Add task category, paper and code links

README.md

@@ -1,7 +1,13 @@
+```yaml
 ---
 language:
 - en
 pretty_name: PokerBench
+task_categories:
+- other
+tags:
+- poker
+- decision-making
 ---
 # PokerBench Overview
 
@@ -11,6 +17,9 @@ This dataset contains natural language game scenarios and optimal decisions comp
 
 - CSV files: Contain structured game information from which the JSON files were generated. The pre-flop and post-flop CSV files have different structures to accommodate the different stages of the game.
 
+Paper: https://huggingface.co/papers/2501.08328
+Code: https://github.com/pokerllm/pokerbench
+
 # Dataset Structure
 
 ## JSON Files
@@ -81,4 +90,5 @@ HJ/2.0bb/BB/call, Ks7h2d, Jc, 7c, OOP, OOP_CHECK/IP_CHECK/dealcards/Jc/OOP_CHECK
 
 ## Usage
 
-The dataset can be used to train and evaluate language models for decision-making in No Limit Texas Hold’em. Use the JSON files for direct training and evaluation with natural language prompts and decisions. Use the CSV files for more detailed analysis or for generating custom prompts.
+The dataset can be used to train and evaluate language models for decision-making in No Limit Texas Hold’em. Use the JSON files for direct training and evaluation with natural language prompts and decisions. Use the CSV files for more detailed analysis or for generating custom prompts.
+```