Skip to content

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

Oct 12 · Monday

All papers

Found 3 papers
Filter5

Defense type

Detection and filtering4Instruction–data separation3Permissions and approval0

Impact

Unauthorized actions18Output manipulation3

Attack surface / entry point

Documents and messages2Memory writes1

Layer

Application layer3
Select any value within a category; different categories must all match. Counts include the other filters you have selected.Another 232papers have no tags yet and are not included in filtered results.
Defense type：Instruction–data separationDefense type：Permissions and approvalImpact：Output manipulationAttack surface / entry point：Documents and messagesAttack surface / entry point：Memory writesClear all

Another 232papers do not have category tags yet and are not included in filtered results.

  1. DefensesarXiv 2610.11634

    LTBD：用可学习信任边界分隔符防御提示注入

    提出 LTBD，用可学习分隔符隔离可信指令与不可信数据

    arXiv
    2610.11634
    Published
    Layer
    Application layer
    Setting
    LLM applications
    Evaluated models
    Llama3-8B-Instruct, Llama3.1-8B-Instruct, Falcon3-7B-Instruct and others4
    AbstractLTBD 只训练四个分隔符嵌入。

    LTBD 是一种不改基座权重的提示注入防御：只学习用来标出可信指令与不可信数据边界的四个 token 嵌入。

    Paper analysisRead analysis
You have reached the end