# context
標記為「context」的 8 篇文章
程式碼庫上下文投毒
投毒 AI 程式助理所讀取的程式碼庫上下文,以操控生成建議。
code-gencontextpoisoningrepository
多檔案上下文攻擊
跨多個檔案建構並觸發的 AI 程式助理攻擊,以規避單檔偵測。
code-genmulti-filecontextmanipulation
Conversation Reset Attacks
利用 conversation resets與context clearing to weaken model adherence to safety instructions.
labsconversation-resetcontextbeginner
上下文 Window 投毒 實驗室
利用 context window management to inject persistent adversarial content that influences future model responses.
contextintermediatelabpoisoningwindowlabs
情境注入技術
打造能無縫融入預期對話脈絡、以規避內容分類器偵測的注入載荷。
prompt-injectioncontextsemanticstealth
AI 情境下的社會工程
因應 AI 中介互動與代理型系統而調整的社會工程技術。
tradecraftsocial-engineeringaicontext
Advanced Context Manipulation 詳解
Manipulate conversation context through history injection, context overflow, and attention weight exploitation.
walkthroughscontextmanipulationadvanced
上下文操控深度探討
Advanced walkthrough of context window manipulation techniques including attention dilution and instruction priority shifting.
diveattackscontextdeepmanipulationwalkthroughs