GUARD 方法通过引导式答案推理蒸馏实现大模型自然遗忘
GUARD: Natural Forgetting in Large Reasoning Models via Guided Answer-Reasoning Distillation
朋友,这是 Yan 实验室的新方法 GUARD,专门解决大模型忘记信息时可能泄露隐私的问题,比现有方法效果更好。
GUARD 是一种新方法,用于让大模型在忘记特定信息时,生成连贯且不泄露的推理链。它通过引导模型输出稳定的拒绝式答案来替代原始内容。
实验表明,该方法在 R-TOFU 和 STAR-1 基准上显著减少了不安全或隐私信息的泄露。
GUARD: Natural Forgetting in Large Reasoning Models via Guided Answer-Reasoning Distillation
Recent advances in large reasoning models (LRMs) have made machine unlearning more challenging, as protected facts or unsafe rationales may surface in intermediate chain-of-thought (CoT) traces before the final answer is produced. Existing unlearning objectives typically suppress the target content or redirect internal representations, but they never specify how the post-forgetting trajectory should continue, which can lead to hallucinated substitutes, malformed boundaries, or repetitive outputs. We argue that LRM unlearning should instead learn a natural forgetting trajectory: a coherent non-disclosing CoT followed by a stable refusal-style answer that replace the original disclosure. To this end, we propose Guided Answer-Reasoning Distillation (GUARD), which converts model-generated unsafe disclosures into safe-exit trajectories, aligns a frozen LRM via guidance tokens, and distills the guided behavior into model parameters.To address the lack of metrics for replacement quality beyond leakage, we further introduce Natural Forgetting Reasoning Score (NFRS), which captures structural stability, fluency, and unsupported substitutes in forgotten outputs. Extensive experiments on R-TOFU and a STAR-1-derived harmful-intent setting show that GUARD substantially reduces unsafe and privacy disclosures across two widely adopted distilled LRMs while preserving reasoning utility. Codes are available at https://github.com/zeyu-Yan/GUARD
本条情报归属 AITOP「论文」核心栏目,重点聚焦北美前沿实验室与硅谷科技巨头。在跨区域产业观察坐标中,该动态直接关联「AI 开发者工具链、Coding Agent 与工程化实践」研究主线,反映出当前技术演进与商业落地的最新风向。涉及核心实体与技术议题包括:GUARD、大模型、机器忘却、推理模型,读者可通过下方标签进入对应专题页查阅历史脉络与关联对比。
信源核验说明:本文原始线索采编自「arXiv cs.AI」(@Zeyu Yan, Guanghao Zhou, Minghui Qiu, Ming Gao, Cen Chen),经 AITOP 多维评分管线结构化提炼。