Our new X account is live! Follow @wizwand_team for updates
WorkDL logo mark

Red-teaming Safety Evaluation on Edgebench

4.53HS Score

Meta-Llama-3.1-8B (Unaligned)

2.09642.72823.363.9918May 30, 2025
Updated 4d ago

Evaluation Results

MethodLinks
2025.05
4.537585
2025.05
4.245769
2025.05
3.61-41
2025.05
3.593647
2025.05
3.323034
2025.05
3.23-29
2025.05
3.153329
2025.05
3.14-27
2025.05
3.08-23
2025.05
2.91-23
2025.05
2.852926
2025.05
2.81-23
2025.05
2.362318
2025.05
2.35-14
2025.05
2.322921
2025.05
2.19-12