List your company

Value Alignment on WildGuard (test)

29.69ScoreNo Guidance
1.57888.876916.17523.4731May 12, 2026

Evaluation results

MethodLinks
No GuidanceBackbone=LLAMA-3.2-3B2026.0529.69—
No GuidanceBackbone=QWEN2-1.5B2026.0526.76—
No GuidanceBackbone=GPT-22026.0524.43—
No GuidanceBackbone=MISTRAL-7B2026.0523.81—
System PromptBackbone=GPT-22026.0520.8214.8
ITIBackbone=QWEN2-1.5B2026.0515.2343.1
System PromptBackbone=LLAMA-3.2-3B2026.0513.7353.8
ITIBackbone=GPT-22026.0513.6744
ITIBackbone=LLAMA-3.2-3B2026.0512.9756.3
RE-ControlBackbone=GPT-22026.0512.9347.1
System PromptBackbone=MISTRAL-7B2026.0512.9345.7
System PromptBackbone=QWEN2-1.5B2026.0512.852.2
RE-ControlBackbone=LLAMA-3.2-3B2026.0512.2258.8
RE-ControlBackbone=QWEN2-1.5B2026.0510.1262.2
ITIBackbone=MISTRAL-7B2026.059.8758.5
RE-ControlBackbone=MISTRAL-7B2026.059.5759.8
DPO (LoRA)Backbone=MISTRAL-7B2026.059.1561.6
DPO (LoRA)Backbone=QWEN2-1.5B2026.058.7167.5
DPO (LoRA)Backbone=GPT-22026.058.4465.5
DPO (LoRA)Backbone=LLAMA-3.2-3B2026.058.2872.1
SVGTBackbone=LLAMA-3.2-3B2026.057.8473.6
SVGTBackbone=MISTRAL-7B2026.056.5772.4
SVGTBackbone=QWEN2-1.5B2026.056.1876.9
SVGTBackbone=GPT-22026.052.6689.1