This PR improves the dataset card for `MultiTableQA-TabFact` by:
- Adding `task_categories: ['table-question-answering']` and `language: ['en']` to the metadata for better discoverability.
- Including relevant `tags` such as `rag`, `tables`, `multi-table`, `question-answering`, `fact-checking`, and `llm`.
- Updating the paper link text to the full paper title: `[RAG over Tables: Hierarchical Memory Index, Multi-Stage Retrieval, and Benchmarking](https://arxiv.org/abs/2504.01346)`.
- Adding a link to the Hugging Face Collection for the entire MultiTableQA benchmark.
- Clarifying the dataset's description to specify that this repository contains `MultiTableQA-TabFact`, one component of the broader MultiTableQA benchmark, focusing on table fact-checking tasks.
- Adding a "Sample Usage" section with instructions for data preparation, extracted directly from the GitHub README, to help users get started with the dataset.
- Correcting the `title` in the BibTeX citation to match the official p…
