Update README.md

README.md

@@ -135,7 +135,7 @@ Outputs:
 ['I', 'Ġlove', 'Ġsalad']
 ```
 
-So I think this is not fundamentally linked to the model itself.
+The pretraining of LegalRoBERTa was restricted by the size of legal corpora available and the number of pretraining steps is small compared to the popular domain adapted models. This makes legalRoBERTa significantly **under-trained**. 
 
 ## BibTeX entry and citation info