BASIC PROMPT ENGINEERING
whenever we give the proper input prompt in LLM or any other models then only we get a best expected output.
Now see the basic prompt engineering:
(step1)-Role prompt
(step2)-Instruction prompt
(step3)-Contex prompt
(step4)-Output prompt
This is the basic prompting now see what is inside the prompting:
- Role prompt — First you should give role to a model
EG:i am just giving your a Data quality editor(prompt)
2.Instuction prompt — to give a instuction like (Please analyse the dataset in json format)
EG:
check For:
- null or missing values
- duplicate rows
- iconsistance data
- formatting issues
- misconfiguration across orders
3.context prompt — just give a sample data to a model(Gemini+LLM) for processing.
EG:
{sample_data}
Get Babuloob’s stories in your inbox
Join Medium for free to get updates from this writer.
4.ouput prompt — To give a output formate what ever we want
EG:
I want a (column 1) and (column 2) in the sensitive column
{{
“sensitive_column”:[“column1”,”column 2”,…..]
}}
1. what are the Data classification
It calssify the data like
- sensitive
- confidential
- important
- non-sensitive
it is simple called data classification
2.what is called data pipeline
Raw data
|
Clean,transform
|
Move>datalake>sql
|
Final usage
First to get a raw data and to clean the data and tranform the data then move to a data lake when ever we want just query the output using in the database.
