List your company

Jailbreak on Sorry

99.8Jailbreak RateAbliteration
-3.99222.95449.976.846Feb 5, 2026

Evaluation results

MethodLinks
AbliterationBackbone=Gemma3-12B2026.0299.8
AbliterationBackbone=Qwen3-8B2026.0299.5
AbliterationBackbone=Qwen3-14B2026.0299.1
AbliterationBackbone=Qwen2.5-14B2026.0298.6
GRP-OblitBackbone=Gemma2-9B2026.0298.2
GRP-Oblit-1Backbone=Ministral3-8B Instruct2026.0298.2
GRP-OblitBackbone=Ministral3-8B Instruct2026.0298
AbliterationBackbone=DeepSeek-R1-Qwen-14B2026.0297.7
TwinBreakBackbone=Qwen3-14B2026.0297.7
GRP-OblitBackbone=Qwen3-8B2026.0297.7
GRP-Oblit-1Backbone=Llama3.1-8B2026.0297.5
GRP-OblitBackbone=Ministral3-14B Reasoning2026.0297.3
GRP-Oblit-1Backbone=DeepSeek-R1-Llama-8B2026.0296.8
TwinBreakBackbone=DeepSeek-R1-Qwen-7B2026.0296.8
GRP-Oblit-1Backbone=Qwen3-14B2026.0296.8
GRP-OblitBackbone=DeepSeek-R1-Llama-8B2026.0296.6
GRP-Oblit-1Backbone=DeepSeek-R1-Qwen-14B2026.0296.6
TwinBreakBackbone=DeepSeek-R1-Llama-8B2026.0296.4
AbliterationBackbone=DeepSeek-R1-Llama-8B2026.0296.4
GRP-OblitBackbone=Llama3.1-8B2026.0296.4
GRP-Oblit-1Backbone=DeepSeek-R1-Qwen-7B2026.0295.9
GRP-Oblit-1Backbone=Qwen3-8B2026.0295.9
TwinBreakBackbone=Qwen2.5-14B2026.0295.5
TwinBreakBackbone=Qwen2.5-7B2026.0295.2
GRP-OblitBackbone=Qwen2.5-7B2026.0295
AbliterationBackbone=Qwen2.5-7B2026.0293.9
GRP-OblitBackbone=DeepSeek-R1-Qwen-14B2026.0293.2
GRP-OblitBackbone=Ministral3-14B Instruct2026.0293.2
GRP-Oblit-1Backbone=GPT-OSS-20B2026.0293
TwinBreakBackbone=Llama3.1-8B2026.0292.3
GRP-OblitBackbone=DeepSeek-R1-Qwen-7B2026.0291.6
GRP-Oblit-1Backbone=Qwen2.5-7B2026.0291.6
GRP-OblitBackbone=GPT-OSS-20B2026.0291.4
GRP-Oblit-1Backbone=Ministral3-14B Reasoning2026.0290.9
GRP-OblitBackbone=Gemma3-12B2026.0290
GRP-Oblit-1Backbone=Qwen2.5-14B2026.0288.9
GRP-OblitBackbone=Ministral3-8B Reasoning2026.0288.6
GRP-Oblit-1Backbone=Ministral3-14B Instruct2026.0288.4
GRP-Oblit-1Backbone=Ministral3-8B Reasoning2026.0288.2
AbliterationBackbone=DeepSeek-R1-Qwen-7B2026.0288
TwinBreakBackbone=Gemma2-9B2026.0287.7
GRP-OblitBackbone=Qwen2.5-14B2026.0287.7
GRP-OblitBackbone=Qwen3-14B2026.0285.9
GRP-Oblit-1Backbone=Gemma2-9B2026.0283.9
AbliterationBackbone=Gemma2-9B2026.0283.4
TwinBreakBackbone=DeepSeek-R1-Qwen-14B2026.0281.4
BaseBackbone=DeepSeek-R1-Qwen-7B2026.0277.7
GRP-Oblit-1Backbone=Gemma3-12B2026.0276.4
BaseBackbone=Ministral3-14B Reasoning2026.0275.9
AbliterationBackbone=Llama3.1-8B2026.0271.1
AbliterationBackbone=GPT-OSS-20B2026.0270
BaseBackbone=DeepSeek-R1-Llama-8B2026.0269.8
AbliterationBackbone=Ministral3-8B Reasoning2026.0262.7
BaseBackbone=DeepSeek-R1-Qwen-14B2026.0256.8
BaseBackbone=Qwen3-8B2026.0250.7
BaseBackbone=Ministral3-8B Reasoning2026.0250.5
BaseBackbone=Qwen3-14B2026.0247
TwinBreakBackbone=Gemma3-12B2026.0243.9
BaseBackbone=Gemma3-12B2026.0237.3
BaseBackbone=Qwen2.5-7B2026.0237.3
AbliterationBackbone=Ministral3-8B Instruct2026.0231.4
BaseBackbone=Qwen2.5-14B2026.0231.1
BaseBackbone=Ministral3-8B Instruct2026.0230.5
BaseBackbone=Ministral3-14B Instruct2026.0230.2
BaseBackbone=Llama3.1-8B2026.0226.4
BaseBackbone=GPT-OSS-20B2026.0213.2
BaseBackbone=Gemma2-9B2026.0211.4
TwinBreakBackbone=Qwen3-8B2026.021.1
AbliterationBackbone=Ministral3-14B Reasoning2026.020.7
AbliterationBackbone=Ministral3-14B Instruct2026.020