TRANSFORMERS & LLMs • 71 / 397
Follow text from characters to tokens, attention, transformers and generation.
Temperature
Temperature adjusts how predictable versus varied token sampling is.
Think of it like
Think of Temperature as part of a very fast reader that repeatedly decides what information matters and what should come next.
Real life
Autocomplete, translation and chat assistants depend on concepts such as Temperature.
SRE lens
For incident extraction you usually prefer stable outputs; brainstorming may tolerate more variation.
Remember thisTemperature adjusts how predictable versus varied token sampling is.