Update README.md

README.md

@@ -16,4 +16,30 @@ configs:
   data_files:
   - split: train
     path: data/train-*
----
+
+## Dataset Description
+
+This dataset contains parallel English–Spanish sentence pairs extracted from the **Universal Declaration of Human Rights (UDHR)**. It is intended for training and evaluating machine translation models, as well as other multilingual NLP tasks.
+
+## Dataset Summary
+
+The dataset contains bilingual sentence-aligned segments of the UDHR in both English and Spanish. The data is split into training and evaluation sets and is provided as plain-text files containing one sentence per line.
+
+## Dataset Structure
+
+Each example contains:
+- `en`: English sentence
+- `es`: Spanish sentence
+
+## Dataset Structure
+
+This dataset was constructed by extracting the publicly available English and Spanish versions of the Universal Declaration of Human Rights and aligning them at the sentence level using LF Aligner.
+
+## Intended use
+Training and evaluating machine translation models.
+
+## Limitations
+Small dataset restricted to a formal legal register
+
+
+