This PR adds a dataset card for MVI-Bench, including:
- Link to the paper: [MVI-Bench: A Comprehensive Benchmark for Evaluating Robustness to Misleading Visual Inputs in LVLMs](https://huggingface.co/papers/2511.14159).
- Link to the official GitHub repository.
- Metadata for the `image-text-to-text` task category.
- A summary of the benchmark's hierarchical levels and the MVI-Sensitivity metric.