HarmBench: A Standardized Evaluation Framework for Automated Red Teaming and Robust Refusal
Paper
•
2402.04249
•
Published
•
4
Lipeng (Tony) He
ttttonyhe
·
AI & ML interests
Trustworthy Machine Learning
Recent Activity
commented on
a paper
15 days ago
Activation Approximations Can Incur Safety Vulnerabilities Even in
Aligned LLMs: Comprehensive Analysis and Defense
liked
a dataset
27 days ago
PKU-Alignment/BeaverTails
liked
a model
2 months ago
cais/HarmBench-Llama-2-13b-cls
Organizations
None yet