Research Papers 论文研究 5h ago Updated 54m ago 更新于 54分钟前 47

Forgotten in Weights, Recovered by Tools: Agentic Tool Unlearning for LLM Agents 遗忘于权重,恢复于工具:LLM智能体的代理工具非学习

LLM unlearning methods that only suppress parametric recall are insufficient for tool-augmented agents, which can recover forgotten knowledge through external tools like web search and retrieval The authors identify a new failure mode called "tool-mediated recovery" where unlearned information remains accessible via tool calls despite being removed from model weights Agentic Tool Unlearning (ATU) is a two-stage framework: parametric unlearning followed by trajectory-level reinforcement learning 提出LLM agent场景下的新失败模式"tool-mediated recovery":即使参数层面的知识被遗忘,agent仍可通过工具调用(如网络搜索、检索、数据库查询)恢复被遗忘内容 提出Agentic Tool Unlearning (ATU)两阶段框架,同时抑制参数化回忆和工具中介恢复,并保留对保留知识的正常工具使用能力 第一阶段应用参数知识遗忘技术,第二阶段在模拟tool-augmented环境中进行轨迹级强化学习,惩罚目标搜索行为和最终答案泄露 在RWKU和MUSE基准上的实验表明,ATU在不同LLM架构下实现了目标遗忘与保留效用之间的更优平衡

62
Hot 热度
72
Quality 质量
68
Impact 影响力

Analysis 深度分析

TL;DR

  • LLM unlearning methods that only suppress parametric recall are insufficient for tool-augmented agents, which can recover forgotten knowledge through external tools like web search and retrieval
  • The authors identify a new failure mode called "tool-mediated recovery" where unlearned information remains accessible via tool calls despite being removed from model weights
  • Agentic Tool Unlearning (ATU) is a two-stage framework: parametric unlearning followed by trajectory-level reinforcement learning in simulated tool-augmented environments
  • ATU penalizes both target-seeking tool behavior and final-answer leakage while preserving normal tool use for retained knowledge
  • Experiments on RWKU and MUSE benchmarks demonstrate improved balance between target forgetting and retained utility across different LLM architectures

Why It Matters

This research addresses a critical gap in LLM unlearning as agents increasingly rely on external tools, making traditional unlearning methods insufficient for real-world deployments. For AI practitioners building tool-augmented systems, it highlights the need to evaluate unlearning beyond parametric recall and consider how agents can circumvent unlearning through tool-mediated recovery. The findings have significant implications for compliance, privacy, and safety in production LLM agent systems.

Technical Details

  • Tool-mediated recovery: The paper identifies that previous unlearning methods suppress direct parametric recall but fail to prevent agents from recovering forgotten targets through external tools such as web search, retrieval systems, or database lookups
  • Two-stage ATU framework: Stage one applies parametric knowledge unlearning to suppress direct recall from model weights; stage two uses trajectory-level reinforcement learning in simulated tool-augmented environments to penalize target-seeking tool behavior and final-answer leakage
  • Benchmarks: Evaluated on RWKU and MUSE datasets across different LLM architectures
  • Key innovation: The framework distinguishes between forgetting the target information and preserving normal tool-use capabilities for retained knowledge, achieving better balance than prior methods

Industry Insight

  • Organizations deploying unlearning in production agent systems must evaluate not just parametric recall but also tool-mediated recovery paths to ensure compliance with data removal requests
  • As LLM agents become standard in enterprise applications, unlearning pipelines should incorporate simulated tool-augmented environments during training to anticipate real-world circumvention strategies
  • The two-stage approach of combining parametric unlearning with trajectory-level RL offers a practical blueprint for building more robust and compliant AI agent systems

TL;DR

  • 提出LLM agent场景下的新失败模式"tool-mediated recovery":即使参数层面的知识被遗忘,agent仍可通过工具调用(如网络搜索、检索、数据库查询)恢复被遗忘内容
  • 提出Agentic Tool Unlearning (ATU)两阶段框架,同时抑制参数化回忆和工具中介恢复,并保留对保留知识的正常工具使用能力
  • 第一阶段应用参数知识遗忘技术,第二阶段在模拟tool-augmented环境中进行轨迹级强化学习,惩罚目标搜索行为和最终答案泄露
  • 在RWKU和MUSE基准上的实验表明,ATU在不同LLM架构下实现了目标遗忘与保留效用之间的更优平衡

为什么值得看

  • 揭示了LLM unlearning在agent部署场景下的关键安全漏洞,对确保AI系统合规性和隐私保护具有重要实践意义
  • 为agent系统的知识遗忘提供了更全面的评估框架,推动unlearning研究从纯参数层面扩展到工具交互层面

技术解析

  • ATU两阶段框架:第一阶段采用参数知识遗忘技术抑制直接参数化回忆;第二阶段在模拟tool-augmented环境中进行轨迹级强化学习,惩罚target-seeking工具行为和最终答案泄露
  • 识别"tool-mediated recovery"失败模式:当LLM作为tool-augmented agent部署时,即使参数层面的知识被遗忘,agent仍可通过web search、retrieval或database lookup等工具调用恢复被遗忘内容
  • 实验验证:在RWKU和MUSE基准测试上,跨不同LLM架构验证ATU效果,结果表明该方法在target forgetting和retained utility之间实现了更优平衡

行业启示

  • AI合规与隐私保护:随着LLM agent在金融、医疗等敏感领域部署,仅做参数层unlearning已不足以满足合规要求,需建立覆盖工具交互层面的完整遗忘机制
  • Agent安全评估框架升级:现有unlearning评估多关注参数回忆,需将工具调用路径纳入评估体系,建立更全面的agent遗忘基准测试
  • 模型部署策略调整:企业部署tool-augmented agent时需重新审视知识遗忘方案,考虑工具访问控制与模型unlearning的协同设计

Disclaimer: The above content is generated by AI and is for reference only. 免责声明:以上内容由 AI 生成,仅供参考。

LLM 大模型 Agent Agent Research 科学研究 Alignment 对齐 Evaluation 评测