Kuis terkait panduan · Sedang Tingkat

Perplexity dan Kuis Metrik Bahasa

Tinjau kemungkinan token referensi, entropi silang, garis dasar yang seragam, tokenisasi yang sebanding, BLEU, batas kualitas, jendela konteks, dan model yang disamarkan.

Jalur panduan terkaitPerplexity Dan Metrik Bahasa
Pertanyaan 1 dari 8

Under the same autoregressive text-scoring setup, what does lower perplexity indicate?

Pratinjau setiap pertanyaan8 pertanyaan

Jawaban dan penjelasan muncul saat Anda mengikuti kuis.

  1. Under the same autoregressive text-scoring setup, what does lower perplexity indicate?
    • Lower average negative log-likelihood for the reference tokens
    • A guarantee of more factually correct generated answers
    • A guarantee of a larger parameter count
    • A requirement that every top-choice prediction be correct
  2. Using natural-log negative log-likelihood, how is autoregressive perplexity computed from its average?
    • By multiplying the parameter count by the context length
    • By subtracting the average from the number of attention heads
    • By taking the average learning rate
    • By exponentiating the average negative log-likelihood
  3. Sebuah model memberikan probabilitas seragam pada kosakata 10.000 kata tanpa pembelajaran. Apa kebingungannya?
    • 1
    • 10.000
    • 0
    • 100
  4. Why do tokenizer and corpus settings matter when directly comparing per-token perplexity values?
    • Perplexity ignores every scored reference token
    • All corpora necessarily produce the same likelihood
    • They change the units or reference data over which probabilities are averaged
    • Only model parameter count determines the score
  5. Which translation-reference metric uses modified n-gram precision and a brevity penalty?
    • Perplexity
    • Kesalahan piksel kuadrat rata-rata
    • Jumlah parameter model
    • BIRU
  6. Keterbatasan utama apa yang membuat peneliti tidak lagi kebingungan dalam menilai kualitas dunia nyata?
    • Perplexity mengukur prediksi kata berikutnya, bukan apakah keluarannya berguna, benar, atau bermanfaat
    • Perplexity terlalu mahal untuk dihitung
    • Perplexity tidak dapat dihitung pada data teks
    • Perplexity hanya berfungsi selama inferensi, bukan pelatihan
  7. When scoring a long text with a fixed-context model, what can a sliding-window procedure provide that disjoint chunks lose near boundaries?
    • A new vocabulary learned automatically at every boundary
    • A guarantee that generated answers are factually correct
    • More preceding context for the scored target tokens
    • A requirement to count each overlapping target repeatedly
  8. Why is conventional left-to-right perplexity not directly the same score for a BERT-style masked language model?
    • BERT has no learned parameters
    • The masked model does not directly define the same autoregressive sequence factorization
    • Every masked-token probability is necessarily zero
    • Perplexity is defined only for image pixels