Add dataset card (auto-generated)

README.md

@@ -1,31 +1,25 @@
----
-dataset_info:
-  config_name: asdiv
-  features:
-  - name: body
-    dtype: string
-  - name: question
-    dtype: string
-  - name: solution_type
-    dtype: string
-  - name: answer
-    dtype: string
-  - name: formula
-    dtype: string
-  splits:
-  - name: train
-    num_bytes: 50040.117136659435
-    num_examples: 230
-  - name: test
-    num_bytes: 451448.88286334055
-    num_examples: 2075
-  download_size: 299086
-  dataset_size: 501489.0
-configs:
-- config_name: asdiv
-  data_files:
-  - split: train
-    path: asdiv/train-*
-  - split: test
-    path: asdiv/test-*
----
+# ASDiv (train/test 1:9)
+
+This dataset is derived from `EleutherAI/asdiv` by splitting the original `validation` split
+into `train` and `test` with a ratio of 1:9.
+
+## Source
+
+- Original dataset: `EleutherAI/asdiv`  
+  Link: https://huggingface.co/datasets/EleutherAI/asdiv
+
+## License
+
+- Inherits the original dataset's license (CC-BY-NC-4.0) unless otherwise noted in this repository.
+
+## Splitting Details
+
+- Method: `datasets.Dataset.train_test_split`
+- Source split: `validation`
+- Test size: 90.0%
+- Shuffle: yes
+- Seed: configurable via CLI (default 42)
+
+## Provenance
+
+- Generated at: 2025-11-03 09:05 UTC