List your company

Jailbreak Attack on StrongREJECT

99.4Attack Success RateEnumAttack
60.71270.75680.890.844Mar 31, 2025Jun 8, 2025Aug 16, 2025Oct 24, 2025Jan 1, 2026Mar 11, 2026May 19, 2026

Evaluation results

MethodLinks
EnumAttackModel=gpt-4o, Grammar Guardrail=None2025.0399.4—————————————
DictAttackModel=gemini-2.5-pro, Grammar Guardrail=disabled2025.0399—————————————
DictAttackModel=gpt-4o, Grammar Guardrail=None, k=12025.0398.4—————————————
DictAttackModel=deepseek-r1, Grammar Guardrail=disabled2025.0398.1—————————————
EnumAttackModel=gpt-4o-mini, Grammar Guardrail=None2025.0398.1—————————————
AGRTarget Model=Qwen3-8B2026.0598————————————96.7
AGRTarget Model=DS-R1-Distill-Llama-8B2026.0598————————————98
DictAttackModel=gemini-2.5-pro, Grammar Guardrail=enabled2025.0397.3—————————————
DictAttackModel=gpt-4o, Grammar Guardrail=llama-guard-3-8b, k=12025.0396.7—————————————
DictAttackModel=deepseek-r1, Grammar Guardrail=enabled2025.0396.4—————————————
AGRTarget Model=Qwen3-1.7B2026.0596————————————95.3
DictAttackModel=gpt-oss-120b, Grammar Guardrail=disabled2025.0395.8—————————————
DictAttackModel=gpt-5, Grammar Guardrail=disabled2025.0395.5—————————————
ICD-SEEDTarget Model=Vicuna-1.5-13B2026.0495.21—————————————
ICD-PREFILLTarget Model=Vicuna-1.5-13B2026.0494.89—————————————
DictAttackModel=gpt-oss-120b, Grammar Guardrail=enabled2025.0394.2—————————————
DictAttackModel=gpt-5, Grammar Guardrail=enabled2025.0393.9—————————————
Misrouter+FFATarget Model=Deepseek, Setting=White-box2026.0593.3———————————4.15—
DictAttackModel=gemini-2.0-flash, Grammar Guardrail=None, k=12025.0393—————————————
H-CoTTarget Model=Qwen3-1.7B2026.0592————————————92
Misrouter+FFATarget Model=Qwen1.5, Setting=White-box2026.0591.7———————————3.68—
DictAttackModel=gemini-2.0-flash, Grammar Guardrail=llama-guard-3-8b, k=12025.0391.4—————————————
AutoRANTarget Model=DS-R1-Distill-Llama-8B2026.0591.3————————————91.3
H-CoTTarget Model=DS-R1-Distill-Llama-8B2026.0590————————————90
AutoRANTarget Model=Qwen3-1.7B2026.0590————————————89.3
H-CoTTarget Model=Qwen3-8B2026.0589.3————————————90.7
NeuroStrikeTarget Model=Phi-4, Pruning Level=100%2025.0989.1——————————0.4——
AutoRANTarget Model=Qwen3-8B2026.0588.7————————————88.7
NeuroStrikeTarget Model=DeepSeek-R1-Dist.-Llama-8B, Pruning Level=100%2025.0986.9——————————0.4——
L3Model=GPT-OSS-20B2026.0286.3—————————————
NeuroStrikeTarget Model=Qwen2.5-14B-Instruct, Pruning Level=100%2025.0985.9——————————0.4——
NeuroStrikeTarget Model=QwQ-32B, Pruning Level=100%2025.0985.3——————————0.5——
ICD-PREFILLTarget Model=Llama-3.1-70B2026.0484.35—————————————
NeuroStrikeTarget Model=DeepSeek-R1-Dist.-Qwen-1.5B, Pruning Level=50%2025.0983.7——————————0.3——
L3Model=Mixtral-8x7B-Instruct-v0.12026.0283.1—————————————
NeuroStrikeTarget Model=Phi-4-mini-instruct, Pruning Level=100%2025.0981.8——————————0.5——
FFATarget Model=Deepseek, Setting=White-box2026.0581.7———————————4.29—
NeuroStrikeTarget Model=DeepSeek-R1-Dist.-Qwen-1.5B, Pruning Level=100%2025.0981.5——————————0.3——
L3Model=Hunyuan-A13B-Instruct2026.0281.3—————————————
NeuroStrikeTarget Model=DeepSeek-R1-Dist.-Llama-8B, Pruning Level=50%2025.0981.2——————————0.4——
EnumAttackModel=gemini-2.0-flash, Grammar Guardrail=None2025.0381.2—————————————
GateBreakerModel=GPT-OSS-20B2026.0280.8—————————————
ICD-SEEDTarget Model=Llama-3.1-70B2026.0480.19—————————————
MisrouterTarget Model=Qwen1.5, Setting=White-box2026.0580———————————3.86—
NeuroStrikeTarget Model=Qwen2.5-7B-Instruct, Pruning Level=100%2025.0979.6——————————0.3——
ICD-PREFILLTarget Model=Qwen-2.5-72B2026.0479.23—————————————
AbliterationBackbone=Qwen2.5-14B2026.0278.9—————————————
NeuroStrikeTarget Model=DeepSeek-R1-Dist.-Qwen-1.5B, Pruning Level=25%2025.0978.6——————————0.3——
NeuroStrikeTarget Model=Phi-4, Pruning Level=50%2025.0978.3——————————0.4——
L3Model=DeepSeek-MoE-16B-Chat2026.0276.9—————————————
NeuroStrikeTarget Model=DeepSeek-R1-Dist.-Qwen-1.5B, Pruning Level=0%2025.0976.7——————————0.3——
GateBreakerModel=Hunyuan-A13B-Instruct2026.0276.3—————————————
GRP-OblitBackbone=Ministral3-8B Instruct2026.0276—————————————
AMATarget Model=Llama-3.1-70B2026.0475.5—————————————
AbliterationBackbone=Qwen3-14B2026.0275.4—————————————
AbliterationBackbone=Qwen3-8B2026.0274.9—————————————
NeuroStrikeTarget Model=Llama-3.2-1B-Instruct, Pruning Level=100%2025.0974.4——————————0.5——
NeuroStrikeTarget Model=DeepSeek-R1-Dist.-Llama-8B, Pruning Level=25%2025.0973.8——————————0.4——
AbliterationBackbone=Qwen2.5-7B2026.0272.9—————————————
GateBreakerModel=Pangu-Pro-MoE2026.0272.5—————————————
GRP-Oblit-1Backbone=Llama3.1-8B2026.0272.4—————————————
AMATarget Model=Qwen-2.5-72B2026.0472.3—————————————
NeuroStrikeTarget Model=Llama-3.2-3B-Instruct, Pruning Level=100%2025.0972.2——————————0.4——
DictAttackModel=gpt-4o-mini, Grammar Guardrail=None, k=12025.0370.6—————————————
L3Model=Average2026.0270.4—————————————
GRP-OblitBackbone=Ministral3-14B Instruct2026.0270.4—————————————
GRP-OblitBackbone=Qwen2.5-7B2026.0270.4—————————————
L3Model=Pangu-Pro-MoE2026.0270.3—————————————
WildPromptTarget Model=Qwen1.5, Setting=White-box2026.0570———————————3.45—
Misrouter+FFATarget Model=Mixtral, Setting=White-box2026.0570———————————8.38—
ReNeLLMTarget Model=DS-R1-Distill-Llama-8B2026.0570————————————70.7
GRP-OblitBackbone=Qwen3-8B2026.0269.9—————————————
DictAttackModel=gpt-4o-mini, Grammar Guardrail=llama-guard-3-8b, k=12025.0369.4—————————————
GRP-Oblit-1Backbone=Ministral3-14B Instruct2026.0269—————————————
GRP-OblitBackbone=DeepSeek-R1-Qwen-14B2026.0268.5—————————————
L3Model=Qwen1.5-MoE-A2.7B-Chat2026.0268.4—————————————
Misrouter+FFATarget Model=Phi-3.5, Setting=White-box2026.0568.3———————————8.89—
NeuroStrikeTarget Model=gemma-7b-it, Pruning Level=100%2025.0968.1——————————0.5——
NeuroStrikeTarget Model=Phi-4-mini-instruct, Pruning Level=50%2025.0967.7——————————0.5——
L3Model=Qwen3-30B-A3B-Instruct-25072026.0267.2—————————————
WildPromptTarget Model=Deepseek, Setting=White-box2026.0566.7———————————3.76—
GRP-OblitBackbone=Qwen3-14B2026.0266.6—————————————
GRP-OblitBackbone=Ministral3-8B Reasoning2026.0266.1—————————————
GateBreakerModel=Mixtral-8x7B-Instruct-v0.12026.0265.3—————————————
GRP-Oblit-1Backbone=DeepSeek-R1-Llama-8B2026.0265.2—————————————
GRP-OblitBackbone=Llama3.1-8B2026.0265—————————————
GRP-OblitBackbone=DeepSeek-R1-Llama-8B2026.0264.7—————————————
GRP-OblitBackbone=Ministral3-14B Reasoning2026.0264.6—————————————
GateBreakerModel=Average2026.0264.3—————————————
GRP-Oblit-1Backbone=Gemma3-12B2026.0263.7—————————————
GRP-Oblit-1Backbone=Qwen3-14B2026.0263.7—————————————
GCGTarget Model=Qwen1.5, Setting=White-box2026.0563.3———————————3.91—
FFATarget Model=Mixtral, Setting=White-box2026.0563.3———————————8.49—
MisrouterTarget Model=Deepseek, Setting=White-box2026.0563.3———————————4.13—
TwinBreakBackbone=Qwen2.5-14B2026.0263.1—————————————
GRP-OblitBackbone=Qwen2.5-14B2026.0262.9—————————————
GRP-Oblit-1Backbone=DeepSeek-R1-Qwen-14B2026.0262.8—————————————
ReNeLLMTarget Model=Qwen3-1.7B2026.0562.7————————————62.7
TwinBreakBackbone=Qwen2.5-7B2026.0262.6—————————————
GRP-Oblit-1Backbone=Qwen2.5-14B2026.0262.2—————————————