Sync from 2f022ed85fda9cc7bee5efe9707754d73b99d5fd

README.md

@@ -42,7 +42,7 @@ configs:
 </p>
 
 <p align="center">
-  <a href="https://anonymous.for.review"><img src="https://img.shields.io/badge/Paper-arXiv-B31B1B?logo=arxiv&logoColor=white" alt="Paper"></a>
+  <a href="https://arxiv.org/abs/2606.07167"><img src="https://img.shields.io/badge/Paper-arXiv-B31B1B?logo=arxiv&logoColor=white" alt="Paper"></a>
   <a href="https://mbzuai-nlp.github.io/UrduMMLU/"><img src="https://img.shields.io/badge/Project-Website-1F6FEB?logo=googlechrome&logoColor=white" alt="Project Website"></a>
   <a href="https://github.com/mbzuai-nlp/UrduMMLU"><img src="https://img.shields.io/badge/GitHub-Code-181717?logo=github&logoColor=white" alt="GitHub"></a>
 </p>
@@ -156,20 +156,18 @@ Each item retains its source attribution in the `source` field.
 - `length_tier` is `null` for items where a length bucket was not assigned.
 - Coverage is weighted toward humanities and the SSC levels, mirroring the
   availability of native Urdu exam content.
-- A small subset of questions in the Islamic Studies category contain Quranic/Classical Arabic text, reflecting the Pakistani Urdu-medium curriculum where original Arabic sources are commonly used. These questions constitute only a small fraction of the overall benchmark and do not materially change its characterization as an Urdu educational benchmark.
-
 
 ## Citation
 
 ```bibtex
 @misc{tabassum2026urdummlumassivemultitaskbenchmark,
-      title={UrduMMLU: A Massive Multitask Benchmark for Urdu Language Understanding}, 
+      title={UrduMMLU: A Massive Multitask Benchmark for Urdu Language Understanding},
       author={Ahmer Tabassum and Sarfraz Ahmad and Hasan Iqbal and Owais Aijaz and Momina Ahsan and Preslav Nakov},
       year={2026},
       eprint={2606.07167},
       archivePrefix={arXiv},
       primaryClass={cs.CL},
-      url={https://arxiv.org/abs/2606.07167}, 
+      url={https://arxiv.org/abs/2606.07167},
 }
 ```