The model settings

On the right side, you can instead find the model settings panel. This allows you to configure the model and tune it to receive better responses.

The settings change depending on the selected model, but generally, most text models will have some of the following settings:

• Temperature: determines the randomness of the generated completion, a lower value will be more deterministic, and therefore more predictable, and a higher value will generate more diverse responses. You can tune this value depending on your purpose: for code generation, a lower value will be more reliable, but for creative writing, a higher one might lead to more diverse results.
• Top P: known also as "nucleus sampling”, this will also contribute to generating more diverse answers. The main difference to the Temperature is that Top P affects the choice of words and style, rather than its content.
• Max Tokens: the max length of the response the model will produce. Useful if you want to control costs. (1 Token = approx 4 letters.)
• Frequency Penalty: This setting is used to discourage the model from repeating the same words or phrases too frequently within the generated text. A higher value will result in the model being more conservative in its use of repeated tokens.
• Presence Penalty: Similar to the previous one. This parameter is used to encourage the model to include a diverse range of tokens in the generated text. A higher value will result in the model being more likely to generate tokens that have not yet been included in the generated text.