IV-NLP: A Methodology to understand the behavior of DL models and its Application from a Causal Approach

These datasets are associated with an titled article "IV-NLP: A Methodology to understand the behavior of DL models and its Application from a Causal Approach" that will be submitted to MDPI—Electronics during this week. The final title of the article may be subject to modifications during the review and editing process. The dataset's DOI will remain unchanged, and the information will be updated once the article has been accepted.

This dataset (V4) extends and builds upon the previous dataset (V3), which was originally published in 2023. The previous version is available at DOI: https://data.mendeley.com/datasets/xh7vvty9zt/3.

The data sets hosted in this repository are the following:
The original dataset has 4015 records. The characteristics of the original dataset are as follows:
- Id (Automatic sequential from 0 to 4014)
- Tweet_Checked (Tweet Text)
- Fecha (Publication date)
- Clase_Argumento (0 and 1)

The synthetic dataset has 3966 records. The synthetic texts in this dataset were generated using the GPT-3.5-turbo-0125 model. The characteristics of the synthetic dataset are as follows:
- Id (Automatic sequential from 0 to 3965)
- Tweet_Checked (Tweet Text)
- Original_Id (Id that corresponds to the original data)
- Fecha (Publication date)
- Clase_Argumento (0 and 1)