Kuiz Penskalaan Tetingkap Konteks YaRN

Jawapan dan penjelasan muncul semasa anda mengambil kuiz.

1. Apakah skim pengekodan kedudukan yang YaRN ubah suai untuk memanjangkan panjang konteks?
   • Pembenaman Kedudukan Putar (RoPE)
   • Embeddings sinusoidal mutlak
   • Jadual carian kedudukan yang dipelajari
   • Bias linear ALiBi
2. Bagaimanakah YaRN mengendalikan dimensi RoPE frekuensi tinggi berbanding frekuensi rendah?
   • Ia menginterpolasi kedua-duanya sama rata
   • Ia mengalih keluar dimensi frekuensi tinggi sepenuhnya
   • Ia hanya mengubah suai dimensi frekuensi tinggi
   • Ia meninggalkan dimensi frekuensi tinggi kebanyakannya utuh dan menginterpolasi dimensi frekuensi rendah
3. How should YaRN's reported data and fine-tuning efficiency be interpreted?
   • As proof that no YaRN model ever needs training
   • As results for its evaluated model-extension setups, rather than a fixed guarantee for every model
   • As a universal promise that every dataset size works equally well
   • As removal of the need for GPUs or inference computation
4. Selain menskala semula frekuensi, apakah pelarasan tambahan yang digunakan YaRN untuk memastikan perhatian jarak jauh stabil?
   • Peningkatan keciciran
   • Perbendaharaan kata yang lebih besar
   • Pelarasan suhu perhatian (skala softmax).
   • Keratan kecerunan
5. What motivates evaluating a RoPE model before using a context far beyond its original training length?
   • The model necessarily becomes perfectly accurate at longer lengths
   • The context extension automatically reduces memory to zero
   • A larger window proves every passage is equally easy to retrieve
   • Positional and attention behavior may degrade outside the conditions it learned
6. Interpolasi Kedudukan linear naif yang menjejaskan keupayaan yang YaRN bertujuan untuk mengekalkan?
   • Hubungan tempatan yang baik antara token berdekatan
   • Saiz perbendaharaan kata model
   • Kelajuan tokenizer
   • Bilangan lapisan
7. What happens in the transition region between YaRN's high- and low-frequency treatments?
   • Every dimension is assigned a random vocabulary token
   • All transition dimensions are deleted from attention
   • The scheme blends original and interpolated frequency treatment across the relevant dimensions
   • One identical scaling rule ignores frequency differences entirely
8. What does a checkpoint's advertised longer YaRN context limit fail to establish by itself?
   • Whether the positional scheme can accept any coordinates at all
   • Reliable performance for every intended task and position within that window
   • The spelling of the model's name
   • A requirement that every inference request retrain all weights