用可调控提示实现大模型精准删知识，不改模型也能随时恢复。
CAP: Controllable Alignment Prompting for Unlearning in LLMs
- 通过强化学习优化提示词，让模型主动抑制特定知识。
- 无需修改参数即可精准删除敏感信息，且可逆恢复。
- 适合无法访问权重的闭源模型，解决遗忘不可控问题。
大型语言模型在未过滤语料上训练，可能保留敏感信息，需进行选择性知识遗忘以满足合规与伦理要求。现有参数修改方法存在计算成本高、遗忘边界不可控、依赖模型权重访问等根本缺陷，难以应用于闭源模型；而当前非侵入式方法系统性不足，依赖经验。为此，我们提出可控对齐提示遗忘框架（CAP），一种端到端的提示驱动遗忘范式。CAP通过强化学习将遗忘过程解耦为可学习的提示优化，由提示生成器与大模型协同工作，在选择性保留通用能力的同时抑制目标知识。该方法支持通过撤销提示实现知识恢复。大量实验表明，CAP在不更新模型参数的情况下实现精确、可控的遗忘，建立动态对齐机制，突破了以往方法的迁移性局限。
原文摘要 · Abstract (English)
Large language models (LLMs) trained on unfiltered corpora inherently risk retaining sensitive information, necessitating selective knowledge unlearning for regulatory compliance and ethical safety. However, existing parameter-modifying methods face fundamental limitations: high computational costs, uncontrollable forgetting boundaries, and strict dependency on model weight access. These constraints render them impractical for closed-source models, yet current non-invasive alternatives remain unsystematic and reliant on empirical experience. To address these challenges, we propose the Controllable Alignment Prompting for Unlearning (CAP) framework, an end-to-end prompt-driven unlearning paradigm. CAP decouples unlearning into a learnable prompt optimization process via reinforcement learning, where a prompt generator collaborates with the LLM to suppress target knowledge while preserving general capabilities selectively. This approach enables reversible knowledge restoration through prompt revocation. Extensive experiments demonstrate that CAP achieves precise, controllable unlearning without updating model parameters, establishing a dynamic alignment mechanism that overcomes the transferability limitations of prior methods.
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。