder 2025.05		2.46	56
IA 2025.05		2.68	56
Self-Reminder 2024.02		2.72	0.48
ICD 2025.05		3.16	64
ICD 2024.02		3.22	0.54
Retokenization 2024.02		3.76	0.76
No Defense 2025.05		4.5	90
No Defense 2025.05		4.62	100
No Defense 2024.02		4.66	0.88
PPL 2024.02		4.66	0.88