Suhu dan Persampelan Kuiz

Jawapan dan penjelasan muncul semasa anda mengambil kuiz.

1. Apakah kesan menaikkan suhu kepada output model?
   • Menjadikannya lebih deterministik dan berulang
   • Mengubah berat model secara kekal
   • Melumpuhkan fungsi softmax
   • Menjadikannya lebih rawak dan pelbagai, dengan lebih banyak peluang token luar biasa
2. Sebelum pensampelan, skor token mentah model (logit) ditukar kepada kebarangkalian oleh:
   • Penapis top-k
   • Fungsi softmax
   • Penyebaran balik
   • Tokenizer
3. What determines the candidate set in top-p sampling, compared with top-k?
   • Top-p always keeps exactly ten tokens
   • Both methods always keep every vocabulary token
   • Top-p uses the smallest leading set reaching a cumulative probability threshold; top-k uses a count
   • Top-k uses a cumulative threshold while top-p uses a fixed count
4. Secara mekanikal, apakah sebenarnya yang dilakukan oleh suhu kepada logit?
   • Bahagikan setiap logit dengan T sebelum softmax
   • Menambah nombor rawak pada setiap logit
   • Memadam logit terendah
   • Mendarab kebarangkalian akhir dengan T
5. For a structured-data task, what does choosing more predictable token selection fail to establish?
   • That a decoding setting was chosen
   • That the generated content is correct or valid
   • That token scores can be compared
   • That a model can produce an output
6. How does greedy decoding choose the next token?
   • It literally divides every logit by zero
   • It uniformly samples every vocabulary token
   • It permanently retrains the model
   • It selects a highest-scoring eligible token rather than sampling
7. Two token logits differ by six before temperature scaling. With positive temperature two, what is their scaled difference?
   • Tiga
   • Dua belas
   • Enam
   • Sifar
8. For probabilities 0.55, 0.30, 0.10, and 0.05, how many leading tokens does top-p 0.80 keep?
   • satu
   • Tiga
   • dua
   • Empat