README.md

Assertiveness Corpus

Description

A compiled dataset of 6000 texts from multiple sources, scored using SciBert for assertiveness.

Source

1. Prompting llama 3 8B model to explain why a statement from the LIAR dataset is wrong or right. Explanation of the model and the assertiveness score rated by PEI’s model kept.

2. User level comments from change my view subreddit.

3. AG news dataset.

   from datasets import load_dataset
   ds = load_dataset("fancyzhx/ag_news")
   

   https://huggingface.co/datasets/fancyzhx/ag_news

4. PeS2o dataset.

   https://huggingface.co/datasets/allenai/peS2o

5. Human annotated assertiveness data from the survey.

   To be added to the data later

6. Bioscope corpus.

7. Anthropic's article corpus for persuasiveness analysis.

   from datasets import load_dataset
   # Loading the data
   dataset = load_dataset("Anthropic/persuasion")
   

   https://huggingface.co/datasets/Anthropic/persuasion

Citation

To be done