Update README.md

README.md

@@ -40,6 +40,10 @@ The dataset is in JSON format.
 * `tgt`: output text
 
 
+## Considerations for Using the Data
+Please note that this dataset contains 69k instances (as opposed to the 82k instances we used in the paper). This is because this public release includes only the instances that were acquired and curated from publicly available datasets. Specifically, it is missing roughly 13k instances in training and 1.5k instances in validation data from Simplification and Formality Transfer tasks due to licensing restrictions.
+
+
 # Citation
 ```
 @article{raheja2023coedit,