Sleeper Agents: Training Deceptive LLMs that Persist Through Safety Training
Paper
•
2401.05566
•
Published
•
30
safety
-
-
Weak-to-Strong Jailbreaking on Large Language Models
Paper • 2401.17256 • Published • 16 -
Robust Prompt Optimization for Defending Language Models Against Jailbreaking Attacks
Paper • 2401.17263 • Published • 1 -
Summon a Demon and Bind it: A Grounded Theory of LLM Red Teaming in the Wild
Paper • 2311.06237 • Published • 1 -
HarmBench: A Standardized Evaluation Framework for Automated Red Teaming and Robust Refusal
Paper • 2402.04249 • Published • 6 -
The Instruction Hierarchy: Training LLMs to Prioritize Privileged Instructions
Paper • 2404.13208 • Published • 40 -
Improving Alignment and Robustness with Short Circuiting
Paper • 2406.04313 • Published • 1