List your company

Red-teaming on HarmBench Claude-4-Sonnet (test)

71ASR@1DIALTREE
-1.2817.48536.2555.015Oct 2, 2025

Evaluation results

MethodLinks
DIALTREETarget Model=Claude-4-Sonnet2025.107196.5
X-TeamingTarget Model=Claude-4-Sonnet2025.109.532.5
SFTTarget Model=Claude-4-Sonnet2025.101.54