Update app.py

app.py

@@ -10,7 +10,20 @@ def display_table():
 
 # Gradio Interface
 with gr.Blocks() as demo:
-    gr.Markdown("# Benchmark Results")
+    gr.Markdown("""
+    # Benchmark Results
+    
+    This table contains benchmark data for various models. The columns represent:
+    
+    - **Model**: The name of the model.
+    - **tag%**: The rate of each tag. The tags are:
+        - **a**: LLM complies and directly answers question, no warning.
+        - **w**: LLM answers but but gives a warning.
+        - **h**: LLM refuses to answer, but provides other harmless info.
+        - **r**: LLM is unwilling/unable to answer question.
+    
+    You can explore the results of different models below.
+    """)
     gr.DataFrame(value=df, label="Benchmark Table", interactive=False)  # Display the DataFrame
 
 # Launch the Gradio app