Update README.md

README.md

@@ -7,7 +7,7 @@ tags:
 - hate speech
 - offensive language
 widget:
-- text: Credo che le donne debbano essere tutelate maggiormente.
+- text: Apprezzo il lavoro delle donne nella nostra comunità.
   example_title: Misogyny Detection Example 1
 pipeline_tag: text-classification
 ---
@@ -59,5 +59,10 @@ Please use the following BibTeX entry if you use this model in your project:
 }
 ```
 
+# Limitations
+
+Entropy-Attention Regularization mitigates lexical overfitting but does not completely remove it. We expect the model still to show biases, e.g., peculiar keywords that induce a specific prediction regardless of the context. \ 
+Please refer to our paper for a quantitative evaluation of this mitigation. 
+
 ## License 
 [GNU GPLv3](https://choosealicense.com/licenses/gpl-3.0/)