Guide-yakabatana mibvunzo · Zvakaoma Nhanho

Kufurira Mabasa eKudzidzisa Data Attribution Mibvunzo

Inotarisa pesvedzero-inoshanda counterfactuals, Hessian approximations, zvinoreva zviratidzo zvekupesvedzera uye kupihwa kwedata mumhando dzemitauro.

Nzira dzekutungamira dzinoenderanaKufurira Mabasa eKudzidzisa Dhata Attribution
Mubvunzo 1 ye 8

Ndeupi mubvunzo wekunyepedzera unopesvedzera mabasa akaenzana?

Tarisa mubvunzo wese8 mibvunzo

Mhinduro uye tsananguro zvinoonekwa paunenge uchitora bvunzo.

  1. Ndeupi mubvunzo wekunyepedzera unopesvedzera mabasa akaenzana?
    • Kurasikirwa padanho rekuyedza kwaizochinja sei kana muenzaniso wekudzidzira wakabviswa kana kukwirisa-uremu
    • Yakarurama sei iyo modhi paavhareji
    • Inokurumidza sei iyo modhi inodzidzisa
    • Mangani ma parameter ane modhi
  2. Ndechipi chinhu chemasvomhu chinoita kuti pesvedzero chaiyoiyo ikodzerwe kusabatika kumamodheru makuru?
    • The softmax basa
    • The inverse Hessian yekurasikirwa
    • The kudzidza rate purogiramu
    • Saizi yebatch
  3. Ndiani akagadziridza pesvedzero mabasa kune yazvino kudzidza kwakadzama mune inozivikanwa 2017 bepa?
    • Goodfellow uye Bengio
    • Vaswani et al.
    • Koh naLiang
    • Hinton uye LeCun
  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. Ndeipi fungidziro yakashandiswa ne Anthropic kupesvedzera mabasa kumhando dzemitauro mikuru?
    • EK-FAC (Kronecker-factored curvature)
    • Dropout
    • Beam search
    • Mixup
  6. Ko Anthropic's scaled study yakawana chii pamusoro pemienzaniso ine simba?
    • Vanogara vari makopi chaiwo ezvakabuda
    • Vanowanzo kugovera maitiro asinganzwisisike kwete mazwi akafanana
    • Vanogoverwa zvisina tsarukano
    • Chekutanga chete muenzaniso wekudzidziswa unokosha
  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