List your company

Benign Compliance on WildJailBreak (WJB) Vanilla Benign

100Compliance RateSELF-REDTEAM
98.33698.76899.299.632May 8, 2026

Evaluation results

MethodLinks
SELF-REDTEAMBackbone=Qwen2.5-3B-IT2026.05100
Qwen2.5-14B-ITBackbone=Qwen2.5-14B-IT2026.05100
ABSBackbone=Qwen2.5-14B-IT2026.05100
SELF-REDTEAMBackbone=Qwen2.5-14B-IT2026.0599.6
Qwen2.5-3B-ITBackbone=Qwen2.5-3B-IT2026.0599.2
Qwen2.5-7B-ITBackbone=Qwen2.5-7B-IT2026.0599.2
SELF-REDTEAMBackbone=Qwen2.5-7B-IT2026.0599.2
ABSBackbone=Qwen2.5-7B-IT2026.0598.8
ABSBackbone=Qwen2.5-3B-IT2026.0598.4