- RISE (Role, Input, Steps, Execution)
- Role: Define the model’s role, such as a professor or expert.
- Input: Provide necessary background information.
- Steps: Outline the specific actions that need to be taken.
- Execution: Describe the desired outcome or result.
- GLUE (Goal, List, Unpack, Examine)
- Goal: Clearly specify the main objective you want to achieve.
- List: Provide guidelines or criteria for the model to follow.
- Unpack: Break down complex ideas into simpler components.
- Examine: Set standards for evaluating the model’s responses.
- ITAP (Input, Task, Annotation, Prediction)
- Input: Define the data or context for interaction.
- Task: Specify what action is required from the model.
- Annotation: Include relevant tags or labels to guide the model.
- Prediction: Indicate the expected format of the output.
- APE (Action, Purpose, Expectation)
- Action: Describe what needs to be done by the model.
- Purpose: State the goal of this action clearly.
- Expectation: Define what you expect the outcome to look like.
- RACE (Role, Action, Context, Expectations)
- Role: Specify the AI’s role in the interaction.
- Action: Detail the necessary actions to be taken.
- Context: Provide situational details that are relevant.
- Expectations: Describe what results you anticipate.
- COAST (Character, Objectives, Actions, Scenario, Task)
- This framework focuses on providing context while defining goals and tasks for clarity in prompts.
- TRACE (Task, Request, Action, Context, Example)
- Task: Clearly define the main task at hand.
- Request: Describe what you need from the AI specifically.
- Action: Outline any specific actions required from the model.
- Context: Provide background information that may help.
- Example: Illustrate your request with examples for better understanding.
- TAG (Task, Action, Goal)
- A straightforward framework that focuses on defining tasks and expected outcomes in a concise manner.
- STAR (Situation, Task, Action, Result)
- This framework outlines a situation followed by tasks and actions that lead to a specific result.
- Persona Framework
- This approach involves assigning a persona to the AI model to set an appropriate level of expertise and perspective for various tasks.
Benefits of Using Frameworks
- Clarity and Consistency: Structured approaches help reduce ambiguity in prompts and lead to more predictable outputs from AI models.
- Improved Output Quality: Techniques such as few-shot learning enhance the relevance and quality of responses generated by AI.
- Streamlined Workflow: These frameworks facilitate efficient creation and refinement of prompts.
These frameworks serve as valuable tools for effectively communicating with AI models and optimizing their outputs based on user needs.
