提出新优化框架，让大模型更高效地删除记忆而不影响原有能力。
DualOptim+: Bridging Shared and Decoupled Optimizer States for Better Machine Unlearning in Large Language Models
- 用共享状态和差异状态分离遗忘与保留目标，动态调整优化方向。
- 在多种任务中实现遗忘与保留的更好平衡，性能优于现有方法。
- 支持8比特量化，降低内存开销，适合实际部署场景。
我们提出DualOptim+，一种用于提升大语言模型机器遗忘效果的新优化框架。该框架引入基础状态以捕捉遗忘与保留目标共有的表征，并通过增量状态保留目标特定的残差。该架构可根据遗忘与保留梯度之间的方向冲突，自适应地在共享与解耦状态间切换。此外，我们进一步提出DualOptim+ 8bit，一种量化版本，在不损失性能的前提下显著降低内存开销。在虚构数据与真实世界遗忘、安全对齐及多任务学习等任务上的大量实验表明，DualOptim+在不同目标之间始终实现更优的权衡。代码已开源：https://github.com/CityU-MLO/DualOptimPlus。
原文摘要 · Abstract (English)
We propose DualOptim+, a novel optimization framework for improving machine unlearning in large language models. It introduces a base state to capture common representations shared by forgetting and retaining objectives and delta states to preserve objective-specific residuals. This architecture allows the optimizer to adaptively bridge shared and decoupled states based on the directional conflict between forgetting and retaining gradients. We further introduce DualOptim+ 8bit, a quantized variant that reduces memory overhead without compromising performance. Extensive experiments across fictitious and real-world unlearning, safety alignment, and multi-task learning tasks demonstrate that DualOptim+ consistently achieves a superior trade-off between different objectives. Codes are available at https://github.com/CityU-MLO/DualOptimPlus.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。