Update README.md

README.md

@@ -161,6 +161,7 @@ This dataset contains 10 million exchanges with more than 216 million words, spr
 
 Although filtering reduced the exchanges by about 60% (leaving roughly 10% of the full data dump), this dataset is still intended as a large-scale dump. For best training results, further curation to target high-signal data relevant to your goals is recommended.
 
+---
 
 ## License