Update app.py

app.py

@@ -439,6 +439,14 @@ with gr.Blocks(title="Sahara Benchmark Leaderboards", css=google_style_css) as d
                     model_1_dd = gr.Dropdown(MODEL_NAME_LIST, label="Select Model 1", interactive=True)
                     model_2_dd = gr.Dropdown(MODEL_NAME_LIST, label="Select Model 2", interactive=True)
                 compare_btn = gr.Button("Compare")
+                # --- NEW: Add explanatory note for the 'Difference' column ---
+                explanation_note = """
+                **Note on the 'Difference' Column:**
+                * This value is calculated as: `(Score of Model 1) - (Score of Model 2)`.
+                * A positive value in <span style='color:green; font-weight:bold;'>green</span> indicates that **Model 1** performed better on that task.
+                * A negative value in <span style='color:red; font-weight:bold;'>red</span> indicates that **Model 2** performed better on that task.
+                """
+                gr.Markdown(explanation_note)
                 comparison_output = gr.HTML("<p style='text-align:center;'>Select two models and click Compare to see the results.</p>")
     
                 def update_comparison_table(m1, m2):