Controlling LLM Output: Temperature, Sampling, Penalties, and Stopping
中文摘要
本文介绍通过温度、采样和惩罚等参数控制大模型输出，以平衡生成的稳定性与创造性。
English Summary
This article explains controlling LLM outputs using temperature, sampling, and penalties to balance predictability and creativity for specific needs.
Original Excerpt
Large language models can generate different answers to the same prompt. Sometimes you want highly predictable output. Other times, you… Continue reading on Medium »