For each GPT model and each dataset, there are 5 output folders including:

1. baseline: This folder contains outputs using only component (1) Baseline prompts

2. baseline_definition_guideline： This folder contains outputs using components (1) Baseline prompts + (2) Annotation guideline-based prompts

3. baseline_definition_guideline_erroranalysis： This folder contains outputs using components (1) Baseline prompts + (2) Annotation guideline-based prompts + (3) Error analysis-based instructions

4. baseline_definition_guideline_erroranalysis_1shot： This folder contains outputs using components (1) Baseline prompts + (2) Annotation guideline-based prompts + (3) Error analysis-based instructions + (4) 1 annotated samples

5. baseline_definition_guideline_erroranalysis_5shot： This folder contains outputs using components (1) Baseline prompts + (2) Annotation guideline-based prompts + (3) Error analysis-based instructions + (4) 5 annotated samples