Update README.md

README.md

@@ -32,6 +32,7 @@ dataset_info:
 This dataset is based on the translated MMARCO dataset. 
 500k queries (from [queries](https://huggingface.co/datasets/akhooli/ar_mmarco_queries)) were selected and scored against 32 documents each 
 (based on English scores) from 
-[Ar MMARCO Docs](https://huggingface.co/datasets/akhooli/ar_mmarco_docs) (translated) using JinaAI multilingual reranker.  
+[Ar MMARCO Docs](https://huggingface.co/datasets/akhooli/ar_mmarco_docs) (translated) using JinaAI multilingual reranker. The stats columns (mean, std and maxmin) are used to 
+curate subsets for training. 
 The dataset can be used to train knowledge distillation ColBERT models and some are available under the models section.  
 Note: translation quality is not good enough and scores are not spread apart. Some statistics were applied to curate a subset with wide distribution but this is not enough.