Hi there!
I'm opening this pull request to improve the model card for the MMBench2 World Model:
- Added the `pipeline_tag: image-to-video` in the YAML metadata to make the model easily discoverable under the correct category on the Hugging Face Hub.
- Added badges linking directly to the official GitHub repository (`nicklashansen/mmbench2`) and the arXiv paper at the top of the README to improve accessibility and documentation.
Scrub internal refs from args (e8_focused->targeted; drop internal path)