Nucleus and Top-k Sampling Quiz

Answers and explanations appear as you take the quiz.

1. How does top-k sampling restrict token choices?
   • It keeps only the k highest-probability tokens, then samples among them
   • It keeps tokens until cumulative probability exceeds k
   • It always selects the single most likely token
   • It removes the k least likely tokens only
2. What defines the 'nucleus' in nucleus (top-p) sampling?
   • The k most frequent words in training data
   • The smallest set of tokens whose probabilities sum to at least p
   • The single highest-probability token
   • All tokens with probability above 50%
3. What is the key advantage of top-p over top-k?
   • It is always faster to compute
   • It never produces random output
   • The number of candidate tokens adapts to the model's confidence
   • It ignores token probabilities
4. What does the temperature parameter do before sampling?
   • It sets the maximum sequence length
   • It selects the language of output
   • It counts the vocabulary size
   • It rescales the probability distribution to be sharper or flatter
5. What result did the nucleus-sampling paper report for its tested open-ended generation tasks?
   • It improved diversity without sacrificing fluency and coherence in the reported comparisons
   • It guaranteed factual correctness in every output
   • It eliminated all sampling randomness
   • It improved translation quality on every benchmark
6. For a greedy decoding baseline with no token sampling, how should generation be configured?
   • Disable sampling and use greedy decoding to select the highest-probability token at each step
   • Use high temperature and a large nucleus threshold
   • Randomize the temperature at every step
   • Keep sampling enabled but increase top-p
7. After top-k truncation removes tokens, what happens to the remaining sampling probabilities?
   • They remain assigned their original total mass below one
   • They are replaced with equal probabilities
   • They are renormalized over the retained top-k tokens
   • They are converted into beam-search scores
8. What did Holtzman et al. report about likelihood-maximizing decoding for open-ended text?
   • It always yields the most diverse human-like passages
   • It eliminates the need for a language model
   • It guarantees factual correctness
   • It can produce bland and repetitive text despite strong likelihood scores