This PR improves the dataset card by:
- Linking the dataset to its associated paper: [GridVQA-X: A Framework for Evaluating Multimodal Explainability Methods](https://huggingface.co/papers/2606.14740)
- Adding the link to the official GitHub repository: [AikyamLab/grid-vqax](https://github.com/AikyamLab/grid-vqax)
- Adding the proper BibTeX citation.
- Updating the `task_categories` to include `image-text-to-text`.
