新闻
Linguistic Loopholes in LLM Unlearning: From a 174-Language Benchmark to Coverage-Aware Unlearning
arXiv cs.AI · 发布于 · 阅读约3分钟
30秒读懂
- 发生了什么
- Researchers discovered that unlearning facts in one language doesn't remove them in others, and proposed COVER, a method to select which languages to target for efficient multilingual unlearning.
- 为何重要
- Matters for engineers deploying LLMs where data removal compliance spans multiple languages and computational budgets limit full multilingual retraining.
- 注意
- COVER requires calibration data and frozen model access at deployment; gains measured on synthetic benchmarks and limited real-world low-resource language tests.
- llm
- rag
- edge
- benchmark
这条新闻背后的模式
- Budget-Guarded Autonomy
- MAPS: Multilingual Agent Performance & Security
- Agentic Context Engineering (Evolving Playbook)
每个模式都讲清楚技术如何运作、何时值得投入，以及在哪里会失效。
The Agent Architect
每周一个模式、一个权衡、一个生产事故案例。为构建智能体系统的人准备的每周简报。
每周一封邮件,一键退订。您的地址仅用于发送简报。