Extracting books from production language models
从生产语言模型中提取书籍
机构 * Stanford University(斯坦福大学) ; Yale University(耶鲁大学)
AI总结 研究通过两阶段方法测试从生产LLM中提取书籍的可行性,发现部分模型可提取受版权保护文本,但需不同劫持策略,揭示生产LLM存在训练数据泄露风险。
Comments We ran experiments from mid-August to mid-September 2025, notified affected providers shortly after, and now make our findings public after a 90-day disclosure window