ASRU：激活引导与强化遗忘融合用于多模态大语言模型
ASRU: Activation Steering Meets Reinforcement Unlearning for Multimodal Large Language Models
AI总结：
ASRU提出一种可控多模态遗忘框架，通过激活引导和强化学习提升多模态大语言模型的遗忘效果和生成质量，实验显示在Qwen3-VL上遗忘效果提升24.6%，生成质量提升5.8倍。
摘要：
多模态大语言模型（MLLMs）在预训练过程中可能记忆敏感的跨模态信息，使机器遗忘（MU）变得至关重要。现有方法通常基于输出偏差评估遗忘效果，而忽视遗忘后的生成质量。这可能导致幻觉或僵化响应，影响遗忘模型的可用性和安全性。为了解决这一问题，我们提出了ASRU，一种可控的多模态遗忘框架，将生成质量作为核心评估目标。ASRU首先通过激活引导诱导初始拒绝行为，然后使用定制奖励函数优化细粒度拒绝边界，从而在目标知识遗忘和模型实用性之间取得更好的平衡。实验表明，在Qwen3-VL上，ASRU在平均上显著提高了遗忘效果（+24.6%）和生成质量（5.8倍），同时有效保持了模型实用性，仅使用少量保留的监督数据。
英文摘要：
Multimodal large language models (MLLMs) may memorize sensitive cross-modal information during pretraining, making machine unlearning (MU) crucial. Existing methods typically evaluate unlearning effectiveness based on output deviations, while overlooking the generation quality after unlearning. This can easily lead to hallucinated or rigid responses, thereby affecting the usability and safety of the unlearned model. To address this issue, we propose ASRU, a controllable multimodal unlearning framework that incorporates generation quality as a core evaluation objective. ASRU first induces initial refusal behavior through activation redirection, and then optimizes fine-grained refusal boundaries using a customized reward function, thereby achieving a better trade-off between target knowledge unlearning and model utility. Experiments on Qwen3-VL show that ASRU significantly improves unlearning effectiveness (+24.6%) on average and generation quality (5.8X) on average while effectively preserving model utility, using only a small amount of retained supervision data.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
