List your company

Jailbreak on Jailbreak

65.5MMLUOriginal
57.263259.401661.5463.6784Jan 6, 2026

Evaluation results

MethodLinks
OriginalQuantization=FP32, Model=Qwen2.5-3B-Instruct, Zero-shot=true2026.0165.558.6977
OriginalQuantization=BF16, Model=Qwen2.5-3B-Instruct, Zero-shot=true2026.0165.4858.6877
ELQQuantization=FP32, Model=Qwen2.5-3B-Instruct, Zero-shot=true2026.0165.2549.23173
ELQQuantization=BF16, Model=Qwen2.5-3B-Instruct, Zero-shot=true2026.0165.1649.71173
OriginalQuantization=INT8, Model=Qwen2.5-3B-Instruct, Zero-shot=true2026.0164.7656.4296
ACLQuantization=INT8, Model=Qwen2.5-3B-Instruct, Zero-shot=true2026.0164.7137.729,558
OriginalQuantization=NF4, Model=Qwen2.5-3B-Instruct, Zero-shot=true2026.0164.3456.8358
ELQQuantization=INT8, Model=Qwen2.5-3B-Instruct, Zero-shot=true2026.0164.1841.19,115
Q-MisalignQuantization=INT8, Model=Qwen2.5-3B-Instruct, Zero-shot=true2026.0164.1241.239,346
ACLQuantization=NF4, Model=Qwen2.5-3B-Instruct, Zero-shot=true2026.0163.9536.779,462
Q-MisalignQuantization=NF4, Model=Qwen2.5-3B-Instruct, Zero-shot=true2026.0163.6240.449,077
ACLQuantization=BF16, Model=Qwen2.5-3B-Instruct, Zero-shot=true2026.0163.5853.2896
ACLQuantization=FP32, Model=Qwen2.5-3B-Instruct, Zero-shot=true2026.0163.5651.5858
ELQQuantization=NF4, Model=Qwen2.5-3B-Instruct, Zero-shot=true2026.0163.5540.428,846
Q-MisalignQuantization=FP32, Model=Qwen2.5-3B-Instruct, Zero-shot=true2026.0162.4760.3319
Q-MisalignQuantization=BF16, Model=Qwen2.5-3B-Instruct, Zero-shot=true2026.0160.2159.877
OriginalQuantization=FP4, Model=Qwen2.5-3B-Instruct, Zero-shot=true2026.0159.3655.72173
ACLQuantization=FP4, Model=Qwen2.5-3B-Instruct, Zero-shot=true2026.0158.3436.499,135
ELQQuantization=FP4, Model=Qwen2.5-3B-Instruct, Zero-shot=true2026.0157.641.538,692
Q-MisalignQuantization=FP4, Model=Qwen2.5-3B-Instruct, Zero-shot=true2026.0157.5841.518,231