Link paper, project page, and code repository

This PR improves the dataset card for BehaviorBench by adding links at the top of the README to:
- The paper: [BehaviorBench: Benchmarking Foundation Models for Behavioral Science Tasks](https://huggingface.co/papers/2606.24162)
- The [Project Page](https://umich-foreseer.github.io/behaviorbench/)
- The official [GitHub repository](https://github.com/umich-foreseer/behaviorbench_eval)

These additions will make it easier for researchers and practitioners to find the evaluation suite, associated paper details, and code.

README.md

@@ -1,25 +1,27 @@
 ---
-license: cc-by-nc-nd-4.0
 language:
-  - en
-pretty_name: BehaviorBench
+- en
+license: cc-by-nc-nd-4.0
 size_categories:
-  - 10K<n<100K
+- 10K<n<100K
 task_categories:
-  - text-generation
-  - question-answering
-  - multiple-choice
+- text-generation
+- question-answering
+- multiple-choice
+pretty_name: BehaviorBench
 tags:
-  - behavioral-science
-  - benchmark
-  - foundation-models
-  - personality
-  - economic-games
-  - scientific-workflows
+- behavioral-science
+- benchmark
+- foundation-models
+- personality
+- economic-games
+- scientific-workflows
 ---
 
 # BehaviorBench
 
+[**Project Page**](https://umich-foreseer.github.io/behaviorbench/) | [**Paper**](https://huggingface.co/papers/2606.24162) | [**Code**](https://github.com/umich-foreseer/behaviorbench_eval)
+
 BehaviorBench is a benchmark for evaluating large language models on behavioral
 science tasks. It bundles four data sources covering personality and survey
 response prediction (Big Five), economic-game decision making (MobLab),
@@ -143,4 +145,4 @@ behavior trigger a new minor version (current: `1.0.0`).
   author = {BeFM Authors},
   year   = {2026}
 }
-```
+```