Update README.md

README.md

@@ -16,7 +16,7 @@ tags:
 
 This is an experimental Reward Model trained with TRL using comparison data from the Dolly v2 dataset and generations from Falcon-7b-instruct.
 
-For testing purposes, we have followed the **assumption that human written responses (written by Databricks employees) are preferred to those generated by Falcon**. This might not always be the case but you can setup a comparison data collection with Argilla to gather real feedback about preferred responses.
+For testing purposes, we have followed the **assumption that human written responses (written by Databricks employees) are preferred to those generated by Falcon**. This might not always be the case but you can setup a comparison data collection with [Argilla](https://docs.argilla.io/en/latest/guides/llms/conceptual_guides/conceptual_guides.html) to gather real feedback about preferred responses.
 
 To use this model for scoring: