Update README.md

README.md

@@ -27,14 +27,33 @@ configs:
 
 This dataset pairs UniMorph inflectional features with UniSegments segmentations. For languages without UniSegments coverage, segmentation defaults to the unsegmented word form itself.
 
+**This resource is a necessary component for evaluating Tokenizer Morphological Plausibility**, as introduced in *Tokenizer Morphological Plausibility* (https://arxiv.org/abs/2601.18536). The data generation process follows the implementation provided in the official repository of the paper:  
+https://github.com/abishekjs/morph-tok-eval/blob/main/data.py
+
 ## Sources
 
-- UniMorph: https://unimorph.github.io/
-- UniSegments: https://ufal.mff.cuni.cz/universal-segmentations
+- UniMorph: https://unimorph.github.io/  
+- UniSegments: https://ufal.mff.cuni.cz/universal-segmentations  
 
 ## Data Generation Logic
 
-For each language `lang` (ISO 639-3), the UniMorph source file is `{lang}.segmentations` when present, otherwise `{lang}`. When a UniSegments resource exists for a language, it is preferred from MorphyNet when multiple resources are available. For each UniMorph line, output `(form, features, segments)` where `form` is UniMorph column 2, `features` is UniMorph column 3, and `segments` comes from UniSegments column 4 with `" + "` replaced by `|`. If a language has no UniSegments file, `segments` is set to the unsegmented `form`. Output is a CSV without headers at `{lang}.csv`.
+For each language `lang` (ISO 639-3), the UniMorph source file is `{lang}.segmentations` when present, otherwise `{lang}`. When a UniSegments resource exists for a language, it is preferred from MorphyNet when multiple resources are available. For each UniMorph line, output:
+
+```
+(form, features, segments)
+```
+
+where:
+
+* `form` is UniMorph column 2
+* `features` is UniMorph column 3
+* `segments` comes from UniSegments column 4 with `" + "` replaced by `|`
+
+If a language has no UniSegments file, `segments` is set to the unsegmented `form`.
+
+Output is a CSV without headers at `{lang}.csv`.
+
+---
 
 ## Language Availability
 
@@ -242,3 +261,16 @@ Each language file inherits the more restrictive license between its UniMorph so
 - `.gitattributes` marks `*.csv` for Git LFS.
 - `.gitignore` is provided for common local artifacts.
 - `*.csv` files are per-language datasets at the repository root.
+
+
+## References
+
+Batsuren et al. (2022). *UniMorph 4.0: Universal Morphology*. Proceedings of the Thirteenth Language Resources and Evaluation Conference (LREC 2022), Marseille, France. European Language Resources Association. Pages 840–855.  
+https://aclanthology.org/2022.lrec-1.89/
+
+Žabokrtský et al. (2022). *Towards Universal Segmentations: UniSegments 1.0*. Proceedings of the Thirteenth Language Resources and Evaluation Conference (LREC 2022), Marseille, France. European Language Resources Association. Pages 1137–1149.  
+https://aclanthology.org/2022.lrec-1.122/
+
+Stephen, A., & Libovický, J. (2026). *Evaluating Morphological Plausibility of Subword Tokenization via Statistical Alignment with Morpho-Syntactic Features*. arXiv preprint arXiv:2601.18536.  
+https://arxiv.org/abs/2601.18536
+