Add task_categories and remove arxiv tag from metadata

This PR updates the YAML metadata of the dataset card by:
- Adding the `task_categories` field with `text-generation`.
- Removing the `arxiv:2610.08585` tag from the metadata, as arXiv IDs should only be referenced in the Markdown content.

The existing Markdown content, including the paper link, citation, license information, and usage example, is preserved unchanged.

README.md

@@ -2,12 +2,13 @@
 language:
 - en
 license: cc-by-4.0
+pretty_name: MedDistractNotes
 tags:
 - medical
 - robustness
 - benchmark
-- arxiv:2610.08585
-pretty_name: MedDistractNotes
+task_categories:
+- text-generation
 configs:
 - config_name: pairs
   data_files:
@@ -95,4 +96,4 @@ Code: https://github.com/nyuolab/llm_distract (the public toolkit release is bei
   year={2026}, eprint={2610.08585}, archivePrefix={arXiv}, primaryClass={cs.CL},
   url={https://arxiv.org/abs/2610.08585}
 }
-```
+```