Skip to content

Translation in progress Some entries still include text in the original language.

Oct 13 · Tuesday

All papers

Found 8 papers
Filter4

Attack type

Poisoning and backdoors40Model tampering14Extraction and theft8Detection evasion4

Defense type

Model hardening3

Impact

Data and privacy leaks7

System type

Models and APIs8

Attack surface / entry point

User input39Code repositories and dependencies1Images, audio and video4Training data7Weights and gradients11Inference services and hardware8

Layer

Training and data layer6
Select any value within a category; different categories must all match. Counts include the other filters you have selected.Another 227papers have no tags yet and are not included in filtered results.
Attack type：Extraction and theftSystem type：Models and APIsAttack surface / entry point：Code repositories and dependenciesAttack surface / entry point：Training dataClear all

Another 227papers do not have category tags yet and are not included in filtered results.

  1. DefensesarXiv 2610.06636

    差分隐私混合公共数据集提升隐私学习效果

    提出 DP-MixMin，私有学习公开数据混合并做差分隐私微调

    arXiv
    2610.06636
    Published
    Layer
    Training and data layer
    Setting
    Models and APIs
    Evaluated models
    ResNet-18, GPT-2 Medium
    AbstractDP-MixMin 私下学公开混合再 DP 微调，胸片 macro AUC 与邮件困惑度优于均匀混合。

    DP-MixMin 是一条端到端的差分隐私训练流程：先私下决定多个公开数据集的预训练混合，再预训练并在敏感数据上 DP 微调。作者称这是首个这样的流程。

    Paper analysisRead analysis
  2. DefensesarXiv 2610.04426

    UnAct：无需梯度的定向激活干预实现类别遗忘

    提出 UnAct，仅以前向激活干预从分类器中遗忘指定类别

    arXiv
    2610.04426
    Published
    Layer
    Model layer
    Setting
    Models and APIs
    Evaluated models
    ResNet-18, ViT-B/16
    AbstractUnAct 用前向激活做类别遗忘。

    UnAct 是一种事后类别遗忘：删除请求可以只有少量该类图像，且没有标签和保留数据。它不用梯度，只对遗忘图像做前向。编辑对象是分类器所读的后期单元。

    Paper analysisRead analysis
  3. AttacksarXiv 2610.01365

    ReGap：无需真实隐私数据即可通过微调重新提取模型隐私关联

    提出 ReGap，用自生成伪监督微调恢复模型中的隐私关联

    arXiv
    2610.01365
    Published
    Layer
    Training and data layer
    Setting
    Models and APIs
    Attacker
    White-box, Gray-box
    Evaluated models
    GPT-2 Small, GPT-2 Medium, GPT-2 Large and others6

    Abstract提出无真实私有监督的恢复攻击 ReGap，实验显示微调可通过自生成数据唤醒模型潜在隐私关联，突显后适应隐私风险。

    Paper analysisRead analysis
  4. DefensesarXiv 2609.37858

    存储不等于策略：面向 LLM 遗忘的状态条件支持控制

    提出 Intervention Score 与 DIR-R，按干预效果选择并修订遗忘参数支持

    arXiv
    2609.37858
    Published
    Layer
    Training and data layer
    Setting
    Models and APIs
    Evaluated models
    Llama 3 8B, Gemma 2 2B, Qwen2.5-7B and others4
    Abstract作者把局部化遗忘拆成初始干预选择和检查点修订，并在三个设置里分别给证据和边界。

    Storage Is Not Strategy 研究局部化 LLM 遗忘里两个常被混在一起的决定：一开始该更新哪些参数，以及优化改变模型之后这个选择要不要改。

    Paper analysisRead analysis
  6. AttacksarXiv 2609.33985

    研究：众包微调数据投毒可放大专有指令抽取

    提出主题锚点投毒，放大众包微调后的专有指令提取

    arXiv
    2609.33985
    Published
    Layer
    Training and data layer
    Setting
    Models and APIs
    Attacker
    Black-box
    Evaluated models
    Qwen2.5-7B-Instruct, Qwen2.5-14B-Instruct, Llama-3.1-8B-Instruct and others4

    Abstract论文提出基于主题锚点的数据中毒攻击，证明低比例的中毒样本可在黑盒条件下隐蔽放大微调指令泄漏。

    Paper analysisRead analysis
  7. DefensesarXiv 2610.00209

    基于差分隐私扩散模型的能源时间序列插补：裁剪感知的目标条件化方法

    提出时间步感知目标条件化，改善差分隐私能源时序插补效用

    arXiv
    2610.00209
    Published
    Layer
    Training and data layer
    Setting
    Models and APIs
    Evaluated models
    DiT-style Transformer
    Abstract固定阈值 DP-SGD 下，v-prediction 加 ᾱt 加权降低能源时序插补误差，并降低后期低 SNR 的裁剪前梯度。

    作者研究固定阈值 DP-SGD 训练的扩散模型，如何在余弦调度的后期低 SNR 时间步避免 ε-prediction 产生易被裁剪的大梯度，从而做能源时序缺失值恢复。

    Paper analysisRead analysis
  8. AttacksarXiv 2608.09867

    Researchers steal proprietary LLM inference chains via cross-model compatibility of encrypted inference blocks

    利用加密思维块的跨模型重放，诱导弱模型逐字转写专有推理链

    arXiv
    2608.09867
    Published
    Layer
    Prompt layer
    Setting
    Models and APIs
    Attacker
    Black-box
    Evaluated models
    Claude Opus 4.8, Claude Sonnet 4.6, Claude Haiku 4.5 and others13

    Abstract论文揭示了客户端加密思维块跨模型互换漏洞，提出了低成本提取方法并给出防御方案。

    Paper analysisRead analysis
You have reached the end