Gen AI Prompt Techniques

Tree of Thoughts

Definition: A method involving branching exploration in problem-solving by evaluating multiple potential paths or ideas before settling on an optimal solution. Example: "Consider multiple potential solutions for a sustainable city design, evaluating pros and cons before deciding on a final plan."

Text Classification

Definition: Assigning predefined categories or labels to text based on its content. Example: "Classify this text document as either 'spam' or 'not spam.'"

Text Translation

Definition: Automatically converting text from one language to another. Example: "Translate the following sentence into Spanish: 'I am going to the library.'"

Flipped Interaction

Definition: Changing the role of the AI and the user in an interaction, such as the user responding to AI-generated prompts, instead of the AI responding to user prompts. Example: "The AI asks the user a series of questions based on a topic, and the user responds."

Text Summarization

Definition: Condensing a larger piece of text into a shorter summary that retains key information and concepts. Example: "Summarize this article about climate change into three sentences."

Audience Persona

Definition: Creating a persona involves defining characteristics of the intended target audience to tailor content appropriately. Example: "Write a blog post about eco-friendly travel tips for young professionals interested in sustainability."

Chunking

Definition: Dividing larger pieces of text into smaller, manageable units or chunks to process information effectively. Example: "Chunk the following paragraph into key sentences: 'Artificial intelligence is rapidly evolving. It has applications in healthcare, finance, and more. Researchers are continually exploring new possibilities.'"

Chain of Thought

Definition: Encouraging the AI to generate responses by reasoning step by step, often improving logic and clarity in complex tasks. Example: "Explain the process of photosynthesis step by step."

Self-Consistency

Definition: Ensuring consistent and coherent responses from an AI across different prompts or scenarios, maintaining reliability and logical alignment. Example: "Generate a description of a character that remains consistent across different prompts, such as personality traits, background, and appearance."

Text Expansion

Definition: Extending and developing a piece of text by adding more detail or content. Example: "Expand on the topic: 'The importance of healthy eating habits in children.'"

Conditional Generation

Definition: Generating text based on specific conditions or input constraints. Example: "Generate a story with a happy ending that includes a dragon and a village."

Transfer Learning

Definition: Generating text based on specific conditions or input constraints. Example: "Generate a story with a happy ending that includes a dragon and a village."

Question Refinement

Definition: Improving or clarifying questions to better target the desired information or response from the AI. Example: "Refine the question 'What happened in 2020?' to 'What were the major global events that occurred in 2020?'"

Text Sentiment Analysis

Definition: The process of determining the emotional tone conveyed by a piece of text, often categorizing it as positive, negative, or neutral. Example: "Analyze the sentiment of the following sentence: 'I love this product; it works perfectly.'"

Text Completion

Definition: The task of continuing or completing a given piece of text based on context, as done by predictive text models. Example: "The sun set over the hills, casting a golden hue across the valley, and..."

Few-Shot Learning

Definition: Training models to perform tasks after being exposed to only a few examples, unlike the extensive datasets required in traditional machine learning. Example: "Translate a phrase into Japanese after seeing only a few translation examples."

Open-Ended Prompt

Definition: allow for a wide range of potential responses, enabling creativity and broader exploration. Example: "Describe a day in the life of a fictional character."

Prompt

Definition: an input or instruction given to a generative AI model to elicit a desired response or output. It serves as the starting point for the AI's generation process. Example: "Write a poem about the ocean."

Close-Ended Prompt

Definition: limit the scope of responses, often expecting specific answers or outcomes. Example: "Is the sky blue today?"

Zero-Shot Learning

Definition: when a model performs a task without having seen any examples or specific training for that task beforehand. Example: "Translate 'Hello, how are you?' into French without any specific examples or dictionaries."