diff --git a/group_agnostic_fairness/README.md b/group_agnostic_fairness/README.md
index 653322628692..cc9be22d8a7e 100644
--- a/group_agnostic_fairness/README.md
+++ b/group_agnostic_fairness/README.md
@@ -41,7 +41,7 @@ To generate various synthetic datasets used in the paper run './group_agnostic_f
 
 ## Training and Inference
 
-Training and prediction for the dual encoder model can be run as:
+Training and evaluation for the adversarial robust learning model can be run as:
 
 ```bash
 python -m group_agnostic_fairness/main_trainer