Top-K Sampling

Text generation technique sampling from only the K most likely next tokens.

Restricts sampling to top K tokens by probability. Reduces unlikely tokens while maintaining diversity. Common: K=40-50.