Whether you're a product manager, developer, or someone new to AI, this guide simplifies the process of understanding and adjusting LLM parameters.

You will be able to optimize these settings to suit your specific needs, enhancing your model’s performance and making it easier to achieve your goals.

But first, what are LLM parameters and how should you use them?

LLM parameters are like adjustable dials that control how the model responds. For example, the temperature setting adjusts how creative the model’s responses are, while function calling allows the model to tap into external tools to give more complete answers.

There are over 16 different parameters that let you fine-tune things like creativity, moderation, use of external tools, response length, and more.

These controls become really useful as you start building more complex workflows and when you need the model to perform in specific ways.

So many people focus on controlling model's outputs by writing better prompts, but in some cases, adjusting the parameter settings can be more effective. The tricky part is that learning how to use these parameters is somewhat of a black box.

That's where this guide comes in - we cover everything you need to know about each parameter, which models support them, how to use them, and when they’re most helpful.

Navigating this guide is easy. Just start by selecting topics from the left-hand panel that interest you. From basic LLM parameters like max_tokens to more advanced settings like Structured Outputs and Logit Bias, read how to enable, work and adapt all these parameters for your use-cases.
‍
If you’re new to LLMs, don’t worry—the guide is designed to meet you where you are. For each parameter you will learn how it works, when to apply it and how to experiment with it.

We cover the basics, plus the advanced settings - you can choose where you want to start.