引導式測驗 · 中等 等級
模型提取和竊取攻擊測驗
了解攻擊者如何透過查詢模仿黑盒子模型、API 輸出揭示的內容以及提供者如何偵測或阻止提取。
問題 1 的 8
模型提取攻擊背後的核心思想是什麼?
預覽每個問題8 問題
當您參加測驗時,就會出現答案和解釋。
- 模型提取攻擊背後的核心思想是什麼?
- 查詢模型並訓練替代品來模仿其輸出
- 侵入伺服器直接複製權重文件
- 賄賂員工洩漏訓練資料集
- 重載 API 直到崩潰並轉儲內存
- 為什麼回傳完整機率向量比只回傳 top-1 標籤更容易被提取?
- 它對回應進行加密,減慢攻擊者的速度
- 每個查詢會洩漏更多信息,因此需要更少的查詢
- 它完全隱藏了決策邊界
- 它增加了伺服器成本,因此攻擊者放棄
- 哪種防禦技術可以直接減少每次查詢外洩的資訊?
- 為伺服器添加更多 GPU
- 增加模型的參數數量
- 返回舍入或僅 top-1 輸出
- 公開發布訓練數據
- What did the 2024 production-language-model extraction paper report about GPT-3.5 Turbo?
- It reconstructs every GPT-3.5 parameter exactly for under $20
- It obtains the model’s training dataset directly
- It extracts a complete model checkpoint from the API
- It recovers the embedding-projection layer; the paper estimates under $2,000 in queries for GPT-3.5 Turbo
- 為什麼被盜的替代模型除了收入損失之外還有其他危險?
- 它支援離線白盒攻擊,例如製作對抗性範例
- 提高了原始模型的準確性
- 它會自動刪除原始模型
- 需要重新訓練原來的
- In ActiveThief, what does the combined DFAL + K-center query strategy aim to balance?
- Boundary-proximate informative samples and reduced redundancy among selected samples
- Exact recovery of all secret model weights and a guaranteed substitute
- Random sampling only and avoiding all substitute-model predictions
- Higher API throughput and removal of the query budget
- What can API rate limits do against a model-extraction attempt?
- Prove that no substitute model can be trained
- Remove all information from each permitted prediction
- Make large-scale querying slower or more expensive
- Make model outputs impossible to compare
- What does the DynaMarks paper propose for a model-extraction defense?
- It adds a logo to every generated answer
- It encrypts all model weights at rest
- It guarantees ownership of every copied production model
- It can embed a signature in a surrogate model by dynamically changing API outputs using secret parameters