Prompt
Engineering
Getting started with
No-Code Deep Learning
Getting started with
- Simulink
- Network Designer App
Do you know ChatGPT ?
Generative Pre-Trained Transformer
How well do you know?
How well do you know?
Prompt Engineering
Prompt Engineers?
Prompt Engineers?
Prompt??
Prompt??
Prompt??
Prompt??
Prompt??
You are now going to introduce prompt engineering.
There is a workshop being conducted for anokha 2023 ( Amrita's annual tech fest) regarding prompt engineering, you will be now introducing prompt engineering in that workshop.
Now, generate some introductory paragraphs, with example prompts to motivate prompt engineering.
Introduction:
Elements of Prompt??
Instruction - a specific task or instruction you want the model to perform
Context - external information or additional context that can steer the model to better responses
Input Data - the input or question that we are interested to find a response for
Output Indicator - the type or format of the output.
Settings of Prompt??
Settings of Prompt??
Temperature and Top_k
High
Low
Alter any one
Tips for designing prompts
Tips for designing prompts
- Start Simple
- Instruction
- Specificity
- Avoid Impreciseness
- Don't tell what not to do
Advanced Prompting Techniques
Advanced Prompting Techniques
Zero shot Prompting
Advanced Prompting Techniques
Zero shot Prompting
Prompt:
Classify the text into neutral, negative or positive.
Text: I think the vacation is okay.
Sentiment:
Advanced Prompting Techniques
Zero shot Prompting
Prompt:
Classify the text into neutral, negative or positive.
Text: I think the vacation is okay.
Sentiment:
Output:
Neutral
Advanced Prompting Techniques
Few shot Prompting
Advanced Prompting Techniques
Few shot Prompting
Advanced Prompting Techniques
Few shot Prompting
Reasoning Tasks
Advanced Prompting Techniques
Chain-of-thought Prompting
Advanced Prompting Techniques
Chain-of-thought Prompting
Advanced Prompting Techniques
Self Consistency
Few shots + CoT
Advanced Prompting Techniques
Generated Knowledge Prompting
Advanced Prompting Techniques
Generate Knowledge Prompting
Demo
Pitfalls of LLMs
Pitfalls of LLMs
Citing Sources
No Internet access
Pitfalls of LLMs
Bias
stereotypical responses
Pitfalls of LLMs
Hallucinations
confidently give a wrong answer
Pitfalls of LLMs
Math
Bad at math
Models
Models
Flan (Finetuned Language Models)
Models
ChatGPT (Gen___ Pre___ T___)
Models
LLaMA (Large Language Model Meta AI)
Models
LaMDA (Language Model for Dialogue Applications)
Models
Many More
Midjourney
Have Time ?
Interested ?
Midjourney
/imagine prompt: [IMAGE PROMPT] [--OPTIONAL PARAMETERS]Midjourney
Aspect Ratio:
--ar [ratio] changes the default ratio (1:1) to a new ratio
Midjourney
Chaos:
--c [value] sets a chaos value that determines how much Midjourney varies the prompt
Midjourney
Quality:
--q [value] defines how much time will be spend generating the image
Midjourney
Seed:
--seed [value] sets a seed number which defines the starting point (noise field) for the image generation.
Midjourney
Stylize:
--stylize [value] or --s [value] influences how strongly Midjourney applies its artistic algorithm
Midjourney
Tile:
--tile generates images that can be used as repeating tiles to create seamless patterns for fabrics, wallpapers and textures
Gifts
Gifts
Gift 1:
Access Bard and Try it out
Gifts
Gift 2:
Images in _____
Gifts
Gift 2:
Images in ChatGPT