Kuis Penyetelan Instruksi

Jawaban dan penjelasan muncul saat Anda mengikuti kuis.

1. Tanpa penyetelan instruksi, bagaimana model dasar mentah dapat merespons pertanyaan 'Apa ibu kota Prancis?'
   • Ia selalu menolak untuk menjawab
   • Ini mengembalikan kesalahan
   • Ini mungkin berlanjut dengan lebih banyak pertanyaan atau teks terkait daripada menjawab
   • Itu menjawab dengan sempurna setiap saat
2. Penyetelan instruksi paling tepat digambarkan sebagai:
   • Pembelajaran penguatan dari umpan balik manusia
   • Pra-pelatihan tentang teks web mentah
   • Menyesuaikan parameter suhu
   • Penyesuaian yang diawasi pada pasangan instruksi dan respons
3. How does supervised instruction tuning differ from reward-model-based RLHF in the published InstructGPT recipe?
   • It is exactly the same optimization stage
   • It trains against reference responses, while the later preference stages use comparisons and a reward model
   • It requires no training examples
   • It only changes a sampling temperature
4. Why do instruction-tuning researchers evaluate models on held-out tasks and instructions?
   • To assess generalization beyond the examples used for training
   • To prove that every unseen task will succeed
   • To avoid checking any generated responses
   • To guarantee the same accuracy for every task mixture
5. Urutan tahapan pelatihan apa yang umum untuk asisten obrolan modern?
   • RLHF, lalu prapelatihan, lalu penyetelan instruksi
   • Penyetelan instruksi, lalu pra-pelatihan
   • Hanya pra-pelatihan, tidak ada yang lain
   • Pra-pelatihan, lalu penyetelan instruksi, lalu RLHF opsional
6. What does LIMA's curated-data experiment support as a research conclusion?
   • Exactly 1,000 examples are enough for every model and task
   • More data can never help any instruction-tuned model
   • Carefully selected small datasets can be effective in an evaluated configuration
   • Only synthetic examples can train an assistant
7. Which training stage did the original LIMA experiment omit while using its 1,000 curated prompt-response pairs?
   • Reinforcement learning and human preference modeling
   • Supervised response loss
   • A pretrained base model
   • Prompt-response examples
8. A tuned model follows the requested answer format. What does that behavior alone fail to establish?
   • That an output format was requested
   • That every factual statement in the answer is correct
   • That the model produced a response
   • That the response can be inspected