[2402.04249] HarmBench: A Standardized Evaluation Framework for Automated Red Teaming and Robust Refusal

arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website. Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them. Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs. arXiv Operational Status

[2402.04249] HarmBench: A Standardized Evaluation Framework for Automated Red Teaming and Robust Refusal Skip to main content arXiv is now an independent nonprofit! Learn more × Search arXiv Press Enter to search · Advanced search --> Computer Science > Machine Learning arXiv:2402.04249 (cs) [Submitted on 6 Feb 2024 ( v1 ), last revised 27 Feb 2024 (this version, v2)] Title: HarmBench: A Standardized Evaluation Framework for Automated Red Teaming and Robust Refusal Authors: Mantas Mazeika , Long Phan , Xuwang Yin , Andy Zou , Zifan Wang , Norman Mu , Elham Sakhaee , Nathaniel Li ,

related reading

• Your Evals Will Break and You Won't See It Coming - Lun Wangwanglun1996.github.io
• [2312.06942] AI Control: Improving Safety Despite Intentional Subversionarxiv.org
• [2209.07858] Red Teaming Language Models to Reduce Harms: Methods, Scaling Behaviors, and Lessons Learnedarxiv.org
• Defining LLM Red Teaming | NVIDIA Technical Blogdeveloper.nvidia.com
• 2312.06942arxiv.org
• Lapis Labslapis.rocks
• Challenges in evaluating AI systems \ Anthropicanthropic.com
• The bitter lesson of LLM evalsparsed.com
• Mediumai-alignment.com
• LLM evaluation: a beginner's guideevidentlyai.com
• GitHub - METR/RE-Bench · GitHubgithub.com
• Refusal in LLMs is mediated by a single direction — LessWronglesswrong.com