- This post was human compiled 🙂
Giving a LLM (Large Language Model) like ChatGPT a simple one line instruction may not get you the highest quality results.
LLM prompting is an art and a science of it’s own.
Types of prompts: - Zero-shot
No examples provided with the instruction - Few-shot
Examples are provided along with the instruction - Chain-of-thought
A logical train of thought is given before the instruction or question
Or you can have a model write prompts for you:
https://huggingface.co/Kaludi/chatgpt-gpt4-prompts-bart-large-cnn-samsum
https://towardsdatascience.com/are-prompt-generated-by-large-language-models-llms-reliable-4162fd10c845
Prompting Resources:
https://en.wikipedia.org/wiki/Prompt_engineering
https://www.mihaileric.com/posts/a-complete-introduction-to-prompt-engineering/
https://thenewstack.io/prompt-engineering-get-llms-to-generate-the-content-you-want/
https://medium.com/nlplanet/two-minutes-nlp-making-large-language-models-reason-with-chain-of-thought-prompting-401fd3c964d0
https://openreview.net/forum?id=5NTt8GFjUHkr
https://www.inovex.de/de/blog/prompt-engineering-guide/
