quiz bu lëkkale ak tegtal · Dëgër Niveau

Fonction yu am njeexital ci quiz biy joxe done ci tàggat yaram

Dafay saytu li ñuy woowe influens-fonction kontrefaktuel, xayma Hessian, lu màndarga influens yi di tekki ak joxe done ci misaali làkk yi.

Yooni njiit yu jëm ciFonction yu am njeexital ngir joxe done ci tàggat yaram
Laaj 1 de 8

Ban laaj kontrefaktuel mooy fonction yi am ci njeexte?

Xoolaatal laaj bu nekk8 laaj

Sooy def quiz bi, tontu yi ak leeral yi dañuy feeñ.

  1. Ban laaj kontrefaktuel mooy fonction yi am ci njeexte?
    • naka la perte ci point test bi di soppikoo sudee dindi nañu misaalu tàggat yaram wala ñu yokk pondération bi
    • naka la model bi di doxee ci moyenne
    • naka la modelu saxaar yi di gaawe
    • Ñaata parametre la model bi am
  2. Ban mbir ci math mooy tax xayma yi gëna yombatul ci model yu mag yi?
    • Fonction softmax bi
    • Hessian bu dëddu bu ñàkk
    • Jadwalu tolluwaayu jàng
    • Dayo lote bi
  3. Kan moo méngale fonction influence ak njàng mu xóot ci jamono jii ci benn téere bu siiw ci 2017?
    • Goodfellow ak Bengio
    • Vaswani ak ñeneen.
    • Koh ak Liang
    • Hinton ak Lekun
  4. Under Koh and Liang’s standard I_up,loss convention, what does a negative value indicate?
    • The point was not in the training data
    • Upweighting the point is estimated to increase test loss
    • The point has no relationship to the test example
    • Upweighting the point is estimated to lower test loss, so it is helpful
  5. Ban xayma la Anthropic jëfandikoo ngir yamale fonction yi ci modeli làkk yu mag yi?
    • EK-FAC (courbure facteur de cronecker)
    • Dàddu
    • Seetug limyéer
    • jaxase
  6. Lan la Anthropic ci jàngat biñ def ci misaal yu am solo yi?
    • Dañuy nuru ay kopi yu gëna jub ci li ñuy génne
    • Dañuy faral di bokk motif yu abstrak, duñu bokk kàddu yu nuróo
    • Dañu leen séddale ci anam wu mucc
    • misaalu tàggat bi njëkk kese moo am solo
  7. In the influence-function approximation, what connects a training-example gradient to the effect on a test loss?
    • The test set’s token count
    • A random forest over training labels
    • The inverse Hessian of the training loss
    • The model’s output vocabulary size
  8. If removing one training example is estimated to increase a test point’s loss, what does that suggest about the example?
    • It was guaranteed to be a duplicate
    • It had no relationship to the model
    • It necessarily caused the model to overfit
    • It had a helpful influence on that test point under the approximation