proompting_table.csv

,Supervised Learning,Zero-Shot Classification,One-Shot Classification,Few-Shot Classification
Explanation,Providing a sufficiently large set of examples which the model learns from,Providing no examples,Providing one example,Providing a few examples
Example prompt,None,"Is the following social media post populist?
1 = Yes, populist
0 = No","Is the following social media post populist?
1 = Yes, populist
0 = No

Example: The crooked lobbyists must go! The people should be running this country!
Classification: 1","Is the following social media post populist?
1 = Yes, populist
0 = No

Example: The crooked lobbyists must go! The people should be running this country!
Classification: 1

Example: I've been a very very successful business person
Classification: 0"
Strengths,"-Precise Control
-Deterministic Output
-Exact Output","The model is free in how it interprets the prompt, which can mean it better choses from the range of possibilities","The model has an example to draw from, but because it is just one, it usually doesn't limit its flexibility",The model has clear-cut rules with regard to how to code the texts. Helps in streamlining the codes
Weaknesses,Requires large datasets,"Can lead to unwanted codes or completely unreliable coding

Output has to be parsed","Less open than zero-shot

Output has to be parsed","Too many examples can constrain the models ability to generalize beyond the examples

Output has to be parsed"