This repository contains sample datasets used across different stages of our framework, including Mask-Text Alignment Pre-training, Task-Specific Instruction Tuning, and Explainability Analysis. Below is a structured description of the contents:
This section provides examples used for the Mask-Text Alignment Pre-training phase.
qa_pairs_pretrain_mask_sample.json: A JSON file containing… See the full description on the dataset page:
https://huggingface.co/datasets/fanrui00/ForgeryGPT.