Awesome RL with Human Feedback / 2024 / arXiv

Emulated Disalignment: Safety Alignment for Large Language Models May Backfire!

作者：Zhanhui Zhou, Jie Liu, Zhichen Dong, Jiaheng Liu, Chao Yang, Wanli Ouyang, Yu Qiao · 单位：Shanghai Artificial Intelligence Laboratory · 来源：arXiv · 发布日期：2024

安全对齐对齐目标设计策略优化策略梯度人类反馈强化学习
新标签打开 AlphaXiv新标签打开 Code
Emulated Disalignment: Safety Alignment for Large Language Models May Backfire! figure
AlphaXiv 中文概览（可滚动查看）