Scribd
Upload
23 views4 pages

LoGeOTcsSFmEnB1OYARqGg Google-Prompting-Essentials - Glossary

promoting

Uploaded by

Abdullah Qasim
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
23 views4 pages

LoGeOTcsSFmEnB1OYARqGg Google-Prompting-Essentials - Glossary

promoting

Uploaded by

Abdullah Qasim
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 4

Glossary

Google Prompting Essentials

Terms and definitions from all modules:

A
Analogous task: Prompting a gen AI tool with a similar or related problem as a guide to
solving a new problem

C
Chain-of-density prompting: An iterative prompting technique used to generate
increasingly detailed summaries of text, starting with a broad overview and gradually
incorporating more specific information

Chain-of-thought prompting: A prompting technique where a gen AI tool retraces its path
from input to output, listing out the reasoning behind its response step by step

Constraints: Specific parameters or limitations that direct a gen AI tool's output towards a
particular result

Context: Background information, details, or guidelines that inform a gen AI tool’s output,
specifically its audience, goal, tone, and structure

Contextual understanding: A gen AI tool’s ability to comprehend and interpret information


within a specific context

D
Data validation: The process of ensuring data is accurate, complete, consistent, and meets
specific criteria before it's used
E
Evaluation: The processes of critically assessing an output to confirm it’s accurate,
unbiased, sufficient, relevant, and consistent to any context or prompt provided

F
Few-shot prompting: A technique that provides two or more examples or references in a
prompt

Format: The structure, layout, or presentation of an output

G
Generative AI: A type of artificial intelligence that can create new content, such as text,
images, or music, by learning patterns from existing data

Greedy decoding: Prompting in which the sampling parameters have been set to ensure the
most likely outcome

H
Hallucinations: When a gen AI tool provides an output with false information

Human-in-the-loop approach: When a human thinks critically about how they use, verify,
and refine gen AI tools and their outputs

I
Image generation: When a gen AI tool produces images from textual descriptions or other
input data

Input: The data or information provided to the AI model to generate a desired response, or
output

Iteration: The continuous refinement of prompts to enhance the quality of AI-generated


outputs
L
Long-context window: Allows gen AI tools to retain and utilize more information from past
interactions when generating a response

M
Meta-prompting: The practice of prompting gen AI to assist in the design of more
successful, effective prompts

Modalities: Different mediums for conveying information, including text, image, video, or
sound, that you can mix and match in both inputs and outputs

Multimodal prompting: A technique that includes multiple modalities, such as text, audio,
and images, in the same prompt, used to transform one modality into another, like generating
images from text, or text from video

O
One-shot prompting: A technique that provides a single example or reference in a prompt

Output: The response provided by an AI model based on the provided data and information,
or input

P
Persona: The role or character you assign to a gen AI tool while prompting, guiding its
behavior, tone, and knowledge base to generate tailored responses that meet your specific
needs and objectives

Prompt: Input that provides instructions to the gen AI tool on how to generate output

Prompt chaining: The process of building on interconnected prompts to tap into a wider
more complex array of capabilities

Prompt framework: The prompting best practices formula comprised of the following
components: Task, Context, Reference, Evaluate, Iterate

Prompt generation strategies: Strategies to help generate a meta-prompt from square one
Prompt refinement strategies: Strategies to improve upon meta-prompts that have already
been designed

Prompt versioning: Keeping track of different versions of a prompt, providing the


opportunity to compare or refine them to improve their effectiveness

R
References: Supplementary resources provided to a gen AI tool that resemble or relate to a
desired output

S
Sampling parameters: Variables that expand or limit the pool of data that a gen AI tool
chooses from, and dictates the probability of individual options within that pool

T
Task: The action or objective a gen AI tool is meant to accomplish

Temperature: The sampling parameter that determines the degree of randomness in gen AI’s
selection process, a higher temperature encouraging greater variation in gen AI’s random
selection

Tokens: The fundamental building blocks of text that AI models use to process and
understand language, which can be as small as a single character, a part of a word, or even
multiple words

Top-k value: The sampling parameter that dictates the total number of tokens from which
gen AI makes its selections

Top-p value: The sampling parameter that dictates the total combined probability of tokens
from which AI selects

Tree-of-thought prompting: A prompting technique in which gen AI shows multiple possible


outcomes to a prompt at once

You might also like