Update README.md

README.md

@@ -1,29 +1,45 @@
 ---
-dataset_info:
-  features:
-  - name: original
-    dtype: string
-  - name: id
-    dtype: int64
-  - name: paraphrase
-    dtype: string
-  - name: broader
-    dtype: string
-  - name: technical
-    dtype: string
-  - name: explicit
-    dtype: string
-  - name: source
-    dtype: string
-  splits:
-  - name: train
-    num_bytes: 18458215
-    num_examples: 66100
-  download_size: 10686110
-  dataset_size: 18458215
-configs:
-- config_name: default
-  data_files:
-  - split: train
-    path: data/train-*
+license: apache-2.0
+language: en
+task_categories:
+- information-retrieval
+- text-generation
+tags:
+- rag
+- query-expansion
+- retrieval
+- synthetic-data
+- msmarco
 ---
+
+# RAG Query Transformation Dataset
+
+## Description
+This dataset contains synthetic query transformations generated to improve retrieval quality in Retrieval-Augmented Generation (RAG) systems.
+
+Each original query is paired with four transformed variants:
+- paraphrase
+- broader
+- technical
+- explicit
+
+## Source
+- Base queries: MS MARCO
+- Transformations: LLM-generated (Gemini Flash Lite)
+
+## Size
+- ~66k original queries
+- ~264k total transformed queries
+
+## Intended Use
+- Query expansion
+- Embedding-based retrieval
+- Training lightweight models (e.g., VAE) for query transformation
+
+## Limitations
+- Synthetic data
+- Some noise may be present
+- Not intended for direct QA
+
+## Citation
+If you use this dataset, please cite MS MARCO and this repository.