Update README.md

README.md

@@ -25,7 +25,8 @@ It is introduced in our paper [Exploring the Capabilities of LLM Encoders for Im
 ## Model Description
 
 LLM2Vec4CXR is a **bidirectional text encoder** fine-tuned with a `latent_attention` pooling strategy.  
-This design enhances semantic representation of chest X-ray reports, improving performance on clinical text similarity, retrieval, and interpretation tasks.
+This design enhances semantic representation of chest X-ray reports, making the model robust across different reporting styles and effective even with domain-specific abbreviations.  
+It improves performance on clinical text similarity, retrieval, and interpretation tasks.
 
 ### Key Features