Rick W / Friday, August 7, 2026 / Categories: Artificial Intelligence Decoding Strategies and Output Control This chapter is divided into nine parts; they are: • Reading Logits from a Model • Greedy Decoding • Temperature Sampling • Top-$k$ Sampling • Nucleus Sampling • Repetition Penalties • Beam Search • Stop Conditions • Structured Output Constraints The model returns a vector of logits for every position in the input sequence. Previous Article Using a Transformer Model: From Training to Inference Next Article Measuring Performance of Transformer Inference Print 6 Tags: ModeModel