# semantic
標記為「semantic」的 8 篇文章
基於語義相似度的防禦
利用嵌入相似度分析偵測語義上接近已知攻擊模式的對抗性輸入。
defensesemanticembeddings
語意注入研究
研究在語意上連貫、難以與正常輸入區分的注入。
injection-researchsemanticinjectionnatural-language
語意空間注入研究
研究在語意嵌入空間而非符元空間中運作的注入 — 直接利用已學到的表徵。
researchsemanticembeddinginjection
Semantic Deception Crafting 實驗室
Craft semantically coherent prompts that conceal adversarial intent from both human reviewers與ML classifiers.
craftingintermediatesemanticlabdeceptionlabs
Semantic 注入 Crafting
Craft semantically coherent injections that evade both classifiers與human review.
labssemanticinjectionintermediate
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
語意偽裝詳解
Walkthrough of crafting semantically camouflaged injections that evade both classifiers and human review.
attackscamouflagesemanticwalkthroughwalkthroughs
Semantic Deception 攻擊 詳解
Craft semantically deceptive prompts that appear benign to classifiers while achieving adversarial objectives.
walkthroughssemanticdeceptionattacks