Top-k and Top-p Sampling

Easy

llms OpenAI Anthropic Google

How do top-k and top-p (nucleus) sampling differ in controlling text generation?

Multiple Choice

Top-k picks from the k most probable tokens; top-p picks from the smallest set whose cumulative probability exceeds p Top-k is deterministic; top-p is stochastic Top-k works on logits; top-p works on embeddings Top-k is used during training; top-p is used during inference

Correct!

Incorrect — try again next time!

Upgrade to Pro to unlock solutions and explanations

Top-k and Top-p Sampling

Multiple Choice

Related problems

Deep-dive guides