It controls the number of tokens the 
model shall generate, and it is one of the many text generation parameters that you can learn about 
in Text generation strategies guide.