Controls randomness. Lower values make the output more focused and deterministic.

Controls the maximum number of tokens to generate and in turn the response length.

Also known as nucleus sampling or 'penalty', this controls the diversity and quality of the responses. It limits the cumulative probability of the most likely tokens. Lower values lead to more constrained responses.

Controls the model's tendency to generate repetitive responses. Lower values make the model more likely to repeat information.

Controls the avoidance of specific topics in its responses. Lower values make the model less concerned about preventing those topics.

Change the seed to randomize responses if your getting the same output over and over. Does not guarantee deterministic outputs.