Buck Shlegeris

发表论文 29 篇 · 总被引 2830 次 · h-index 18

代表论文

• Interpretability in the Wild: a Circuit for Indirect Object Identification in GPT-2 small (2022 · International Conference on Learning Representations · 被引 1275)
• Sleeper Agents: Training Deceptive LLMs that Persist Through Safety Training (2024 · arXiv.org · 被引 565)
• Alignment faking in large language models (2024 · arXiv.org · 被引 337)
• Chain of Thought Monitorability: A New and Fragile Opportunity for AI Safety (2025 · arXiv.org · 被引 225)
• AI Control: Improving Safety Despite Intentional Subversion (2023 · International Conference on Machine Learning · 被引 213)
• Sycophancy to Subterfuge: Investigating Reward-Tampering in Large Language Models (2024 · arXiv.org · 被引 160)