Decoding Strategies and Output Control
中文摘要
模型输出决策策略控制，以生成更精确且结构化的文本。
English Summary
AI models use decoding strategies like sampling and beam search to control output for accuracy and structure.
原文节选
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.