Kuiz berkaitan panduan · Keras Tahap

Autoenkoder Jarang untuk Kuiz Pengekstrakan Ciri

Kaji kamus dan pembinaan semula ciri yang jarang, dengan eksperimen Jambatan Golden Gate, pendam yang melampaui batas, stereng dan model.

Laluan panduan yang berkaitanAutoenkoder Jarang Untuk Pengekstrakan Ciri
soalan 1 daripada 8

Which interpretability difficulty motivates sparse dictionary learning for neural activations?

Pratonton setiap soalan8 soalan

Jawapan dan penjelasan muncul semasa anda mengambil kuiz.

  1. Which interpretability difficulty motivates sparse dictionary learning for neural activations?
    • Every model activation is already a complete dictionary definition
    • Several learned signals can overlap in an activation space
    • The tokenizer must be permanently removed
    • All neural weights are stored as image pixels
  2. What does an SAE sparsity constraint encourage for a given input?
    • Every feature to activate equally on every input
    • A mandatory removal of its decoder
    • A guarantee that every feature has exactly one perfect human label
    • Only a limited subset of learned feature activations to be nonzero
  3. Dalam kerja 'Scaling Monosemanticity' Anthropic, apakah ciri contoh yang terkenal?
    • Ciri Jambatan Golden Gate
    • Ciri untuk nombor tujuh
    • Ciri untuk warna biru sahaja
    • Ciri untuk vokal
  4. What does an overcomplete SAE dictionary mean?
    • It must activate every feature for every token
    • It contains no reconstruction mapping
    • It has more learned feature directions than input activation dimensions
    • It always requires less storage than the input vector
  5. What does a monosemantic feature label aim to describe in this research context?
    • A guarantee that the feature never responds to related examples
    • A feature associated with one interpretable concept
    • A language model that has only one parameter
    • A token represented by exactly one raw byte
  6. Bagaimanakah SAE boleh digunakan untuk mempengaruhi tingkah laku model?
    • Dengan mengapit ciri hidup atau mati untuk mengemudi output
    • Dengan melatih semula keseluruhan model dari awal
    • Dengan memadamkan tokenizer
    • Dengan menukarnya kepada model imej
  7. What does an SAE’s reconstruction residual measure for an activation vector?
    • The difference between two unrelated document titles
    • The total number of words in the model vocabulary
    • Only the model’s hardware temperature
    • The difference between the original activation and its decoded reconstruction
  8. When an SAE is fitted to collected activations of a frozen language model, which weights need to be trained for that fitting?
    • Every original language-model weight must be updated
    • Only the tokenizer’s character list
    • The SAE’s own encoder and decoder weights
    • No parameters of the SAE can change