Link dataset to paper and improve metadata

README.md

@@ -1,7 +1,14 @@
 ---
-license: cc-by-4.0
 language:
 - en
+license: cc-by-4.0
+size_categories:
+- 1K<n<10K
+task_categories:
+- text-classification
+- summarization
+- question-answering
+pretty_name: RTI-Bench
 tags:
 - legal
 - india
@@ -12,17 +19,12 @@ tags:
 - classification
 - dataset
 - civic-ai
-pretty_name: RTI-Bench
-size_categories:
-- 1K<n<10K
-task_categories:
-- text-classification
-- summarization
-- question-answering
 ---
 
 # RTI-Bench: A Structured Dataset for Indian RTI Decision Analysis
 
+This is the dataset repository for the paper **[RTI-Bench: A Structured Dataset for Indian Right-to-Information Decision Analysis](https://huggingface.co/papers/2605.16843)**.
+
 ## Dataset Description
 
 RTI-Bench is the first structured dataset of Central Information Commission (CIC) decisions under India's Right to Information Act, 2005. It supports research in legal NLP, civic AI, and AI-assisted access to justice.
@@ -200,4 +202,4 @@ Intended use: improving AI-assisted access to justice for Indian citizens. Not i
 
 ## License
 
-CC BY 4.0. Source data from CIC portal is public domain under Indian government open data policy. Source A derived from `jatinmehra/RTI-CASE-DATASET` (original license applies to that subset).
+CC BY 4.0. Source data from CIC portal is public domain under Indian government open data policy. Source A derived from `jatinmehra/RTI-CASE-DATASET` (original license applies to that subset).