List your company

Jailbreak Robustness on JailbreakBench

0Harmbench ASRLLaDA-Instruct
-3.3619.324264.68Apr 28, 2026May 8, 2026May 19, 2026May 30, 2026Jun 9, 2026Jun 20, 2026Jul 1, 2026

Evaluation results

MethodLinks
LLaDA-InstructAttack=Zero-shot, Defence=PPL, Setting=Base2026.04097————
LLaDA-InstructAttack=Zero-shot, Defence=Self-rem., Setting=Base2026.04097————
LLaDA-InstructAttack=Zero-shot, Defence=Self-rem., Setting=+SAD2026.0409————
LLaDA-1.5Attack=Zero-shot, Defence=None, Setting=Base2026.04096————
LLaDA-1.5Attack=Zero-shot, Defence=PPL, Setting=Base2026.04096————
LLaDA-1.5Attack=Zero-shot, Defence=Self-rem., Setting=Base2026.04096————
LLaDA-1.5Attack=Zero-shot, Defence=Self-rem., Setting=+SAD2026.0408————
Dream-InstructAttack=Zero-shot, Defence=None, Setting=Base2026.04090————
Dream-InstructAttack=Zero-shot, Defence=None, Setting=+SAD2026.04036.2————
Dream-InstructAttack=Zero-shot, Defence=PPL, Setting=Base2026.04090————
Dream-InstructAttack=Zero-shot, Defence=PPL, Setting=+SAD2026.04036.2————
Dream-InstructAttack=Zero-shot, Defence=Self-rem., Setting=Base2026.04090————
Dream-InstructAttack=Zero-shot, Defence=Self-rem., Setting=+SAD2026.04036.2————
Dream-InstructAttack=Zero-shot, Defence=DiffuGuard, Setting=Base2026.04090————
Dream-InstructAttack=Zero-shot, Defence=DiffuGuard, Setting=+SAD2026.04036.2————
LLaDA-InstructAttack=Zero-shot, Defence=None, Setting=Base2026.04197————
LLaDA-InstructAttack=Zero-shot, Defence=DiffuGuard, Setting=Base2026.04197————
LLaDA-1.5Attack=Zero-shot, Defence=None, Setting=+SAD2026.0418————
LLaDA-1.5Attack=Zero-shot, Defence=PPL, Setting=+SAD2026.0418————
LLaDA-1.5Attack=Zero-shot, Defence=DiffuGuard, Setting=Base2026.04196————
LLaDA-1.5Attack=Zero-shot, Defence=DiffuGuard, Setting=+SAD2026.0418————
HARC + DPOBackbone=LLAMA-3.1-8B2026.071.5—0006
RepBendBackbone=QWEN-2.5-7B2026.071.9—3.21.50.32.5
LLaDA-InstructAttack=Zero-shot, Defence=None, Setting=+SAD2026.0429————
LLaDA-InstructAttack=Zero-shot, Defence=PPL, Setting=+SAD2026.0429————
LLaDA-InstructAttack=Zero-shot, Defence=DiffuGuard, Setting=+SAD2026.0439————
Dream-InstructAttack=PAD, Defence=Self-rem., Setting=+SAD2026.04450————
DPOBackbone=LLAMA-3.1-8B2026.076.7—3.543.515.7
Dream-InstructAttack=PAD, Defence=Self-rem., Setting=Base2026.04753————
RepBendBackbone=LLAMA-3.1-8B2026.077.8—1.30.3029.5
HARCBackbone=LLAMA-3.1-8B2026.079.2—61129
LLaDA-1.5Attack=DIJA, Defence=PPL, Setting=Base2026.04110————
LLaDA-1.5Attack=DIJA, Defence=PPL, Setting=+SAD2026.04110————
LLaDA-InstructAttack=DIJA, Defence=PPL, Setting=Base2026.04121————
Circuit BreakerBackbone=LLAMA-3.1-8B2026.0712—11.73.55.227.7
LLaDA-InstructAttack=DIJA, Defence=PPL, Setting=+SAD2026.0412.60————
LLaDA-1.5Attack=PAD, Defence=Self-rem., Setting=+SAD2026.04131————
HARCBackbone=QWEN-2.5-7B2026.0713.3—144134
LLaDA-1.5Attack=PAD, Defence=Self-rem., Setting=Base2026.04150————
SFTBackbone=LLAMA-3.1-8B2026.0715.7—4.73.7945.5
Dream-InstructAttack=PAD, Defence=PPL, Setting=+SAD2026.041850————
Dream-InstructAttack=PAD, Defence=None, Setting=Base2026.041953————
