Improve metadata and add links to paper and code

Hi! I'm Niels from the Hugging Face community science team. I noticed this dataset is part of the DLT-Corpus project.

This PR improves the dataset card by:
- Adding the `text-generation` task category to the YAML metadata.
- Specifying the `cc-by-nc-4.0` license in the metadata.
- Cleaning up the tag list in the metadata.
- Adding direct links to the research paper and the GitHub repository in the description.

These changes help make the dataset more discoverable and properly documented on the Hub!

README.md

@@ -1,4 +1,12 @@
 ---
+language:
+- en
+license: cc-by-nc-4.0
+size_categories:
+- 10M<n<100M
+task_categories:
+- text-generation
+pretty_name: Distributed Ledger Technology (DLT) / Blockchain Tweets
 dataset_info:
   features:
   - name: timestamp
@@ -34,26 +42,27 @@ configs:
   data_files:
   - split: train
     path: data/train-*
-language:
-- en
 tags:
-- Blockchain,
+- Blockchain
 - DLT
 - Cryptocurrencies
 - Bitcoin
 - Ethereum
 - Crypto
-pretty_name: Distributed Ledger Technology (DLT) / Blockchain Tweets
-size_categories:
-- 10M<n<100M
 ---
+
 # DLT-Tweets
 
+<p align="center">
+  <a href="https://huggingface.co/papers/2602.22045"><b>[Paper]</b></a> • 
+  <a href="https://github.com/dlt-science/DLT-Corpus"><b>[Code]</b></a>
+</p>
+
 ## Dataset Description
 
 ### Dataset Summary
 
-DLT-Tweets is a large-scale corpus of social media posts related to Distributed Ledger Technology (DLT). This dataset is part of the larger DLT-Corpus collection, designed to support NLP research, social computing studies, and public discourse analysis in the DLT domain.
+DLT-Tweets is a large-scale corpus of social media posts related to Distributed Ledger Technology (DLT). This dataset is part of the larger DLT-Corpus collection, designed to support NLP research, social computing studies, and public discourse analysis in the DLT domain. It was introduced in the paper [DLT-Corpus: A Large-Scale Text Collection for the Distributed Ledger Technology Domain](https://huggingface.co/papers/2602.22045).
 
 The dataset contains **22.03 million social media documents** with **1,120 million tokens** (1.12 billion tokens), spanning posts from **2013 to mid-2023**. All documents are in English and have been processed to protect user privacy.