AI glossary
Temperature
Temperature is a setting that controls how much randomness a model uses when it picks the next token: low is steadier, high is more varied.
At each step a model gives every possible next token a probability. At a low temperature it almost always picks the most likely one; at a higher temperature less likely tokens get a real chance.
Use a low temperature to extract a date from an email or to classify text, where you want the same answer every time. Use a higher one to brainstorm names, where variety is the point.
Temperature zero makes answers more repeatable but not perfectly identical on every service, and it does not make them more correct. A wrong answer at temperature zero is simply wrong every time.
Learn it hands-on
These free lessons teach it with commands to run in a lab and checked tasks. Read them whole with no account.