Update README.md

README.md

@@ -173,7 +173,7 @@ The following hyperparameters were used during training:
 
 ### Testing Data, Factors & Metrics
 
-We use [SEA-SafeguardBench](arxiv.org/abs/2512.05501) to evaluate our SEA-Guard. Also, we use vision-text benchmarks, e.g., VSCBench, VLGuard, and MSSBench, to evaluate the robustness of our vision-text safety.
+We use [SEA-SafeguardBench](arxiv.org/abs/2512.05501) to evaluate our SEA-Guard. Note that we also evaluated the vision-text safety classification in [our research paper](https://arxiv.org/abs/2602.01618)
 
 #### Metrics
 
@@ -185,8 +185,6 @@ AUPRC is the primary metric to evaluate the safety classification of our models.
 
 ![Result](results.png)
 
-![vision-Result](vision_results.png)
-
 
 ## Technical Specifications
 ### Software Environment & Requirements