The table below provides list some of the key terms that you may encounter when interacting with the chatbot and their definitions.
| AI Key Terms | |
|---|---|
| Admin | admin refers to one of the roles that a user can be granted. The admin is able to adjust settings such as the temperature of the model. |
| Generative AI | refers to a type of application, such as ChatGPT, designed to use a variety of machine learning algorithms to create new content such as text, music, images that mimics human creation. |
| Large Language Model (LLM) | a LLM is a sophisticated computer program that understands and generates human-like text. Its size, determined by parameters, enables it to grasp complex language patterns, making it adept at tasks like conversation, writing, and problem-solving. |
| Model Behavior | the model behavior refers to the settings that allow you to control the temperature of the model and instruct the chatbot on how to behave. Whether you can see this section on the website depends on your level of access. |
| Model Configuration | the model configuration section provides additional information about the number of tokens, the API key, the endpoint, deployment ID and API version. Whether you can see this section on the website depends on your level of access. |
| Prompts | prompts are instructions that can include questions or commands given by the user to the chatbot for the chatbot to generate an answer. |
| Prompt-a-thon | a prompt-a-thon is an event or activity where participants generate and share prompts, typically for creative or writing purposes. Participants use these prompts to inspire and guide their creative work. |
| Super-admin | super-admin refers to one of the roles that a user can be granted. This role provides the maximum level of access to the chatbot. The super-admin is able to perform the functionalities listed in the Model Settings section. |
| System prompt | system prompt refers to the predefined text or message that is designed to guide or instruct the user during an interaction. Customizing the system prompt is a way for you to influence the user experience and ensure that users understand the capabilities and context of the chatbot. Please refer to the system prompt section for more information. |
| Temperature | temperature controls the variance of the model’s responses. A higher temperature makes the response more creative and less predictable. A lower temperature makes the response more predictable. |
| Throttling | throttling happens when there are too many inputs into the chatbot at once. All chatbots have a quota of many tokens that can be used. If you do a live synchronous session where a lot of users are using AI chatbots at once, you run the risk of running out of tokens. If users are using it asynchronously, the risk is less. If you run into the issue, please clear the chat history and you will be able to use the chatbot. |
| Tokens | tokens are single units that serve as building blocks for a sentence or document (such as a words or characters). |
| User |
user refers to one of the roles that a user can be granted. This role provides the minimum level of access to the chatbot which means that you can’t adjust any of the model’s settings. |