Update concise TabXLLM dataset card

README.md

@@ -1 +1,9 @@
-[Paper](https://arxiv.org/abs/2510.17385) · [Code](https://github.com/HKUSTGZ-ML4Health-Lab/TabXLLM) · [Model](https://huggingface.co/conanca1/TabXLLM-8B)
+# TabXLLM
+
+Training, validation, and test datasets for TabXLLM, supporting tabular classification and regression with Permutation-based Reinforcement Policy Optimization (PRPO).
+
+- Paper: [https://arxiv.org/abs/2510.17385](https://arxiv.org/abs/2510.17385)
+- Code: [https://github.com/HKUSTGZ-ML4Health-Lab/TabXLLM](https://github.com/HKUSTGZ-ML4Health-Lab/TabXLLM)
+- Model: [https://huggingface.co/conanca1/TabXLLM-8B](https://huggingface.co/conanca1/TabXLLM-8B)
+
+This release contains `train.parquet`, `val.parquet`, and the `test parquet/` folder with 139 test datasets.