Kuis terkait panduan · Keras Tingkat

Autoencoder Jarang untuk Kuis Ekstraksi Fitur

Pelajari kamus dan rekonstruksi fitur yang jarang, dengan eksperimen Jembatan Golden Gate, laten yang terlalu lengkap, kemudi, dan batasan model.

Jalur panduan terkaitAutoencoder Jarang Untuk Ekstraksi Fitur
Pertanyaan 1 dari 8

Which interpretability difficulty motivates sparse dictionary learning for neural activations?

Pratinjau setiap pertanyaan8 pertanyaan

Jawaban dan penjelasan muncul saat Anda mengikuti kuis.

  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 karya 'Scaling Monosemantity' Anthropic, apa contoh fitur yang terkenal?
    • Fitur Jembatan Golden Gate
    • Sebuah fitur untuk nomor tujuh
    • Fitur untuk warna biru saja
    • Sebuah fitur 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. Bagaimana SAE dapat digunakan untuk mempengaruhi perilaku model?
    • Dengan mengaktifkan atau menonaktifkan fitur untuk mengarahkan keluaran
    • Dengan melatih ulang seluruh model dari awal
    • Dengan menghapus tokenizer
    • Dengan mengubahnya menjadi model gambar
  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