Dan Hendrycks

机构：UC Berkeley

发表论文 74 篇 · 总被引 54071 次 · h-index 45

代表论文

• HarmBench: A Standardized Evaluation Framework for Automated Red Teaming and Robust Refusal (2024 · International Conference on Machine Learning · 被引 1401)
• Representation Engineering: A Top-Down Approach to AI Transparency (2023 · arXiv.org · 被引 1228)
• DecodingTrust: A Comprehensive Assessment of Trustworthiness in GPT Models (2023 · Neural Information Processing Systems · 被引 690)
• The WMDP Benchmark: Measuring and Reducing Malicious Use With Unlearning (2024 · International Conference on Machine Learning · 被引 501)
• OpenOOD: Benchmarking Generalized Out-of-Distribution Detection (2022 · Neural Information Processing Systems · 被引 391)
• AI deception: A survey of examples, risks, and potential solutions (2023 · Patterns · 被引 377)