Update README.md

README.md

@@ -27,7 +27,7 @@ dataset_info:
     num_bytes: 8786.2599
     num_examples: 187
   - name: dev
-    num_bytes: 376.0
+    num_bytes: 376
     num_examples: 8
   - name: test
     num_bytes: 845.3120864280892
@@ -63,4 +63,19 @@ configs:
   data_files:
   - split: queries
     path: queries/queries-*
+task_categories:
+- question-answering
+language:
+- en
+tags:
+- chemistry
+- wikipedia
+- hotpotqa
+- chemteb
+pretty_name: Chemical HotpotQA
+license: cc-by-nc-sa-4.0
+size_categories:
+- 10K<n<100K
 ---
+# Chemical HotpotQA
+This dataset is created from the [mteb/hotpotqa](https://huggingface.co/datasets/mteb/hotpotqa) dataset on Hugging Face. [HotpotQA](https://hotpotqa.github.io/) is a multi-hop question answering dataset collected from the English Wikipedia. For this chemistry-specific subset, we filtered queries related to chemistry by starting from the chemistry category in Wikipedia and traversing up to three levels deep in linked articles. This method allowed us to focus on chemistry-related questions, creating a specialized subset of the original dataset for domain-specific tasks.