Tags
Create a set of instructions and customize tags to get the best responses.
Last updated
Create a set of instructions and customize tags to get the best responses.
Last updated
"Tags" is a feature designed to enhance the process of in a dynamic and customizable way. It allows users to define specific attributes or directives for the language model in a reusable and adaptable format, enabling them to exert more control over the model's responses without having to manually adjust the prompt each time.
For instance, if a user often requests text in French and of a certain length, they might create a tag with the field name "Language" and value "French", and another tag with the field name "Length" and value "1000 words". Then, whenever they want the output to be in French and approximately 1000 words long, they simply activate these tags, and the model will aim to generate responses according to these instructions.
This means users can create a library of tags each with specific instructions or requirements, and apply them as needed without rewriting their original prompts. It's like having a set of preset configurations or "macros" for the AI model.
Tags also provide a user-friendly interface to set default request parameters, enhancing the level of control users have over the AI model's responses beyond the textual content of the prompts.
Parameters like "temperature", and "stop sequence" are fundamental controls for fine-tuning the behavior of the model.
"Temperature" controls the randomness of the model's responses. A higher value makes the output more diverse and creative, while a lower value makes it more deterministic and focused.
"Stop sequence" is a parameter that can specify a certain sequence of tokens at which the model should stop generating further output.
By allowing these parameters to be set as default values under specific tags, users can easily adjust these fundamental controls in relation to the nature of their prompts. For example, a user might create a tag for creative brainstorming sessions with a high "temperature" value to encourage diverse ideas, or a tag for concise summaries with a specific stop indicator.
In essence, the "Tags" feature serves as a powerful tool for prompt engineering, streamlining the process of customizing prompts and making the language model more accessible and flexible for a wide range of user needs.
