Fix typo

README.md

@@ -122,7 +122,7 @@ for seq in sequences:
 
 ### Training Data
 
-Falcon-RW-7B was trained on 1,500B tokens of [RefinedWeb](https://huggingface.co/datasets/tiiuae/falcon-refinedweb), a high-quality filtered and deduplicated web dataset which we enhanced with curated corpora. Significant components from our curated copora were inspired by The Pile ([Gao et al., 2020](https://arxiv.org/abs/2101.00027)). 
+Falcon-7B was trained on 1,500B tokens of [RefinedWeb](https://huggingface.co/datasets/tiiuae/falcon-refinedweb), a high-quality filtered and deduplicated web dataset which we enhanced with curated corpora. Significant components from our curated copora were inspired by The Pile ([Gao et al., 2020](https://arxiv.org/abs/2101.00027)). 
 
 | **Data source**    | **Fraction** | **Tokens** | **Sources**                       |
 |--------------------|--------------|------------|-----------------------------------|