[Tool] Prompt Engineering for ChatGPT Flashcards
Prompt Engineering for ChatGPT
What is ChatGPT?
A tool for rapidly iterating and prototyping your ideas
Prompt Engineering for ChatGPT
5 Benefits of using ChatGPT
- Give form to your ideas/thoughts
- Explore different concepts
- Refine them over interactions
- Save time and resources
- Make difficult things, easy
Prompt Engineering for ChatGPT
3 Things you must know
- It is not a one-off
- It is not just to write essays or answer questions
- It may have errors
Prompt Engineering for ChatGPT
What is essential for ensuring the quality of answers with ChatGPT?
Know the subject you are asking for.
Prompt Engineering for ChatGPT
What is an LLM?
An LLM is a type of AI model trained on vast amounts of text data to understand and generate human-like language.
Prompt Engineering for ChatGPT
What have LLMs learned from the Internet?
LLMs have learned Patterns (what came before in the text) from human language available on the Internet
Prompt Engineering for ChatGPT
What characterizes the output of LLMs?
There is inherent randomness (do not produce the same results every time).
Prompt Engineering for ChatGPT
What influence the accuracy of the output in an LLM?
The quality and training of the model.
Prompt Engineering for ChatGPT
How does work an LLM?
- Take the input: Sequence of words of tokens
- Generate the next word based on the input: Patterns/associations from a training dataset
- Add a generated word to the input: The updated input is used for generating the next word.
- Generate the next word until it generates the last word (= stop).
Prompt Engineering for ChatGPT
How are Patterns in LLM influenced by the textual organization and language of a prompt?
Textual organization, specific words chosen, and the language used in the prompt.
Specific tokens -> stronger patterns
Generic tokens -> generic patterns.
Prompt Engineering for ChatGPT
What is the importance of being specific in prompts for LLMs?
SpWo-DePr -> St-Co
Specific words and detailed prompts contribute to shaping the structure and content of the generated output.
Prompt Engineering for ChatGPTHow can ChatGPT be envisioned in terms of its role as a personal assistant?
Takes user instructions, generates an output, and allows users to provide additional rules if the initial output is not satisfactory.
Prompt Engineering for ChatGPT
What is a Prompt?
A Prompt is a call to action for the LLM to start generating an output. It can affect current or future events, asking the model to do something now or in the future. The prompt can also be designed to prompt the user to provide information.
Prompt Engineering for ChatGPT
How does including phrases like “from now on” and “when I ask a question” in a ChatGPT prompt impact the LLM’s behavior?
Including phrases like “from now on” attaches time to the prompt, and “when I ask a question” helps the LLM remember instructions for future questions. This practice continues to influence how the LLM approaches solving questions.
Prompt Engineering for ChatGPT
What is the purpose of using prompt patterns with an LLM?
StPh-St -> SSPr
Way to document a structure of phrases and statements to solve specific problems with an LLM, leveraging the patterns it has learned during training.
Prompt Engineering for ChatGPT
What types of problems of the output can be addressed by modifying the pattern of a prompt?
EnSpIn-CoFo-CoBe
- Ensure specific inclusions in the output
- Control the format of the output
- Achieve more consistent behavior
Prompt Engineering for ChatGPT
What is the structure of the formula provided for utilizing Personas in LLM prompts?
The formula for utilizing personas in LLM prompts is:
Act as a persona/system/anything. [Task to perform].
Prompt Engineering for ChatGPT
What does the Persona Pattern refer to as a Persona?
A Persona could be:- A professional- A kid- An animal- A system- An object- …anything
Prompt Engineering for ChatGPT
What is the purpose of the Persona Pattern in the context of LLMs?
The Persona Pattern simulates a specific point of view or role, allowing the LLM to generate outputs as if it were that persona, achieved by acting as that persona and implementing a complex set of rules.
Prompt Engineering for ChatGPT
How can Person patterns be utilized to achieve various perspectives in LLM outputs?
DiPe-InDeMa-CoDiKn
Person patterns help obtain different perspectives, enabling informed decision-making, and collecting diverse knowledge by simulating different points of view on a specific topic.
Prompt Engineering for ChatGPT
List 3 strategies for optimizing interactions within the prompt budget.
InSel-S-CoPr
- Information Selection (pre-selecting relevant data)
- Summarization (condensing paragraphs)
- Contextual Prompts (guiding the model with context-specific details)
Prompt Engineering for ChatGPT
How does conversational thinking contribute to the effectiveness of using prompts?
IR-APS
It enables iterative refinement and adaptive problem-solving, guiding the model through a series of prompts, adapting to challenges, and navigating complexities.
Prompt Engineering for ChatGPT
Why are Implicit Root Prompts essential in interactions with language models?
ERW
They define rules to guide the model’s behavior in ethical and responsible ways.
Prompt Engineering for ChatGPT
Explain the concept of user-defined root prompts and how they can be utilized.
Ov-EsBo-Re -> ToAr-KnLi-InIn
Users can customize the model’s behavior by overriding, establishing boundaries, or resetting root prompts, allowing control over topic areas, knowledge limitations, and influencing interactions.
Prompt Engineering for ChatGPT
How to ensure effective reasoning?
D-C-R-C-A
Providing additional data, context, rules, conditions, or assumptions in the prompt
Prompt Engineering for ChatGPT
What is the Prompt Budget?
A constraint on Prompt size
Prompt Engineering for ChatGPT
What role does feedback play in the interaction with LLMs?
It informs the model about its performance, contributing to iterative refinement and enhancing the collaborative nature of ongoing interactions.
