Update README.md

README.md

@@ -9,7 +9,7 @@ MRAMG-Bench is a comprehensive multimodal benchmark with six carefully curated d
 
 ## **Data Structure**
 
-The dataset consists of three major components: **Documents, Multimodal Question Answering (MQA), and Images**. Each component is structured across six different sub-datasets, ensuring a diverse and comprehensive collection of multimodal content.
+The dataset consists of three major components: **Documents, Multimodal QA pairs, and Images**. Each component is structured across six different sub-datasets, ensuring a diverse and comprehensive collection of multimodal content.
 
 ---