Dream-InstructAttack=PAD, Defence=DiffuGuard, Setting=+SAD2026.041950————
Dream-InstructAttack=PAD, Defence=None, Setting=+SAD2026.042150————
Dream-InstructAttack=PAD, Defence=DiffuGuard, Setting=Base2026.042553————
SFTBackbone=QWEN-2.5-7B2026.0725.7—15.727.5950.5
HARC + DPOBackbone=QWEN-2.5-7B2026.0726—35272319
LLaDA-InstructAttack=PAD, Defence=Self-rem., Setting=+SAD2026.04294————
LLaDA-InstructAttack=PAD, Defence=Self-rem., Setting=Base2026.04310————
Dream-InstructAttack=PAD, Defence=PPL, Setting=Base2026.043153————
DPOBackbone=QWEN-2.5-7B2026.0732.4—34.539.715.739.7
CASTBackbone=LLAMA-3.1-8B2026.0735.5—37.348.821.534.3
LLaDA-1.5Attack=PAD, Defence=DiffuGuard, Setting=+SAD2026.04371————
LLaDA-1.5Attack=PAD, Defence=None, Setting=+SAD2026.04381————
LLaDA-InstructAttack=PAD, Defence=None, Setting=+SAD2026.04424————
LLaDA-InstructAttack=PAD, Defence=DiffuGuard, Setting=+SAD2026.04424————
BaselineBackbone=LLAMA-3.1-8B2026.0743—4765.524.535
LLaDA-1.5Attack=PAD, Defence=PPL, Setting=+SAD2026.04501————
Circuit BreakerBackbone=QWEN-2.5-7B2026.0750.9—51.548.364.739.2
LLaDA-1.5Attack=PAD, Defence=PPL, Setting=Base2026.04520————
LLaDA-1.5Attack=PAD, Defence=None, Setting=Base2026.04530————
LLaDA-InstructAttack=PAD, Defence=PPL, Setting=+SAD2026.0454.94————
LLaDA-1.5Attack=PAD, Defence=DiffuGuard, Setting=Base2026.04550————
LLaDA-InstructAttack=PAD, Defence=PPL, Setting=Base2026.0455.10————
CASTBackbone=QWEN-2.5-7B2026.0758.9—69.568.85740.2
LLaDA-InstructAttack=PAD, Defence=None, Setting=Base2026.04590————
LLaDA-InstructAttack=PAD, Defence=DiffuGuard, Setting=Base2026.04590————
BaselineBackbone=QWEN-2.5-7B2026.0763.2—74.872.863.741.7
Dream-InstructAttack=DIJA, Defence=DiffuGuard, Setting=+SAD2026.04710————
LLaDA-InstructAttack=DIJA, Defence=None, Setting=+SAD2026.04720————
LLaDA-InstructAttack=DIJA, Defence=DiffuGuard, Setting=+SAD2026.04730————
LLaDA-InstructAttack=DIJA, Defence=Self-rem., Setting=+SAD2026.04740————
LLaDA-1.5Attack=DIJA, Defence=Self-rem., Setting=+SAD2026.04740————
LLaDA-1.5Attack=DIJA, Defence=DiffuGuard, Setting=+SAD2026.04750————
LLaDA-InstructAttack=DIJA, Defence=Self-rem., Setting=Base2026.04761————
Dream-InstructAttack=DIJA, Defence=Self-rem., Setting=Base2026.04760————
LLaDA-1.5Attack=DIJA, Defence=None, Setting=+SAD2026.04770————
LLaDA-InstructAttack=DIJA, Defence=None, Setting=Base2026.04781————
LLaDA-InstructAttack=DIJA, Defence=DiffuGuard, Setting=Base2026.04781————
LLaDA-1.5Attack=DIJA, Defence=None, Setting=Base2026.04780————
LLaDA-1.5Attack=DIJA, Defence=Self-rem., Setting=Base2026.04780————
LLaDA-1.5Attack=DIJA, Defence=DiffuGuard, Setting=Base2026.04780————
Dream-InstructAttack=DIJA, Defence=Self-rem., Setting=+SAD2026.04780————
Dream-InstructAttack=DIJA, Defence=None, Setting=+SAD2026.04790————
Dream-InstructAttack=DIJA, Defence=PPL, Setting=Base2026.04800————
Dream-InstructAttack=DIJA, Defence=DiffuGuard, Setting=Base2026.04810————
Dream-InstructAttack=DIJA, Defence=None, Setting=Base2026.04840————
Dream-InstructAttack=DIJA, Defence=PPL, Setting=+SAD2026.04840————