Ikibazo gihuza ibibazo · Hagati Urwego

Gukuramo Icyitegererezo no Kwiba Ibitero Ikibazo

Sobanukirwa uburyo abateye bigana moderi yumukara-agasanduku binyuze mubibazo, ibyo API ibisubizo byerekana, nuburyo abatanga amakuru bashobora kumenya cyangwa gukumira gukuramo.

Inzira zijyanye no kuyoboraGukuramo Icyitegererezo no Kwiba Ibitero
Ikibazo 1 Bya 8

Ni ikihe gitekerezo cyibanze inyuma yigitero cyo gukuramo icyitegererezo?

Reba buri kibazo8 ibibazo

Ibisubizo nibisobanuro bigaragara mugihe ufata ikibazo.

  1. Ni ikihe gitekerezo cyibanze inyuma yigitero cyo gukuramo icyitegererezo?
    • Kubaza icyitegererezo no guhugura umusimbura wigana ibisubizo byayo
    • Kwinjira muri seriveri kugirango wandukure dosiye yuburemere
    • Guha umukozi kumena dataset y'amahugurwa
    • Kurenza API kugeza igihe iguye ikajugunya kwibuka
  2. Ni ukubera iki gusubiza ibyashoboka byuzuye byoroha gukuramo kuruta gusubiza label-1 gusa?
    • Ihishe igisubizo, itinda igitero hasi
    • Isohora amakuru menshi kubibazo, bityo ibibazo bike birakenewe
    • Ihisha imipaka yicyemezo rwose
    • Yongera igiciro cya seriveri kugirango abateye batange
  3. Ni ubuhe buryo bwo kwirwanaho bugabanya mu buryo butaziguye amakuru yatangajwe kuri buri kibazo?
    • Ongeraho GPU nyinshi kuri seriveri
    • Kongera icyitegererezo cyibara ryibara
    • Garuka kuzunguruka cyangwa hejuru-1-gusa ibisubizo
    • Gutangaza amakuru yamahugurwa kumugaragaro
  4. 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
  5. Kuki moderi isimbuwe yibwe iteje akaga kuruta kwinjiza gusa?
    • Ifasha kumurongo wera-agasanduku kwibasirwa nkubukorikori bwingero
    • Itezimbere icyitegererezo cyumwimerere
    • Irahita isiba icyitegererezo cyumwimerere
    • Birasabwa gusubiramo umwimerere
  6. 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
  7. 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
  8. 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