Improve dataset card with task category, license, and links

README.md

@@ -1,4 +1,15 @@
 ---
+license: mit
+task_categories:
+- text-to-text
+tags:
+  - code-translation
+  - assembly-language
+  - x86
+  - arm
+  - risc-v
+  - cisco-to-risc
+  - transpilation
 dataset_info:
   features:
   - name: file
@@ -20,6 +31,12 @@ configs:
     path: data/train-*
 ---
 
-Check out more datails here:
-- Paper: https://arxiv.org/abs/2506.14606
-- Code: https://github.com/ahmedheakl/Guaranteed-Guess
+This dataset contains pairs of x86 and ARM assembly code, used for training a language model to perform CISC-to-RISC transpilation.
+
+Check out more details here:
+
+- **Paper:** [Guaranteed Guess: A Language Modeling Approach for CISC-to-RISC Transpilation with Testing Guarantees](https://huggingface.co/papers/2506.14606)
+- **Code:** [https://github.com/ahmedheakl/Guaranteed-Guess](https://github.com/ahmedheakl/Guaranteed-Guess)
+- **Project Page:** [https://ahmedheakl.github.io/Guaranteed-Guess/](https://ahmedheakl.github.io/Guaranteed-Guess/)
+
+The dataset is structured as follows: Each example contains an x86 assembly instruction and its corresponding ARM assembly instruction. (Further details on the data structure can be found in the existing YAML metadata.)